id stringlengths 9 10 | submitter stringlengths 1 64 ⌀ | authors stringlengths 4 20.7k | title stringlengths 4 246 | comments stringlengths 1 523 ⌀ | journal-ref stringlengths 4 404 ⌀ | doi stringlengths 11 153 ⌀ | report-no stringlengths 2 254 ⌀ | categories stringlengths 5 98 | license stringclasses 9 values | orig_abstract stringlengths 14 3.35k | versions listlengths 1 60 | update_date stringlengths 10 10 | authors_parsed listlengths 1 1.35k | abstract stringlengths 11 3.34k |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2405.14384 | Marion Neumeier | Marion Neumeier, Sebastian Dorn, Michael Botsch, Wolfgang Utschick | Reliable Trajectory Prediction and Uncertainty Quantification with
Conditioned Diffusion Models | Accepted at IEEE/CVF Computer Vision and Pattern Recognition
Conference Workshops (CVPRW) 2024 | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This work introduces the conditioned Vehicle Motion Diffusion (cVMD) model, a
novel network architecture for highway trajectory prediction using diffusion
models. The proposed model ensures the drivability of the predicted trajectory
by integrating non-holonomic motion constraints and physical constraints into
the generative prediction module. Central to the architecture of cVMD is its
capacity to perform uncertainty quantification, a feature that is crucial in
safety-critical applications. By integrating the quantified uncertainty into
the prediction process, the cVMD's trajectory prediction performance is
improved considerably. The model's performance was evaluated using the publicly
available highD dataset. Experiments show that the proposed architecture
achieves competitive trajectory prediction accuracy compared to
state-of-the-art models, while providing guaranteed drivable trajectories and
uncertainty quantification.
| [
{
"created": "Thu, 23 May 2024 10:01:39 GMT",
"version": "v1"
}
] | 2024-05-24 | [
[
"Neumeier",
"Marion",
""
],
[
"Dorn",
"Sebastian",
""
],
[
"Botsch",
"Michael",
""
],
[
"Utschick",
"Wolfgang",
""
]
] | This work introduces the conditioned Vehicle Motion Diffusion (cVMD) model, a novel network architecture for highway trajectory prediction using diffusion models. The proposed model ensures the drivability of the predicted trajectory by integrating non-holonomic motion constraints and physical constraints into the generative prediction module. Central to the architecture of cVMD is its capacity to perform uncertainty quantification, a feature that is crucial in safety-critical applications. By integrating the quantified uncertainty into the prediction process, the cVMD's trajectory prediction performance is improved considerably. The model's performance was evaluated using the publicly available highD dataset. Experiments show that the proposed architecture achieves competitive trajectory prediction accuracy compared to state-of-the-art models, while providing guaranteed drivable trajectories and uncertainty quantification. |
1512.07331 | Suhas Sreehari | Suhas Sreehari, S. V. Venkatakrishnan, Brendt Wohlberg, Lawrence F.
Drummy, Jeffrey P. Simmons, Charles A. Bouman | Plug-and-Play Priors for Bright Field Electron Tomography and Sparse
Interpolation | 13 pages, 11 figures | null | 10.1109/TCI.2016.2599778 | null | cs.CV eess.IV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Many material and biological samples in scientific imaging are characterized
by non-local repeating structures. These are studied using scanning electron
microscopy and electron tomography. Sparse sampling of individual pixels in a
2D image acquisition geometry, or sparse sampling of projection images with
large tilt increments in a tomography experiment, can enable high speed data
acquisition and minimize sample damage caused by the electron beam.
In this paper, we present an algorithm for electron tomographic
reconstruction and sparse image interpolation that exploits the non-local
redundancy in images. We adapt a framework, termed plug-and-play (P&P) priors,
to solve these imaging problems in a regularized inversion setting. The power
of the P&P approach is that it allows a wide array of modern denoising
algorithms to be used as a "prior model" for tomography and image
interpolation. We also present sufficient mathematical conditions that ensure
convergence of the P&P approach, and we use these insights to design a new
non-local means denoising algorithm. Finally, we demonstrate that the algorithm
produces higher quality reconstructions on both simulated and real electron
microscope data, along with improved convergence properties compared to other
methods.
| [
{
"created": "Wed, 23 Dec 2015 02:06:29 GMT",
"version": "v1"
}
] | 2017-11-09 | [
[
"Sreehari",
"Suhas",
""
],
[
"Venkatakrishnan",
"S. V.",
""
],
[
"Wohlberg",
"Brendt",
""
],
[
"Drummy",
"Lawrence F.",
""
],
[
"Simmons",
"Jeffrey P.",
""
],
[
"Bouman",
"Charles A.",
""
]
] | Many material and biological samples in scientific imaging are characterized by non-local repeating structures. These are studied using scanning electron microscopy and electron tomography. Sparse sampling of individual pixels in a 2D image acquisition geometry, or sparse sampling of projection images with large tilt increments in a tomography experiment, can enable high speed data acquisition and minimize sample damage caused by the electron beam. In this paper, we present an algorithm for electron tomographic reconstruction and sparse image interpolation that exploits the non-local redundancy in images. We adapt a framework, termed plug-and-play (P&P) priors, to solve these imaging problems in a regularized inversion setting. The power of the P&P approach is that it allows a wide array of modern denoising algorithms to be used as a "prior model" for tomography and image interpolation. We also present sufficient mathematical conditions that ensure convergence of the P&P approach, and we use these insights to design a new non-local means denoising algorithm. Finally, we demonstrate that the algorithm produces higher quality reconstructions on both simulated and real electron microscope data, along with improved convergence properties compared to other methods. |
2107.07355 | Stefan Marksteiner | Stefan Marksteiner, Slava Bronfman, Markus Wolf, Eddie Lazebnik | Using Cyber Digital Twins for Automated Automotive Cybersecurity Testing | 6 pages, 3 figures, accepted for the joint SRCNAS/STRIVE workshop at
the 6th IEEE European Symposium on Security and Privacy | 2021 IEEE European Symposium on Security and Privacy Workshops
(EuroS&PW) - Safety vs Security in the Air and on the Ground | 10.1109/EuroSPW54576.2021.00020 | null | cs.CR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Cybersecurity testing of automotive systems has become a practical necessity,
with the wide adoption of advanced driving assistance functions and vehicular
communications. These functionalities require the integration of information
and communication technologies that not only allow for a plethora of on-the-fly
configuration abilities, but also provide a huge surface for attacks. Theses
circumstances have also been recognized by standardization and regulation
bodies, making the need for not only proper cybersecurity engineering but also
proving the effectiveness of security measures by verification and validation
through testing also a formal necessity. In order to keep pace with the rapidly
growing demand of neutral-party security testing of vehicular systems, novel
approaches are needed. This paper therefore presents a methodology to create
and execute cybersecurity test cases on the fly in a black box setting by using
pattern matching-based binary analysis and translation mechanisms to formal
attack descriptions as well as model-checking techniques. The approach is
intended to generate meaningful attack vectors on a system with next-to-zero a
priori knowledge.
| [
{
"created": "Thu, 15 Jul 2021 14:32:10 GMT",
"version": "v1"
}
] | 2021-09-07 | [
[
"Marksteiner",
"Stefan",
""
],
[
"Bronfman",
"Slava",
""
],
[
"Wolf",
"Markus",
""
],
[
"Lazebnik",
"Eddie",
""
]
] | Cybersecurity testing of automotive systems has become a practical necessity, with the wide adoption of advanced driving assistance functions and vehicular communications. These functionalities require the integration of information and communication technologies that not only allow for a plethora of on-the-fly configuration abilities, but also provide a huge surface for attacks. Theses circumstances have also been recognized by standardization and regulation bodies, making the need for not only proper cybersecurity engineering but also proving the effectiveness of security measures by verification and validation through testing also a formal necessity. In order to keep pace with the rapidly growing demand of neutral-party security testing of vehicular systems, novel approaches are needed. This paper therefore presents a methodology to create and execute cybersecurity test cases on the fly in a black box setting by using pattern matching-based binary analysis and translation mechanisms to formal attack descriptions as well as model-checking techniques. The approach is intended to generate meaningful attack vectors on a system with next-to-zero a priori knowledge. |
1810.11274 | Hao Chen | Hao Chen, Daniel Zelazo, Xiangke Wang, and Lincheng Shen | Convergence Analysis of Signed Nonlinear Networks | null | null | null | null | cs.SY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This work analyzes the convergence properties of signed networks with
nonlinear edge functions. We consider diffusively coupled networks comprised of
maximal equilibrium-independent passive (MEIP) dynamics on the nodes, and a
general class of nonlinear coupling functions on the edges. The first
contribution of this work is to generalize the classical notion of signed
networks for graphs with scalar weights to graphs with nonlinear edge functions
using notions from passivity theory. We show that the output of the network can
finally form one or several steady-state clusters if all edges are positive,
and in particular, all nodes can reach an output agreement if there is a
connected subnetwork spanning all nodes and strictly positive edges. When there
are non-positive edges added to the network, we show that the tension of the
network still converges to the equilibria of the edge functions if the relative
outputs of the nodes connected by non-positive edges converge to their
equilibria. Furthermore, we establish the equivalent circuit models for signed
nonlinear networks, and define the concept of equivalent edge functions which
is a generalization of the notion of effective resistance. We finally
characterize the relationship between the convergence property and the
equivalent edge function, when a non-positive edge is added to a strictly
positive network comprised of nonlinear integrators. We show that the
convergence of the network is always guaranteed, if the sum of the equivalent
edge function of the previous network and the new edge function is passive.
| [
{
"created": "Fri, 26 Oct 2018 11:38:58 GMT",
"version": "v1"
},
{
"created": "Thu, 31 Jan 2019 03:07:13 GMT",
"version": "v2"
},
{
"created": "Wed, 27 Mar 2019 05:46:44 GMT",
"version": "v3"
}
] | 2019-03-28 | [
[
"Chen",
"Hao",
""
],
[
"Zelazo",
"Daniel",
""
],
[
"Wang",
"Xiangke",
""
],
[
"Shen",
"Lincheng",
""
]
] | This work analyzes the convergence properties of signed networks with nonlinear edge functions. We consider diffusively coupled networks comprised of maximal equilibrium-independent passive (MEIP) dynamics on the nodes, and a general class of nonlinear coupling functions on the edges. The first contribution of this work is to generalize the classical notion of signed networks for graphs with scalar weights to graphs with nonlinear edge functions using notions from passivity theory. We show that the output of the network can finally form one or several steady-state clusters if all edges are positive, and in particular, all nodes can reach an output agreement if there is a connected subnetwork spanning all nodes and strictly positive edges. When there are non-positive edges added to the network, we show that the tension of the network still converges to the equilibria of the edge functions if the relative outputs of the nodes connected by non-positive edges converge to their equilibria. Furthermore, we establish the equivalent circuit models for signed nonlinear networks, and define the concept of equivalent edge functions which is a generalization of the notion of effective resistance. We finally characterize the relationship between the convergence property and the equivalent edge function, when a non-positive edge is added to a strictly positive network comprised of nonlinear integrators. We show that the convergence of the network is always guaranteed, if the sum of the equivalent edge function of the previous network and the new edge function is passive. |
2207.09869 | Tam\'as Matuszka Ph.D. | Tamas Matuszka, Daniel Kozma | A Novel Neural Network Training Method for Autonomous Driving Using
Semi-Pseudo-Labels and 3D Data Augmentations | null | null | 10.1007/978-3-031-21967-2_18 | null | cs.CV cs.LG | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Training neural networks to perform 3D object detection for autonomous
driving requires a large amount of diverse annotated data. However, obtaining
training data with sufficient quality and quantity is expensive and sometimes
impossible due to human and sensor constraints. Therefore, a novel solution is
needed for extending current training methods to overcome this limitation and
enable accurate 3D object detection. Our solution for the above-mentioned
problem combines semi-pseudo-labeling and novel 3D augmentations. For
demonstrating the applicability of the proposed method, we have designed a
convolutional neural network for 3D object detection which can significantly
increase the detection range in comparison with the training data distribution.
| [
{
"created": "Wed, 20 Jul 2022 13:04:08 GMT",
"version": "v1"
}
] | 2022-12-13 | [
[
"Matuszka",
"Tamas",
""
],
[
"Kozma",
"Daniel",
""
]
] | Training neural networks to perform 3D object detection for autonomous driving requires a large amount of diverse annotated data. However, obtaining training data with sufficient quality and quantity is expensive and sometimes impossible due to human and sensor constraints. Therefore, a novel solution is needed for extending current training methods to overcome this limitation and enable accurate 3D object detection. Our solution for the above-mentioned problem combines semi-pseudo-labeling and novel 3D augmentations. For demonstrating the applicability of the proposed method, we have designed a convolutional neural network for 3D object detection which can significantly increase the detection range in comparison with the training data distribution. |
1807.00948 | De'Aira Bryant | Tobi Ogunyale, De'Aira Bryant and Ayanna Howard | Does Removing Stereotype Priming Remove Bias? A Pilot Human-Robot
Interaction Study | 5 pages, 9 figures, 1 table, to be presented at the 5th Workshop on
Fairness, Accountability, and Transparency in Machine Learning (FAT/ML 2018),
Stockholm, Sweden, July 15, 2018 | null | null | null | cs.RO cs.HC | http://creativecommons.org/licenses/by/4.0/ | Robots capable of participating in complex social interactions have shown
great potential in a variety of applications. As these robots grow more
popular, it is essential to continuously evaluate the dynamics of the
human-robot relationship. One factor shown to have potential impacts on this
critical relationship is the human projection of stereotypes onto social
robots, a practice that is implicitly known to effect both developers and users
of this technology. As such, in this research, we wished to investigate the
difference in participants' perceptions of the robot interaction if we removed
stereotype priming. This has not yet been a common practice in similar studies.
Given the stereotypes of emotions among ethnic groups, especially in the U.S.,
this study specifically sought to investigate the impact that robot "skin
color" could potentially have on the human perception of a robot's emotional
expressive behavior. A between-subject experiment with 198 individuals was
conducted. The results showed no significant differences in the overall emotion
classification or intensity ratings for the different robot skin colors. These
results lend credence to our hypothesis that when individuals are not primed
with information related to human stereotypes, robots are evaluated based on
functional attributes versus stereotypical attributes. This provides some
confidence that robots, if designed correctly, can potentially be used as a
tool to override stereotype-based biases associated with human behavior.
| [
{
"created": "Tue, 3 Jul 2018 01:48:06 GMT",
"version": "v1"
}
] | 2018-07-04 | [
[
"Ogunyale",
"Tobi",
""
],
[
"Bryant",
"De'Aira",
""
],
[
"Howard",
"Ayanna",
""
]
] | Robots capable of participating in complex social interactions have shown great potential in a variety of applications. As these robots grow more popular, it is essential to continuously evaluate the dynamics of the human-robot relationship. One factor shown to have potential impacts on this critical relationship is the human projection of stereotypes onto social robots, a practice that is implicitly known to effect both developers and users of this technology. As such, in this research, we wished to investigate the difference in participants' perceptions of the robot interaction if we removed stereotype priming. This has not yet been a common practice in similar studies. Given the stereotypes of emotions among ethnic groups, especially in the U.S., this study specifically sought to investigate the impact that robot "skin color" could potentially have on the human perception of a robot's emotional expressive behavior. A between-subject experiment with 198 individuals was conducted. The results showed no significant differences in the overall emotion classification or intensity ratings for the different robot skin colors. These results lend credence to our hypothesis that when individuals are not primed with information related to human stereotypes, robots are evaluated based on functional attributes versus stereotypical attributes. This provides some confidence that robots, if designed correctly, can potentially be used as a tool to override stereotype-based biases associated with human behavior. |
2107.09265 | Ziqi Lu | Ziqi Lu, Qiangqiang Huang, Kevin Doherty, John Leonard | Consensus-Informed Optimization Over Mixtures for Ambiguity-Aware Object
SLAM | null | null | null | null | cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Building object-level maps can facilitate robot-environment interactions
(e.g. planning and manipulation), but objects could often have multiple
probable poses when viewed from a single vantage point, due to symmetry,
occlusion or perceptual failures. A robust object-level simultaneous
localization and mapping (object SLAM) algorithm needs to be aware of this pose
ambiguity. We propose to maintain and subsequently disambiguate the multiple
pose interpretations to gradually recover a globally consistent world
representation. The max-mixtures model is applied to implicitly and efficiently
track all pose hypotheses, but the resulting formulation is non-convex, and
therefore subject to local optima. To mitigate this problem, temporally
consistent hypotheses are extracted, guiding the optimization into the global
optimum. This consensus-informed inference method is applied online via
landmark variable re-initialization within an incremental SLAM framework,
iSAM2, for robust real-time performance. We demonstrate that this approach
improves SLAM performance on both simulated and real object SLAM problems with
pose ambiguity.
| [
{
"created": "Tue, 20 Jul 2021 05:23:20 GMT",
"version": "v1"
},
{
"created": "Wed, 8 Sep 2021 04:32:34 GMT",
"version": "v2"
}
] | 2021-09-09 | [
[
"Lu",
"Ziqi",
""
],
[
"Huang",
"Qiangqiang",
""
],
[
"Doherty",
"Kevin",
""
],
[
"Leonard",
"John",
""
]
] | Building object-level maps can facilitate robot-environment interactions (e.g. planning and manipulation), but objects could often have multiple probable poses when viewed from a single vantage point, due to symmetry, occlusion or perceptual failures. A robust object-level simultaneous localization and mapping (object SLAM) algorithm needs to be aware of this pose ambiguity. We propose to maintain and subsequently disambiguate the multiple pose interpretations to gradually recover a globally consistent world representation. The max-mixtures model is applied to implicitly and efficiently track all pose hypotheses, but the resulting formulation is non-convex, and therefore subject to local optima. To mitigate this problem, temporally consistent hypotheses are extracted, guiding the optimization into the global optimum. This consensus-informed inference method is applied online via landmark variable re-initialization within an incremental SLAM framework, iSAM2, for robust real-time performance. We demonstrate that this approach improves SLAM performance on both simulated and real object SLAM problems with pose ambiguity. |
2102.00423 | Reza Hadi Mogavi | Reza Hadi Mogavi, Xiaojuan Ma, Pan Hui | Characterizing Student Engagement Moods for Dropout Prediction in
Question Pool Websites | Accepted in the 24th ACM Conference on Computer-Supported Cooperative
Work and Social Computing (CSCW 2021) | null | 10.1145/3449086 | null | cs.HC cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Problem-Based Learning (PBL) is a popular approach to instruction that
supports students to get hands-on training by solving problems. Question Pool
websites (QPs) such as LeetCode, Code Chef, and Math Playground help PBL by
supplying authentic, diverse, and contextualized questions to students.
Nonetheless, empirical findings suggest that 40% to 80% of students registered
in QPs drop out in less than two months. This research is the first attempt to
understand and predict student dropouts from QPs via exploiting students'
engagement moods. Adopting a data-driven approach, we identify five different
engagement moods for QP students, which are namely challenge-seeker,
subject-seeker, interest-seeker, joy-seeker, and non-seeker. We find that
students have collective preferences for answering questions in each engagement
mood, and deviation from those preferences increases their probability of
dropping out significantly. Last but not least, this paper contributes by
introducing a new hybrid machine learning model (we call Dropout-Plus) for
predicting student dropouts in QPs. The test results on a popular QP in China,
with nearly 10K students, show that Dropout-Plus can exceed the rival
algorithms' dropout prediction performance in terms of accuracy, F1-measure,
and AUC. We wrap up our work by giving some design suggestions to QP managers
and online learning professionals to reduce their student dropouts.
| [
{
"created": "Sun, 31 Jan 2021 10:30:19 GMT",
"version": "v1"
},
{
"created": "Tue, 2 Feb 2021 19:15:09 GMT",
"version": "v2"
}
] | 2021-02-05 | [
[
"Mogavi",
"Reza Hadi",
""
],
[
"Ma",
"Xiaojuan",
""
],
[
"Hui",
"Pan",
""
]
] | Problem-Based Learning (PBL) is a popular approach to instruction that supports students to get hands-on training by solving problems. Question Pool websites (QPs) such as LeetCode, Code Chef, and Math Playground help PBL by supplying authentic, diverse, and contextualized questions to students. Nonetheless, empirical findings suggest that 40% to 80% of students registered in QPs drop out in less than two months. This research is the first attempt to understand and predict student dropouts from QPs via exploiting students' engagement moods. Adopting a data-driven approach, we identify five different engagement moods for QP students, which are namely challenge-seeker, subject-seeker, interest-seeker, joy-seeker, and non-seeker. We find that students have collective preferences for answering questions in each engagement mood, and deviation from those preferences increases their probability of dropping out significantly. Last but not least, this paper contributes by introducing a new hybrid machine learning model (we call Dropout-Plus) for predicting student dropouts in QPs. The test results on a popular QP in China, with nearly 10K students, show that Dropout-Plus can exceed the rival algorithms' dropout prediction performance in terms of accuracy, F1-measure, and AUC. We wrap up our work by giving some design suggestions to QP managers and online learning professionals to reduce their student dropouts. |
2104.04071 | Gautam Srivastava | Farrah Huntinghawk, Candace Richard, Sarah Plosker, Gautam Srivastava | Expanding Cybersecurity Knowledge Through an Indigenous Lens: A First
Look | 9 pages, 0 figures | 2020 IEEE CCECE, London, ON, Canada, 2020, pp. 1-4 | 10.1109/CCECE47787.2020.9255753. | null | cs.CY cs.CR | http://creativecommons.org/licenses/by/4.0/ | Decolonization and Indigenous education are at the forefront of Canadian
content currently in Academia. Over the last few decades, we have seen some
major changes in the way in which we share information. In particular, we have
moved into an age of electronically-shared content, and there is an increasing
expectation in Canada that this content is both culturally significant and
relevant. In this paper, we discuss an ongoing community engagement initiative
with First Nations communities in the Western Manitoba region. The initiative
involves knowledge-sharing activities that focus on the topic of cybersecurity,
and are aimed at a public audience. This initial look into our educational
project focuses on the conceptual analysis and planning stage. We are
developing a "Cybersecurity 101" mini-curriculum, to be implemented over
several one-hour long workshops aimed at diverse groups (these public workshops
may include a wide range of participants, from tech-adverse to tech-savvy).
Learning assessment tools have been built in to the workshop program. We have
created informational and promotional pamphlets, posters, lesson plans, and
feedback questionnaires which we believe instill relevance and personal
connection to this topic, helping to bridge gaps in accessibility for
Indigenous communities while striving to build positive, reciprocal
relationships. Our methodology is to approach the subject from a community
needs and priorities perspective. Activities are therefore being tailored to
fit each community.
| [
{
"created": "Tue, 30 Mar 2021 19:25:01 GMT",
"version": "v1"
}
] | 2021-04-12 | [
[
"Huntinghawk",
"Farrah",
""
],
[
"Richard",
"Candace",
""
],
[
"Plosker",
"Sarah",
""
],
[
"Srivastava",
"Gautam",
""
]
] | Decolonization and Indigenous education are at the forefront of Canadian content currently in Academia. Over the last few decades, we have seen some major changes in the way in which we share information. In particular, we have moved into an age of electronically-shared content, and there is an increasing expectation in Canada that this content is both culturally significant and relevant. In this paper, we discuss an ongoing community engagement initiative with First Nations communities in the Western Manitoba region. The initiative involves knowledge-sharing activities that focus on the topic of cybersecurity, and are aimed at a public audience. This initial look into our educational project focuses on the conceptual analysis and planning stage. We are developing a "Cybersecurity 101" mini-curriculum, to be implemented over several one-hour long workshops aimed at diverse groups (these public workshops may include a wide range of participants, from tech-adverse to tech-savvy). Learning assessment tools have been built in to the workshop program. We have created informational and promotional pamphlets, posters, lesson plans, and feedback questionnaires which we believe instill relevance and personal connection to this topic, helping to bridge gaps in accessibility for Indigenous communities while striving to build positive, reciprocal relationships. Our methodology is to approach the subject from a community needs and priorities perspective. Activities are therefore being tailored to fit each community. |
1708.02393 | Chadarat Phipathananunth | Panuchart Bunyakiati and Chadarat Phipathananunth | Cherry-Picking of Code Commits in Long-Running, Multi-release Software | 5 pages | null | 10.1145/3106237.3122818 | null | cs.SE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper presents Tartarian, a tool that supports maintenance of software
with long-running, multi-release branches in distributed version control
systems. When new maintenance code, such as bug fixes and code improvement, is
committed into a branch, it is likely that such code can be applied or reused
with some other branches. To do so, a developer may manually identify a commit
and cherry pick it. Tartarian can support this activity by providing commit
hashtags, which the developer uses as metadata to specify their intentions when
committing the code. With these tags, Tartarian uses dependency graph, that
represents the dependency constraints of the branches, and Branch Identifier,
which matches the commit hashtags with the dependency graph, to identify the
applicable branches for the commits. Using Tartarian, developers may be able to
maintain software with multiple releases more efficiently.
| [
{
"created": "Tue, 8 Aug 2017 07:43:31 GMT",
"version": "v1"
}
] | 2017-08-09 | [
[
"Bunyakiati",
"Panuchart",
""
],
[
"Phipathananunth",
"Chadarat",
""
]
] | This paper presents Tartarian, a tool that supports maintenance of software with long-running, multi-release branches in distributed version control systems. When new maintenance code, such as bug fixes and code improvement, is committed into a branch, it is likely that such code can be applied or reused with some other branches. To do so, a developer may manually identify a commit and cherry pick it. Tartarian can support this activity by providing commit hashtags, which the developer uses as metadata to specify their intentions when committing the code. With these tags, Tartarian uses dependency graph, that represents the dependency constraints of the branches, and Branch Identifier, which matches the commit hashtags with the dependency graph, to identify the applicable branches for the commits. Using Tartarian, developers may be able to maintain software with multiple releases more efficiently. |
2006.14784 | Peter Vaillancourt | Peter Z. Vaillancourt, J. Eric Coulter, Richard Knepper, Brandon
Barker | Self-Scaling Clusters and Reproducible Containers to Enable Scientific
Computing | Accepted for publication in the IEEE conference proceedings for HPEC
2020 | null | null | null | cs.DC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Container technologies such as Docker have become a crucial component of many
software industry practices especially those pertaining to reproducibility and
portability. The containerization philosophy has influenced the scientific
computing community, which has begun to adopt - and even develop - container
technologies (such as Singularity). Leveraging containers for scientific
software often poses challenges distinct from those encountered in industry,
and requires different methodologies. This is especially true for HPC. With an
increasing number of options for HPC in the cloud (including NSF-funded cloud
projects), there is strong motivation to seek solutions that provide
flexibility to develop and deploy scientific software on a variety of
computational infrastructures in a portable and reproducible way. The
flexibility offered by cloud services enables virtual HPC clusters that scale
on-demand, and the Cyberinfrastructure Resource Integration team in the XSEDE
project has developed a set of tools which provides scalable infrastructure in
the cloud. We now present a solution which uses the Nix package manager in an
MPI-capable Docker container that is converted to Singularity. It provides
consistent installations, dependencies, and environments in each image that are
reproducible and portable across scientific computing infrastructures. We
demonstrate the utility of these containers with cluster benchmark runs in a
self-scaling virtual cluster using the Slurm scheduler deployed in the
Jetstream and Aristotle Red Cloud OpenStack clouds. We conclude this technique
is useful as a template for scientific software application containers to be
used in the XSEDE compute environment, other Singularity HPC environments, and
cloud computing environments.
| [
{
"created": "Fri, 26 Jun 2020 03:57:19 GMT",
"version": "v1"
},
{
"created": "Mon, 3 Aug 2020 23:40:15 GMT",
"version": "v2"
}
] | 2020-08-05 | [
[
"Vaillancourt",
"Peter Z.",
""
],
[
"Coulter",
"J. Eric",
""
],
[
"Knepper",
"Richard",
""
],
[
"Barker",
"Brandon",
""
]
] | Container technologies such as Docker have become a crucial component of many software industry practices especially those pertaining to reproducibility and portability. The containerization philosophy has influenced the scientific computing community, which has begun to adopt - and even develop - container technologies (such as Singularity). Leveraging containers for scientific software often poses challenges distinct from those encountered in industry, and requires different methodologies. This is especially true for HPC. With an increasing number of options for HPC in the cloud (including NSF-funded cloud projects), there is strong motivation to seek solutions that provide flexibility to develop and deploy scientific software on a variety of computational infrastructures in a portable and reproducible way. The flexibility offered by cloud services enables virtual HPC clusters that scale on-demand, and the Cyberinfrastructure Resource Integration team in the XSEDE project has developed a set of tools which provides scalable infrastructure in the cloud. We now present a solution which uses the Nix package manager in an MPI-capable Docker container that is converted to Singularity. It provides consistent installations, dependencies, and environments in each image that are reproducible and portable across scientific computing infrastructures. We demonstrate the utility of these containers with cluster benchmark runs in a self-scaling virtual cluster using the Slurm scheduler deployed in the Jetstream and Aristotle Red Cloud OpenStack clouds. We conclude this technique is useful as a template for scientific software application containers to be used in the XSEDE compute environment, other Singularity HPC environments, and cloud computing environments. |
1710.02282 | Gabriele D'Angelo | Stefano Ferretti, Gabriele D'Angelo, Vittorio Ghini, Moreno Marzolla | The Quest for Scalability and Accuracy in the Simulation of the Internet
of Things: an Approach based on Multi-Level Simulation | Proceedings of the IEEE/ACM International Symposium on Distributed
Simulation and Real Time Applications (DS-RT 2017) | null | 10.1109/DISTRA.2017.8167672 | null | cs.PF cs.DC cs.MA cs.NI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper presents a methodology for simulating the Internet of Things (IoT)
using multi-level simulation models. With respect to conventional simulators,
this approach allows us to tune the level of detail of different parts of the
model without compromising the scalability of the simulation. As a use case, we
have developed a two-level simulator to study the deployment of smart services
over rural territories. The higher level is base on a coarse grained,
agent-based adaptive parallel and distributed simulator. When needed, this
simulator spawns OMNeT++ model instances to evaluate in more detail the issues
concerned with wireless communications in restricted areas of the simulated
world. The performance evaluation confirms the viability of multi-level
simulations for IoT environments.
| [
{
"created": "Fri, 6 Oct 2017 06:05:58 GMT",
"version": "v1"
},
{
"created": "Tue, 7 Aug 2018 07:12:41 GMT",
"version": "v2"
}
] | 2018-08-08 | [
[
"Ferretti",
"Stefano",
""
],
[
"D'Angelo",
"Gabriele",
""
],
[
"Ghini",
"Vittorio",
""
],
[
"Marzolla",
"Moreno",
""
]
] | This paper presents a methodology for simulating the Internet of Things (IoT) using multi-level simulation models. With respect to conventional simulators, this approach allows us to tune the level of detail of different parts of the model without compromising the scalability of the simulation. As a use case, we have developed a two-level simulator to study the deployment of smart services over rural territories. The higher level is base on a coarse grained, agent-based adaptive parallel and distributed simulator. When needed, this simulator spawns OMNeT++ model instances to evaluate in more detail the issues concerned with wireless communications in restricted areas of the simulated world. The performance evaluation confirms the viability of multi-level simulations for IoT environments. |
2002.02071 | Jiangsheng You Dr. | Jason You | Finite Hilbert Transform in Weighted L2 Spaces | null | null | null | null | cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Several new properties of weighted Hilbert transform are obtained. If mu is
zero, two Plancherel-like equations and the isotropic properties are derived.
For mu is real number, a coerciveness is derived and two iterative sequences
are constructed to find the inversion. The proposed iterative sequences are
applicable to the case of pure imaginary constant mu=i*eta with |eta|<pi/4 .
For mu=0.0 and 3.0 , we present the computer simulation results by using the
Chebyshev series representation of finite Hilbert transform. The results in
this paper are useful to the half scan in several imaging applications.
| [
{
"created": "Thu, 6 Feb 2020 02:13:18 GMT",
"version": "v1"
},
{
"created": "Tue, 11 Feb 2020 03:47:58 GMT",
"version": "v2"
}
] | 2020-02-12 | [
[
"You",
"Jason",
""
]
] | Several new properties of weighted Hilbert transform are obtained. If mu is zero, two Plancherel-like equations and the isotropic properties are derived. For mu is real number, a coerciveness is derived and two iterative sequences are constructed to find the inversion. The proposed iterative sequences are applicable to the case of pure imaginary constant mu=i*eta with |eta|<pi/4 . For mu=0.0 and 3.0 , we present the computer simulation results by using the Chebyshev series representation of finite Hilbert transform. The results in this paper are useful to the half scan in several imaging applications. |
1806.00194 | Chen Huang | Chen Huang, Yining Li, Chen Change Loy, Xiaoou Tang | Deep Imbalanced Learning for Face Recognition and Attribute Prediction | 14 pages, 10 figures, 8 tables. Accepted to TPAMI | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Data for face analysis often exhibit highly-skewed class distribution, i.e.,
most data belong to a few majority classes, while the minority classes only
contain a scarce amount of instances. To mitigate this issue, contemporary deep
learning methods typically follow classic strategies such as class re-sampling
or cost-sensitive training. In this paper, we conduct extensive and systematic
experiments to validate the effectiveness of these classic schemes for
representation learning on class-imbalanced data. We further demonstrate that
more discriminative deep representation can be learned by enforcing a deep
network to maintain inter-cluster margins both within and between classes. This
tight constraint effectively reduces the class imbalance inherent in the local
data neighborhood, thus carving much more balanced class boundaries locally. We
show that it is easy to deploy angular margins between the cluster
distributions on a hypersphere manifold. Such learned Cluster-based Large
Margin Local Embedding (CLMLE), when combined with a simple k-nearest cluster
algorithm, shows significant improvements in accuracy over existing methods on
both face recognition and face attribute prediction tasks that exhibit
imbalanced class distribution.
| [
{
"created": "Fri, 1 Jun 2018 04:55:47 GMT",
"version": "v1"
},
{
"created": "Tue, 30 Apr 2019 03:49:42 GMT",
"version": "v2"
}
] | 2019-05-01 | [
[
"Huang",
"Chen",
""
],
[
"Li",
"Yining",
""
],
[
"Loy",
"Chen Change",
""
],
[
"Tang",
"Xiaoou",
""
]
] | Data for face analysis often exhibit highly-skewed class distribution, i.e., most data belong to a few majority classes, while the minority classes only contain a scarce amount of instances. To mitigate this issue, contemporary deep learning methods typically follow classic strategies such as class re-sampling or cost-sensitive training. In this paper, we conduct extensive and systematic experiments to validate the effectiveness of these classic schemes for representation learning on class-imbalanced data. We further demonstrate that more discriminative deep representation can be learned by enforcing a deep network to maintain inter-cluster margins both within and between classes. This tight constraint effectively reduces the class imbalance inherent in the local data neighborhood, thus carving much more balanced class boundaries locally. We show that it is easy to deploy angular margins between the cluster distributions on a hypersphere manifold. Such learned Cluster-based Large Margin Local Embedding (CLMLE), when combined with a simple k-nearest cluster algorithm, shows significant improvements in accuracy over existing methods on both face recognition and face attribute prediction tasks that exhibit imbalanced class distribution. |
2309.12578 | Bokyeong Yoon | Bokyeong Yoon, Yoonsang Han, Gordon Euhyun Moon | SPION: Layer-Wise Sparse Training of Transformer via Convolutional Flood
Filling | null | null | null | null | cs.LG cs.DC | http://creativecommons.org/licenses/by/4.0/ | Sparsifying the Transformer has garnered considerable interest, as training
the Transformer is very computationally demanding. Prior efforts to sparsify
the Transformer have either used a fixed pattern or data-driven approach to
reduce the number of operations involving the computation of multi-head
attention, which is the main bottleneck of the Transformer. However, existing
methods suffer from inevitable problems, such as the potential loss of
essential sequence features due to the uniform fixed pattern applied across all
layers, and an increase in the model size resulting from the use of additional
parameters to learn sparsity patterns in attention operations. In this paper,
we propose a novel sparsification scheme for the Transformer that integrates
convolution filters and the flood filling method to efficiently capture the
layer-wise sparse pattern in attention operations. Our sparsification approach
reduces the computational complexity and memory footprint of the Transformer
during training. Efficient implementations of the layer-wise sparsified
attention algorithm on GPUs are developed, demonstrating a new SPION that
achieves up to 3.08X speedup over existing state-of-the-art sparse Transformer
models, with better evaluation quality.
| [
{
"created": "Fri, 22 Sep 2023 02:14:46 GMT",
"version": "v1"
}
] | 2023-09-25 | [
[
"Yoon",
"Bokyeong",
""
],
[
"Han",
"Yoonsang",
""
],
[
"Moon",
"Gordon Euhyun",
""
]
] | Sparsifying the Transformer has garnered considerable interest, as training the Transformer is very computationally demanding. Prior efforts to sparsify the Transformer have either used a fixed pattern or data-driven approach to reduce the number of operations involving the computation of multi-head attention, which is the main bottleneck of the Transformer. However, existing methods suffer from inevitable problems, such as the potential loss of essential sequence features due to the uniform fixed pattern applied across all layers, and an increase in the model size resulting from the use of additional parameters to learn sparsity patterns in attention operations. In this paper, we propose a novel sparsification scheme for the Transformer that integrates convolution filters and the flood filling method to efficiently capture the layer-wise sparse pattern in attention operations. Our sparsification approach reduces the computational complexity and memory footprint of the Transformer during training. Efficient implementations of the layer-wise sparsified attention algorithm on GPUs are developed, demonstrating a new SPION that achieves up to 3.08X speedup over existing state-of-the-art sparse Transformer models, with better evaluation quality. |
2312.12006 | Md.Rafiul Biswas Mr. | Md. Rafiul Biswas, Ashhadul Islam, Zubair Shah, Wajdi Zaghouani, Samir
Brahim Belhaouari | Can ChatGPT be Your Personal Medical Assistant? | 5 pages, 7 figures, two tables, Accepted on The International
Symposium on Foundation and Large Language Models (FLLM2023) | The International Symposium on Foundation and Large Language
Models (FLLM2023) https://fllm-conference.org/2023/ | null | null | cs.CL cs.SI | http://creativecommons.org/licenses/by/4.0/ | The advanced large language model (LLM) ChatGPT has shown its potential in
different domains and remains unbeaten due to its characteristics compared to
other LLMs. This study aims to evaluate the potential of using a fine-tuned
ChatGPT model as a personal medical assistant in the Arabic language. To do so,
this study uses publicly available online questions and answering datasets in
Arabic language. There are almost 430K questions and answers for 20
disease-specific categories. GPT-3.5-turbo model was fine-tuned with a portion
of this dataset. The performance of this fine-tuned model was evaluated through
automated and human evaluation. The automated evaluations include perplexity,
coherence, similarity, and token count. Native Arabic speakers with medical
knowledge evaluated the generated text by calculating relevance, accuracy,
precision, logic, and originality. The overall result shows that ChatGPT has a
bright future in medical assistance.
| [
{
"created": "Tue, 19 Dec 2023 09:54:27 GMT",
"version": "v1"
}
] | 2023-12-20 | [
[
"Biswas",
"Md. Rafiul",
""
],
[
"Islam",
"Ashhadul",
""
],
[
"Shah",
"Zubair",
""
],
[
"Zaghouani",
"Wajdi",
""
],
[
"Belhaouari",
"Samir Brahim",
""
]
] | The advanced large language model (LLM) ChatGPT has shown its potential in different domains and remains unbeaten due to its characteristics compared to other LLMs. This study aims to evaluate the potential of using a fine-tuned ChatGPT model as a personal medical assistant in the Arabic language. To do so, this study uses publicly available online questions and answering datasets in Arabic language. There are almost 430K questions and answers for 20 disease-specific categories. GPT-3.5-turbo model was fine-tuned with a portion of this dataset. The performance of this fine-tuned model was evaluated through automated and human evaluation. The automated evaluations include perplexity, coherence, similarity, and token count. Native Arabic speakers with medical knowledge evaluated the generated text by calculating relevance, accuracy, precision, logic, and originality. The overall result shows that ChatGPT has a bright future in medical assistance. |
2010.01247 | Zhun Deng | Zhun Deng, Cynthia Dwork, Jialiang Wang, Linjun Zhang | Interpreting Robust Optimization via Adversarial Influence Functions | null | null | null | null | cs.LG cs.AI stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Robust optimization has been widely used in nowadays data science, especially
in adversarial training. However, little research has been done to quantify how
robust optimization changes the optimizers and the prediction losses comparing
to standard training. In this paper, inspired by the influence function in
robust statistics, we introduce the Adversarial Influence Function (AIF) as a
tool to investigate the solution produced by robust optimization. The proposed
AIF enjoys a closed-form and can be calculated efficiently. To illustrate the
usage of AIF, we apply it to study model sensitivity -- a quantity defined to
capture the change of prediction losses on the natural data after implementing
robust optimization. We use AIF to analyze how model complexity and randomized
smoothing affect the model sensitivity with respect to specific models. We
further derive AIF for kernel regressions, with a particular application to
neural tangent kernels, and experimentally demonstrate the effectiveness of the
proposed AIF. Lastly, the theories of AIF will be extended to distributional
robust optimization.
| [
{
"created": "Sat, 3 Oct 2020 01:19:10 GMT",
"version": "v1"
}
] | 2020-10-06 | [
[
"Deng",
"Zhun",
""
],
[
"Dwork",
"Cynthia",
""
],
[
"Wang",
"Jialiang",
""
],
[
"Zhang",
"Linjun",
""
]
] | Robust optimization has been widely used in nowadays data science, especially in adversarial training. However, little research has been done to quantify how robust optimization changes the optimizers and the prediction losses comparing to standard training. In this paper, inspired by the influence function in robust statistics, we introduce the Adversarial Influence Function (AIF) as a tool to investigate the solution produced by robust optimization. The proposed AIF enjoys a closed-form and can be calculated efficiently. To illustrate the usage of AIF, we apply it to study model sensitivity -- a quantity defined to capture the change of prediction losses on the natural data after implementing robust optimization. We use AIF to analyze how model complexity and randomized smoothing affect the model sensitivity with respect to specific models. We further derive AIF for kernel regressions, with a particular application to neural tangent kernels, and experimentally demonstrate the effectiveness of the proposed AIF. Lastly, the theories of AIF will be extended to distributional robust optimization. |
2011.08529 | Zhaoyi Wan | Zhaoyi Wan, Yimin Chen, Sutao Deng, Kunpeng Chen, Cong Yao, Jiebo Luo | Slender Object Detection: Diagnoses and Improvements | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | In this paper, we are concerned with the detection of a particular type of
objects with extreme aspect ratios, namely \textbf{slender objects}. In
real-world scenarios, slender objects are actually very common and crucial to
the objective of a detection system. However, this type of objects has been
largely overlooked by previous object detection algorithms. Upon our
investigation, for a classical object detection method, a drastic drop of
$18.9\%$ mAP on COCO is observed, if solely evaluated on slender objects.
Therefore, we systematically study the problem of slender object detection in
this work. Accordingly, an analytical framework with carefully designed
benchmark and evaluation protocols is established, in which different
algorithms and modules can be inspected and compared. \New Our study reveals
that effective slender object detection can be achieved ~\textbf{with none of}
(1) anchor-based localization; (2) specially designed box representations.
Instead, \textbf{the critical aspect of improving slender object detection is
feature adaptation}. It identifies and extends the insights of existing methods
that are previously underexploited. Furthermore, we propose a feature adaption
strategy that achieves clear and consistent improvements over current
representative object detection methods.
| [
{
"created": "Tue, 17 Nov 2020 09:39:42 GMT",
"version": "v1"
},
{
"created": "Sat, 21 Nov 2020 05:33:07 GMT",
"version": "v2"
},
{
"created": "Thu, 24 Dec 2020 09:14:36 GMT",
"version": "v3"
},
{
"created": "Wed, 7 Apr 2021 02:35:15 GMT",
"version": "v4"
}
] | 2021-04-08 | [
[
"Wan",
"Zhaoyi",
""
],
[
"Chen",
"Yimin",
""
],
[
"Deng",
"Sutao",
""
],
[
"Chen",
"Kunpeng",
""
],
[
"Yao",
"Cong",
""
],
[
"Luo",
"Jiebo",
""
]
] | In this paper, we are concerned with the detection of a particular type of objects with extreme aspect ratios, namely \textbf{slender objects}. In real-world scenarios, slender objects are actually very common and crucial to the objective of a detection system. However, this type of objects has been largely overlooked by previous object detection algorithms. Upon our investigation, for a classical object detection method, a drastic drop of $18.9\%$ mAP on COCO is observed, if solely evaluated on slender objects. Therefore, we systematically study the problem of slender object detection in this work. Accordingly, an analytical framework with carefully designed benchmark and evaluation protocols is established, in which different algorithms and modules can be inspected and compared. \New Our study reveals that effective slender object detection can be achieved ~\textbf{with none of} (1) anchor-based localization; (2) specially designed box representations. Instead, \textbf{the critical aspect of improving slender object detection is feature adaptation}. It identifies and extends the insights of existing methods that are previously underexploited. Furthermore, we propose a feature adaption strategy that achieves clear and consistent improvements over current representative object detection methods. |
1012.5041 | Pablo S\'anchez-Moreno | P. S\'anchez-Moreno, A. Zarzo and J.S. Dehesa | Jensen divergence based on Fisher's information | 8 pages, 8 figures | J. Phys. A: Math. Theor. 45 (2012) 125305 | 10.1088/1751-8113/45/12/125305 | null | cs.IT math.IT physics.data-an | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The measure of Jensen-Fisher divergence between probability distributions is
introduced and its theoretical grounds set up. This quantity, in contrast to
the remaining Jensen divergences, is very sensitive to the fluctuations of the
probability distributions because it is controlled by the (local) Fisher
information, which is a gradient functional of the distribution. So, it is
appropriate and informative when studying the similarity of distributions,
mainly for those having oscillatory character. The new Jensen-Fisher divergence
shares with the Jensen-Shannon divergence the following properties:
non-negativity, additivity when applied to an arbitrary number of probability
densities, symmetry under exchange of these densities, vanishing if and only if
all the densities are equal, and definiteness even when these densities present
non-common zeros. Moreover, the Jensen-Fisher divergence is shown to be
expressed in terms of the relative Fisher information as the Jensen-Shannon
divergence does in terms of the Kullback-Leibler or relative Shannon entropy.
Finally the Jensen-Shannon and Jensen-Fisher divergences are compared for the
following three large, non-trivial and qualitatively different families of
probability distributions: the sinusoidal, generalized gamma-like and
Rakhmanov-Hermite distributions.
| [
{
"created": "Wed, 22 Dec 2010 17:15:17 GMT",
"version": "v1"
}
] | 2013-01-08 | [
[
"Sánchez-Moreno",
"P.",
""
],
[
"Zarzo",
"A.",
""
],
[
"Dehesa",
"J. S.",
""
]
] | The measure of Jensen-Fisher divergence between probability distributions is introduced and its theoretical grounds set up. This quantity, in contrast to the remaining Jensen divergences, is very sensitive to the fluctuations of the probability distributions because it is controlled by the (local) Fisher information, which is a gradient functional of the distribution. So, it is appropriate and informative when studying the similarity of distributions, mainly for those having oscillatory character. The new Jensen-Fisher divergence shares with the Jensen-Shannon divergence the following properties: non-negativity, additivity when applied to an arbitrary number of probability densities, symmetry under exchange of these densities, vanishing if and only if all the densities are equal, and definiteness even when these densities present non-common zeros. Moreover, the Jensen-Fisher divergence is shown to be expressed in terms of the relative Fisher information as the Jensen-Shannon divergence does in terms of the Kullback-Leibler or relative Shannon entropy. Finally the Jensen-Shannon and Jensen-Fisher divergences are compared for the following three large, non-trivial and qualitatively different families of probability distributions: the sinusoidal, generalized gamma-like and Rakhmanov-Hermite distributions. |
1505.07293 | Vijay Badrinarayanan | Vijay Badrinarayanan, Ankur Handa, Roberto Cipolla | SegNet: A Deep Convolutional Encoder-Decoder Architecture for Robust
Semantic Pixel-Wise Labelling | This version was first submitted to CVPR' 15 on November 14, 2014
with paper Id 1468. A similar architecture was proposed more recently on May
17, 2015, see http://arxiv.org/pdf/1505.04366.pdf | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose a novel deep architecture, SegNet, for semantic pixel wise image
labelling. SegNet has several attractive properties; (i) it only requires
forward evaluation of a fully learnt function to obtain smooth label
predictions, (ii) with increasing depth, a larger context is considered for
pixel labelling which improves accuracy, and (iii) it is easy to visualise the
effect of feature activation(s) in the pixel label space at any depth. SegNet
is composed of a stack of encoders followed by a corresponding decoder stack
which feeds into a soft-max classification layer. The decoders help map low
resolution feature maps at the output of the encoder stack to full input image
size feature maps. This addresses an important drawback of recent deep learning
approaches which have adopted networks designed for object categorization for
pixel wise labelling. These methods lack a mechanism to map deep layer feature
maps to input dimensions. They resort to ad hoc methods to upsample features,
e.g. by replication. This results in noisy predictions and also restricts the
number of pooling layers in order to avoid too much upsampling and thus reduces
spatial context. SegNet overcomes these problems by learning to map encoder
outputs to image pixel labels. We test the performance of SegNet on outdoor RGB
scenes from CamVid, KITTI and indoor scenes from the NYU dataset. Our results
show that SegNet achieves state-of-the-art performance even without use of
additional cues such as depth, video frames or post-processing with CRF models.
| [
{
"created": "Wed, 27 May 2015 12:54:17 GMT",
"version": "v1"
}
] | 2015-05-28 | [
[
"Badrinarayanan",
"Vijay",
""
],
[
"Handa",
"Ankur",
""
],
[
"Cipolla",
"Roberto",
""
]
] | We propose a novel deep architecture, SegNet, for semantic pixel wise image labelling. SegNet has several attractive properties; (i) it only requires forward evaluation of a fully learnt function to obtain smooth label predictions, (ii) with increasing depth, a larger context is considered for pixel labelling which improves accuracy, and (iii) it is easy to visualise the effect of feature activation(s) in the pixel label space at any depth. SegNet is composed of a stack of encoders followed by a corresponding decoder stack which feeds into a soft-max classification layer. The decoders help map low resolution feature maps at the output of the encoder stack to full input image size feature maps. This addresses an important drawback of recent deep learning approaches which have adopted networks designed for object categorization for pixel wise labelling. These methods lack a mechanism to map deep layer feature maps to input dimensions. They resort to ad hoc methods to upsample features, e.g. by replication. This results in noisy predictions and also restricts the number of pooling layers in order to avoid too much upsampling and thus reduces spatial context. SegNet overcomes these problems by learning to map encoder outputs to image pixel labels. We test the performance of SegNet on outdoor RGB scenes from CamVid, KITTI and indoor scenes from the NYU dataset. Our results show that SegNet achieves state-of-the-art performance even without use of additional cues such as depth, video frames or post-processing with CRF models. |
2308.14326 | Maximilian St\"abler | Maximilian Staebler, Frank Koester, Christoph Schlueter-Langdon | Towards solving ontological dissonance using network graphs | 5 pages, AMCIS 2023 proceedings | null | null | null | cs.AI cs.SI | http://creativecommons.org/licenses/by/4.0/ | Data Spaces are an emerging concept for the trusted implementation of
data-based applications and business models, offering a high degree of
flexibility and sovereignty to all stakeholders. As Data Spaces are currently
emerging in different domains such as mobility, health or food, semantic
interfaces need to be identified and implemented to ensure the technical
interoperability of these Data Spaces. This paper consolidates data models from
13 different domains and analyzes the ontological dissonance of these domains.
Using a network graph, central data models and ontology attributes are
identified, while the semantic heterogeneity of these domains is described
qualitatively. The research outlook describes how these results help to connect
different Data Spaces across domains.
| [
{
"created": "Mon, 28 Aug 2023 06:10:26 GMT",
"version": "v1"
}
] | 2023-08-29 | [
[
"Staebler",
"Maximilian",
""
],
[
"Koester",
"Frank",
""
],
[
"Schlueter-Langdon",
"Christoph",
""
]
] | Data Spaces are an emerging concept for the trusted implementation of data-based applications and business models, offering a high degree of flexibility and sovereignty to all stakeholders. As Data Spaces are currently emerging in different domains such as mobility, health or food, semantic interfaces need to be identified and implemented to ensure the technical interoperability of these Data Spaces. This paper consolidates data models from 13 different domains and analyzes the ontological dissonance of these domains. Using a network graph, central data models and ontology attributes are identified, while the semantic heterogeneity of these domains is described qualitatively. The research outlook describes how these results help to connect different Data Spaces across domains. |
2012.11334 | Viacheslav Dubeyko | Viacheslav Dubeyko | Cognitive Computing in Data-centric Paradigm | null | null | null | null | cs.AR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Knowledge is the most precious asset of humankind. People extract the
experience from the data that provide for us the reality through the feelings.
Generally speaking, it is possible to see the analogy of knowledge elaboration
between humankind's way and the artificial system's way. Digital data are the
"feelings" of an artificial system, and it needs to invent a method of
extraction of knowledge from the Universe of data.
The cognitive computing paradigm implies that a system should be able to
extract the knowledge from raw data without any human-made algorithm. The first
step of the paradigm is analysis of raw data streams through the discovery of
repeatable patterns of data. The knowledge of relationships among the patterns
provides a way to see the structures and to generalize the concepts with the
goal to synthesize new statements. The cognitive computing paradigm is capable
of mimicking the human's ability to generalize the notions. It is possible to
say that the generalization step provides the basis for discovering the
abstract notions, revealing the abstract relations of patterns and general
rules of structure synthesis.
If anyone continues the process of structure generalization, then it is
possible to build the multi-level hierarchy of abstract notions. Moreover,
discovering the generalized classes of notions is the first step towards a
paradigm of artificial analytical thinking. The most critical possible
responsibility of cognitive computing could be the classification of data and
recognition of input data stream's states. The synthesis of new statements
creates the foundation for the foreseeing the possible data states and
elaboration of knowledge about new data classes by employing synthesis and
checking the hypothesis.
| [
{
"created": "Mon, 14 Dec 2020 22:39:53 GMT",
"version": "v1"
}
] | 2020-12-22 | [
[
"Dubeyko",
"Viacheslav",
""
]
] | Knowledge is the most precious asset of humankind. People extract the experience from the data that provide for us the reality through the feelings. Generally speaking, it is possible to see the analogy of knowledge elaboration between humankind's way and the artificial system's way. Digital data are the "feelings" of an artificial system, and it needs to invent a method of extraction of knowledge from the Universe of data. The cognitive computing paradigm implies that a system should be able to extract the knowledge from raw data without any human-made algorithm. The first step of the paradigm is analysis of raw data streams through the discovery of repeatable patterns of data. The knowledge of relationships among the patterns provides a way to see the structures and to generalize the concepts with the goal to synthesize new statements. The cognitive computing paradigm is capable of mimicking the human's ability to generalize the notions. It is possible to say that the generalization step provides the basis for discovering the abstract notions, revealing the abstract relations of patterns and general rules of structure synthesis. If anyone continues the process of structure generalization, then it is possible to build the multi-level hierarchy of abstract notions. Moreover, discovering the generalized classes of notions is the first step towards a paradigm of artificial analytical thinking. The most critical possible responsibility of cognitive computing could be the classification of data and recognition of input data stream's states. The synthesis of new statements creates the foundation for the foreseeing the possible data states and elaboration of knowledge about new data classes by employing synthesis and checking the hypothesis. |
1603.02381 | Ragesh K Ramachandran | Ragesh K Ramachandran and Spring Berman | The Effect of Communication Topology on Scalar Field Estimation by
Networked Robotic Swarms | null | null | null | null | cs.RO cs.SY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper studies the problem of reconstructing a two-dimensional scalar
field using a swarm of networked robots with local communication capabilities.
We consider the communication network of the robots to form either a chain or a
grid topology. We formulate the reconstruction problem as an optimization
problem that is constrained by first-order linear dynamics on a large,
interconnected system. To solve this problem, we employ an optimization-based
scheme that uses a gradient-based method with an analytical computation of the
gradient. In addition, we derive bounds on the trace of observability Gramian
of the system, which helps us to quantify and compare the estimation capability
of chain and grid networks. A comparison based on a performance measure related
to the H2 norm of the system is also used to study robustness of the network
topologies. Our resultsare validated using both simulated scalar fields and
actual ocean salinity data.
| [
{
"created": "Tue, 8 Mar 2016 04:51:09 GMT",
"version": "v1"
}
] | 2016-03-09 | [
[
"Ramachandran",
"Ragesh K",
""
],
[
"Berman",
"Spring",
""
]
] | This paper studies the problem of reconstructing a two-dimensional scalar field using a swarm of networked robots with local communication capabilities. We consider the communication network of the robots to form either a chain or a grid topology. We formulate the reconstruction problem as an optimization problem that is constrained by first-order linear dynamics on a large, interconnected system. To solve this problem, we employ an optimization-based scheme that uses a gradient-based method with an analytical computation of the gradient. In addition, we derive bounds on the trace of observability Gramian of the system, which helps us to quantify and compare the estimation capability of chain and grid networks. A comparison based on a performance measure related to the H2 norm of the system is also used to study robustness of the network topologies. Our resultsare validated using both simulated scalar fields and actual ocean salinity data. |
2304.00173 | Rami Botros | Rami Botros, Rohit Prabhavalkar, Johan Schalkwyk, Ciprian Chelba, Tara
N. Sainath, Fran\c{c}oise Beaufays | Lego-Features: Exporting modular encoder features for streaming and
deliberation ASR | null | null | null | null | cs.CL cs.AI cs.SD eess.AS | http://creativecommons.org/licenses/by/4.0/ | In end-to-end (E2E) speech recognition models, a representational
tight-coupling inevitably emerges between the encoder and the decoder. We build
upon recent work that has begun to explore building encoders with modular
encoded representations, such that encoders and decoders from different models
can be stitched together in a zero-shot manner without further fine-tuning.
While previous research only addresses full-context speech models, we explore
the problem in a streaming setting as well. Our framework builds on top of
existing encoded representations, converting them to modular features, dubbed
as Lego-Features, without modifying the pre-trained model. The features remain
interchangeable when the model is retrained with distinct initializations.
Though sparse, we show that the Lego-Features are powerful when tested with
RNN-T or LAS decoders, maintaining high-quality downstream performance. They
are also rich enough to represent the first-pass prediction during two-pass
deliberation. In this scenario, they outperform the N-best hypotheses, since
they do not need to be supplemented with acoustic features to deliver the best
results. Moreover, generating the Lego-Features does not require beam search or
auto-regressive computation. Overall, they present a modular, powerful and
cheap alternative to the standard encoder output, as well as the N-best
hypotheses.
| [
{
"created": "Fri, 31 Mar 2023 23:33:21 GMT",
"version": "v1"
}
] | 2023-04-04 | [
[
"Botros",
"Rami",
""
],
[
"Prabhavalkar",
"Rohit",
""
],
[
"Schalkwyk",
"Johan",
""
],
[
"Chelba",
"Ciprian",
""
],
[
"Sainath",
"Tara N.",
""
],
[
"Beaufays",
"Françoise",
""
]
] | In end-to-end (E2E) speech recognition models, a representational tight-coupling inevitably emerges between the encoder and the decoder. We build upon recent work that has begun to explore building encoders with modular encoded representations, such that encoders and decoders from different models can be stitched together in a zero-shot manner without further fine-tuning. While previous research only addresses full-context speech models, we explore the problem in a streaming setting as well. Our framework builds on top of existing encoded representations, converting them to modular features, dubbed as Lego-Features, without modifying the pre-trained model. The features remain interchangeable when the model is retrained with distinct initializations. Though sparse, we show that the Lego-Features are powerful when tested with RNN-T or LAS decoders, maintaining high-quality downstream performance. They are also rich enough to represent the first-pass prediction during two-pass deliberation. In this scenario, they outperform the N-best hypotheses, since they do not need to be supplemented with acoustic features to deliver the best results. Moreover, generating the Lego-Features does not require beam search or auto-regressive computation. Overall, they present a modular, powerful and cheap alternative to the standard encoder output, as well as the N-best hypotheses. |
1809.01906 | Felix Leibfried | Felix Leibfried, Peter Vrancx | Model-Based Regularization for Deep Reinforcement Learning with
Transcoder Networks | Presented at the NIPS Deep Reinforcement Learning Workshop, Montreal,
Canada, 2018 | null | null | null | cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper proposes a new optimization objective for value-based deep
reinforcement learning. We extend conventional Deep Q-Networks (DQNs) by adding
a model-learning component yielding a transcoder network. The prediction errors
for the model are included in the basic DQN loss as additional regularizers.
This augmented objective leads to a richer training signal that provides
feedback at every time step. Moreover, because learning an environment model
shares a common structure with the RL problem, we hypothesize that the
resulting objective improves both sample efficiency and performance. We
empirically confirm our hypothesis on a range of 20 games from the Atari
benchmark attaining superior results over vanilla DQN without model-based
regularization.
| [
{
"created": "Thu, 6 Sep 2018 09:49:18 GMT",
"version": "v1"
},
{
"created": "Tue, 20 Nov 2018 13:30:16 GMT",
"version": "v2"
}
] | 2018-11-21 | [
[
"Leibfried",
"Felix",
""
],
[
"Vrancx",
"Peter",
""
]
] | This paper proposes a new optimization objective for value-based deep reinforcement learning. We extend conventional Deep Q-Networks (DQNs) by adding a model-learning component yielding a transcoder network. The prediction errors for the model are included in the basic DQN loss as additional regularizers. This augmented objective leads to a richer training signal that provides feedback at every time step. Moreover, because learning an environment model shares a common structure with the RL problem, we hypothesize that the resulting objective improves both sample efficiency and performance. We empirically confirm our hypothesis on a range of 20 games from the Atari benchmark attaining superior results over vanilla DQN without model-based regularization. |
2202.13677 | Sean Kauffman | Sean Kauffman, Martin Zimmermann | The Complexity of Evaluating nfer | null | null | null | null | cs.LO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Nfer is a rule-based language for abstracting event streams into a hierarchy
of intervals with data. Nfer has multiple implementations and has been applied
in the analysis of spacecraft telemetry and autonomous vehicle logs. This work
provides the first complexity analysis of nfer evaluation, i.e., the problem of
deciding whether a given interval is generated by applying rules.
We show that the full nfer language is undecidable and that this depends on
both recursion in the rules and an infinite data domain. By restricting either
or both of those capabilities, we obtain tight decidability results. We also
examine the impact on complexity of exclusive rules and minimality. For the
most practical case, which is minimality with finite data, we provide a
polynomial-time algorithm.
| [
{
"created": "Mon, 28 Feb 2022 10:53:09 GMT",
"version": "v1"
},
{
"created": "Fri, 1 Jul 2022 12:37:22 GMT",
"version": "v2"
},
{
"created": "Mon, 21 Nov 2022 12:08:18 GMT",
"version": "v3"
}
] | 2022-11-22 | [
[
"Kauffman",
"Sean",
""
],
[
"Zimmermann",
"Martin",
""
]
] | Nfer is a rule-based language for abstracting event streams into a hierarchy of intervals with data. Nfer has multiple implementations and has been applied in the analysis of spacecraft telemetry and autonomous vehicle logs. This work provides the first complexity analysis of nfer evaluation, i.e., the problem of deciding whether a given interval is generated by applying rules. We show that the full nfer language is undecidable and that this depends on both recursion in the rules and an infinite data domain. By restricting either or both of those capabilities, we obtain tight decidability results. We also examine the impact on complexity of exclusive rules and minimality. For the most practical case, which is minimality with finite data, we provide a polynomial-time algorithm. |
2209.10767 | Srikanth Malla | Srikanth Malla, Chiho Choi, Isht Dwivedi, Joon Hee Choi, Jiachen Li | DRAMA: Joint Risk Localization and Captioning in Driving | WACV 2023 (Winter Conference on Applications of Computer Vision) | null | null | null | cs.CV cs.AI cs.LG cs.RO | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Considering the functionality of situational awareness in safety-critical
automation systems, the perception of risk in driving scenes and its
explainability is of particular importance for autonomous and cooperative
driving. Toward this goal, this paper proposes a new research direction of
joint risk localization in driving scenes and its risk explanation as a natural
language description. Due to the lack of standard benchmarks, we collected a
large-scale dataset, DRAMA (Driving Risk Assessment Mechanism with A captioning
module), which consists of 17,785 interactive driving scenarios collected in
Tokyo, Japan. Our DRAMA dataset accommodates video- and object-level questions
on driving risks with associated important objects to achieve the goal of
visual captioning as a free-form language description utilizing closed and
open-ended responses for multi-level questions, which can be used to evaluate a
range of visual captioning capabilities in driving scenarios. We make this data
available to the community for further research. Using DRAMA, we explore
multiple facets of joint risk localization and captioning in interactive
driving scenarios. In particular, we benchmark various multi-task prediction
architectures and provide a detailed analysis of joint risk localization and
risk captioning. The data set is available at https://usa.honda-ri.com/drama
| [
{
"created": "Thu, 22 Sep 2022 03:53:56 GMT",
"version": "v1"
},
{
"created": "Wed, 5 Oct 2022 21:09:10 GMT",
"version": "v2"
}
] | 2022-10-07 | [
[
"Malla",
"Srikanth",
""
],
[
"Choi",
"Chiho",
""
],
[
"Dwivedi",
"Isht",
""
],
[
"Choi",
"Joon Hee",
""
],
[
"Li",
"Jiachen",
""
]
] | Considering the functionality of situational awareness in safety-critical automation systems, the perception of risk in driving scenes and its explainability is of particular importance for autonomous and cooperative driving. Toward this goal, this paper proposes a new research direction of joint risk localization in driving scenes and its risk explanation as a natural language description. Due to the lack of standard benchmarks, we collected a large-scale dataset, DRAMA (Driving Risk Assessment Mechanism with A captioning module), which consists of 17,785 interactive driving scenarios collected in Tokyo, Japan. Our DRAMA dataset accommodates video- and object-level questions on driving risks with associated important objects to achieve the goal of visual captioning as a free-form language description utilizing closed and open-ended responses for multi-level questions, which can be used to evaluate a range of visual captioning capabilities in driving scenarios. We make this data available to the community for further research. Using DRAMA, we explore multiple facets of joint risk localization and captioning in interactive driving scenarios. In particular, we benchmark various multi-task prediction architectures and provide a detailed analysis of joint risk localization and risk captioning. The data set is available at https://usa.honda-ri.com/drama |
2211.06223 | Linqi Ye Dr. | Linqi Ye, Xueqian Wang, Houde Liu, Bin Liang | The Simplest Balance Controller for Dynamic Walking | null | null | null | null | cs.RO cs.SY eess.SY | http://creativecommons.org/licenses/by/4.0/ | Humans can balance very well during walking, even when perturbed. But it
seems difficult to achieve robust walking for bipedal robots. Here we describe
the simplest balance controller that leads to robust walking for a linear
inverted pendulum (LIP) model. The main idea is to use a linear function of the
body velocity to determine the next foot placement, which we call linear foot
placement control (LFPC). By using the Poincar\'e map, a balance criterion is
derived, which shows that LFPC is stable when the velocity-feedback coefficient
is located in a certain range. And that range is much bigger when stepping
faster, which indicates "faster stepping, easier to balance". We show that
various gaits can be generated by adjusting the controller parameters in LFPC.
Particularly, a dead-beat controller is discovered that can lead to
steady-state walking in just one step. The effectiveness of LFPC is verified
through Matlab simulation as well as V-REP simulation for both 2D and 3D
walking. The main feature of LFPC is its simplicity and inherent robustness,
which may help us understand the essence of how to maintain balance in dynamic
walking.
| [
{
"created": "Fri, 11 Nov 2022 14:19:40 GMT",
"version": "v1"
}
] | 2022-11-14 | [
[
"Ye",
"Linqi",
""
],
[
"Wang",
"Xueqian",
""
],
[
"Liu",
"Houde",
""
],
[
"Liang",
"Bin",
""
]
] | Humans can balance very well during walking, even when perturbed. But it seems difficult to achieve robust walking for bipedal robots. Here we describe the simplest balance controller that leads to robust walking for a linear inverted pendulum (LIP) model. The main idea is to use a linear function of the body velocity to determine the next foot placement, which we call linear foot placement control (LFPC). By using the Poincar\'e map, a balance criterion is derived, which shows that LFPC is stable when the velocity-feedback coefficient is located in a certain range. And that range is much bigger when stepping faster, which indicates "faster stepping, easier to balance". We show that various gaits can be generated by adjusting the controller parameters in LFPC. Particularly, a dead-beat controller is discovered that can lead to steady-state walking in just one step. The effectiveness of LFPC is verified through Matlab simulation as well as V-REP simulation for both 2D and 3D walking. The main feature of LFPC is its simplicity and inherent robustness, which may help us understand the essence of how to maintain balance in dynamic walking. |
2309.16783 | David Widemann | Lakshmi Nair, David Widemann, Brad Turcott, Nick Moore, Alexandra
Wleklinski, Darius Bunandar, Ioannis Papavasileiou, Shihu Wang, Eric Logan | Photonic Accelerators for Image Segmentation in Autonomous Driving and
Defect Detection | null | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by/4.0/ | Photonic computing promises faster and more energy-efficient deep neural
network (DNN) inference than traditional digital hardware. Advances in photonic
computing can have profound impacts on applications such as autonomous driving
and defect detection that depend on fast, accurate and energy efficient
execution of image segmentation models. In this paper, we investigate image
segmentation on photonic accelerators to explore: a) the types of image
segmentation DNN architectures that are best suited for photonic accelerators,
and b) the throughput and energy efficiency of executing the different image
segmentation models on photonic accelerators, along with the trade-offs
involved therein. Specifically, we demonstrate that certain segmentation models
exhibit negligible loss in accuracy (compared to digital float32 models) when
executed on photonic accelerators, and explore the empirical reasoning for
their robustness. We also discuss techniques for recovering accuracy in the
case of models that do not perform well. Further, we compare throughput
(inferences-per-second) and energy consumption estimates for different image
segmentation workloads on photonic accelerators. We discuss the challenges and
potential optimizations that can help improve the application of photonic
accelerators to such computer vision tasks.
| [
{
"created": "Thu, 28 Sep 2023 18:22:41 GMT",
"version": "v1"
},
{
"created": "Tue, 3 Oct 2023 16:34:13 GMT",
"version": "v2"
}
] | 2023-10-04 | [
[
"Nair",
"Lakshmi",
""
],
[
"Widemann",
"David",
""
],
[
"Turcott",
"Brad",
""
],
[
"Moore",
"Nick",
""
],
[
"Wleklinski",
"Alexandra",
""
],
[
"Bunandar",
"Darius",
""
],
[
"Papavasileiou",
"Ioannis",
""
],
[
"Wang",
"Shihu",
""
],
[
"Logan",
"Eric",
""
]
] | Photonic computing promises faster and more energy-efficient deep neural network (DNN) inference than traditional digital hardware. Advances in photonic computing can have profound impacts on applications such as autonomous driving and defect detection that depend on fast, accurate and energy efficient execution of image segmentation models. In this paper, we investigate image segmentation on photonic accelerators to explore: a) the types of image segmentation DNN architectures that are best suited for photonic accelerators, and b) the throughput and energy efficiency of executing the different image segmentation models on photonic accelerators, along with the trade-offs involved therein. Specifically, we demonstrate that certain segmentation models exhibit negligible loss in accuracy (compared to digital float32 models) when executed on photonic accelerators, and explore the empirical reasoning for their robustness. We also discuss techniques for recovering accuracy in the case of models that do not perform well. Further, we compare throughput (inferences-per-second) and energy consumption estimates for different image segmentation workloads on photonic accelerators. We discuss the challenges and potential optimizations that can help improve the application of photonic accelerators to such computer vision tasks. |
2102.08085 | Fouzia Altaf Ms | Fouzia Altaf, Syed M.S. Islam, Naeem K. Janjua, Naveed Akhtar | Boosting Deep Transfer Learning for COVID-19 Classification | 5 pages | null | null | null | cs.CV cs.AI cs.LG | http://creativecommons.org/licenses/by-nc-nd/4.0/ | COVID-19 classification using chest Computed Tomography (CT) has been found
pragmatically useful by several studies. Due to the lack of annotated samples,
these studies recommend transfer learning and explore the choices of
pre-trained models and data augmentation. However, it is still unknown if there
are better strategies than vanilla transfer learning for more accurate COVID-19
classification with limited CT data. This paper provides an affirmative answer,
devising a novel `model' augmentation technique that allows a considerable
performance boost to transfer learning for the task. Our method systematically
reduces the distributional shift between the source and target domains and
considers augmenting deep learning with complementary representation learning
techniques. We establish the efficacy of our method with publicly available
datasets and models, along with identifying contrasting observations in the
previous studies.
| [
{
"created": "Tue, 16 Feb 2021 11:15:23 GMT",
"version": "v1"
}
] | 2021-02-17 | [
[
"Altaf",
"Fouzia",
""
],
[
"Islam",
"Syed M. S.",
""
],
[
"Janjua",
"Naeem K.",
""
],
[
"Akhtar",
"Naveed",
""
]
] | COVID-19 classification using chest Computed Tomography (CT) has been found pragmatically useful by several studies. Due to the lack of annotated samples, these studies recommend transfer learning and explore the choices of pre-trained models and data augmentation. However, it is still unknown if there are better strategies than vanilla transfer learning for more accurate COVID-19 classification with limited CT data. This paper provides an affirmative answer, devising a novel `model' augmentation technique that allows a considerable performance boost to transfer learning for the task. Our method systematically reduces the distributional shift between the source and target domains and considers augmenting deep learning with complementary representation learning techniques. We establish the efficacy of our method with publicly available datasets and models, along with identifying contrasting observations in the previous studies. |
2011.01671 | Christian Berger | Christian Berger, Hans P. Reiser, Jo\~ao Sousa, Alysson Bessani | AWARE: Adaptive Wide-Area Replication for Fast and Resilient Byzantine
Consensus | This paper consists of 16 pages in total. This paper is the accepted
version to be published in IEEE Transactions on Dependable and Secure
Computing (2020). For the published version refer to DOI
https://doi.org/10.1109/TDSC.2020.3030605 | null | 10.1109/TDSC.2020.3030605 | null | cs.DC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | With upcoming blockchain infrastructures, world-spanning Byzantine consensus
is getting practical and necessary. In geographically distributed systems, the
pace at which consensus is achieved is limited by the heterogenous latencies of
connections between replicas. If deployed on a wide-area network,
consensus-based systems benefit from weighted replication, an approach that
utilizes extra replicas and assigns higher voting power to well connected
replicas. This enables more choice in quorum formation and replicas can
leverage proportionally smaller quorums to advance, thus decreasing consensus
latency. However, the system needs a solution to autonomously adjust to its
environment if network conditions change or faults occur. We present Adaptive
Wide-Area REplication (AWARE), a mechanism which improves the geographical
scalability of consensus with nodes being widely spread across the world.
Essentially, AWARE is an automated and dynamic voting weight tuning and leader
positioning scheme, which supports the emergence of fast quorums in the system.
It employs a reliable self-monitoring process and provides a prediction model
seeking to minimize the system's consensus latency. In experiments using
several AWS EC2 regions, AWARE dynamically optimizes consensus latency by
self-reliantly finding a fast weight configuration yielding latency gains
observed by clients located across the globe.
| [
{
"created": "Tue, 3 Nov 2020 12:58:39 GMT",
"version": "v1"
}
] | 2020-11-04 | [
[
"Berger",
"Christian",
""
],
[
"Reiser",
"Hans P.",
""
],
[
"Sousa",
"João",
""
],
[
"Bessani",
"Alysson",
""
]
] | With upcoming blockchain infrastructures, world-spanning Byzantine consensus is getting practical and necessary. In geographically distributed systems, the pace at which consensus is achieved is limited by the heterogenous latencies of connections between replicas. If deployed on a wide-area network, consensus-based systems benefit from weighted replication, an approach that utilizes extra replicas and assigns higher voting power to well connected replicas. This enables more choice in quorum formation and replicas can leverage proportionally smaller quorums to advance, thus decreasing consensus latency. However, the system needs a solution to autonomously adjust to its environment if network conditions change or faults occur. We present Adaptive Wide-Area REplication (AWARE), a mechanism which improves the geographical scalability of consensus with nodes being widely spread across the world. Essentially, AWARE is an automated and dynamic voting weight tuning and leader positioning scheme, which supports the emergence of fast quorums in the system. It employs a reliable self-monitoring process and provides a prediction model seeking to minimize the system's consensus latency. In experiments using several AWS EC2 regions, AWARE dynamically optimizes consensus latency by self-reliantly finding a fast weight configuration yielding latency gains observed by clients located across the globe. |
2403.18133 | Erkan Karabulut | Erkan Karabulut, Victoria Degeler, Paul Groth | AE SemRL: Learning Semantic Association Rules with Autoencoders | null | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | Association Rule Mining (ARM) is the task of learning associations among data
features in the form of logical rules. Mining association rules from
high-dimensional numerical data, for example, time series data from a large
number of sensors in a smart environment, is a computationally intensive task.
In this study, we propose an Autoencoder-based approach to learn and extract
association rules from time series data (AE SemRL). Moreover, we argue that in
the presence of semantic information related to time series data sources,
semantics can facilitate learning generalizable and explainable association
rules. Despite enriching time series data with additional semantic features, AE
SemRL makes learning association rules from high-dimensional data feasible. Our
experiments show that semantic association rules can be extracted from a latent
representation created by an Autoencoder and this method has in the order of
hundreds of times faster execution time than state-of-the-art ARM approaches in
many scenarios. We believe that this study advances a new way of extracting
associations from representations and has the potential to inspire more
research in this field.
| [
{
"created": "Tue, 26 Mar 2024 22:28:43 GMT",
"version": "v1"
}
] | 2024-03-28 | [
[
"Karabulut",
"Erkan",
""
],
[
"Degeler",
"Victoria",
""
],
[
"Groth",
"Paul",
""
]
] | Association Rule Mining (ARM) is the task of learning associations among data features in the form of logical rules. Mining association rules from high-dimensional numerical data, for example, time series data from a large number of sensors in a smart environment, is a computationally intensive task. In this study, we propose an Autoencoder-based approach to learn and extract association rules from time series data (AE SemRL). Moreover, we argue that in the presence of semantic information related to time series data sources, semantics can facilitate learning generalizable and explainable association rules. Despite enriching time series data with additional semantic features, AE SemRL makes learning association rules from high-dimensional data feasible. Our experiments show that semantic association rules can be extracted from a latent representation created by an Autoencoder and this method has in the order of hundreds of times faster execution time than state-of-the-art ARM approaches in many scenarios. We believe that this study advances a new way of extracting associations from representations and has the potential to inspire more research in this field. |
1704.03928 | Noah Stephens-Davidowitz | Huck Bennett, Alexander Golovnev, Noah Stephens-Davidowitz | On the Quantitative Hardness of CVP | null | FOCS 2017 | null | null | cs.CC cs.DS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | $ \newcommand{\eps}{\varepsilon}
\newcommand{\problem}[1]{\ensuremath{\mathrm{#1}} }
\newcommand{\CVP}{\problem{CVP}} \newcommand{\SVP}{\problem{SVP}}
\newcommand{\CVPP}{\problem{CVPP}} \newcommand{\ensuremath}[1]{#1} $For odd
integers $p \geq 1$ (and $p = \infty$), we show that the Closest Vector Problem
in the $\ell_p$ norm ($\CVP_p$) over rank $n$ lattices cannot be solved in
$2^{(1-\eps) n}$ time for any constant $\eps > 0$ unless the Strong Exponential
Time Hypothesis (SETH) fails. We then extend this result to "almost all" values
of $p \geq 1$, not including the even integers. This comes tantalizingly close
to settling the quantitative time complexity of the important special case of
$\CVP_2$ (i.e., $\CVP$ in the Euclidean norm), for which a $2^{n +o(n)}$-time
algorithm is known. In particular, our result applies for any $p = p(n) \neq 2$
that approaches $2$ as $n \to \infty$.
We also show a similar SETH-hardness result for $\SVP_\infty$; hardness of
approximating $\CVP_p$ to within some constant factor under the so-called
Gap-ETH assumption; and other quantitative hardness results for $\CVP_p$ and
$\CVPP_p$ for any $1 \leq p < \infty$ under different assumptions.
| [
{
"created": "Wed, 12 Apr 2017 20:55:59 GMT",
"version": "v1"
},
{
"created": "Thu, 5 Oct 2017 19:05:01 GMT",
"version": "v2"
}
] | 2019-01-28 | [
[
"Bennett",
"Huck",
""
],
[
"Golovnev",
"Alexander",
""
],
[
"Stephens-Davidowitz",
"Noah",
""
]
] | $ \newcommand{\eps}{\varepsilon} \newcommand{\problem}[1]{\ensuremath{\mathrm{#1}} } \newcommand{\CVP}{\problem{CVP}} \newcommand{\SVP}{\problem{SVP}} \newcommand{\CVPP}{\problem{CVPP}} \newcommand{\ensuremath}[1]{#1} $For odd integers $p \geq 1$ (and $p = \infty$), we show that the Closest Vector Problem in the $\ell_p$ norm ($\CVP_p$) over rank $n$ lattices cannot be solved in $2^{(1-\eps) n}$ time for any constant $\eps > 0$ unless the Strong Exponential Time Hypothesis (SETH) fails. We then extend this result to "almost all" values of $p \geq 1$, not including the even integers. This comes tantalizingly close to settling the quantitative time complexity of the important special case of $\CVP_2$ (i.e., $\CVP$ in the Euclidean norm), for which a $2^{n +o(n)}$-time algorithm is known. In particular, our result applies for any $p = p(n) \neq 2$ that approaches $2$ as $n \to \infty$. We also show a similar SETH-hardness result for $\SVP_\infty$; hardness of approximating $\CVP_p$ to within some constant factor under the so-called Gap-ETH assumption; and other quantitative hardness results for $\CVP_p$ and $\CVPP_p$ for any $1 \leq p < \infty$ under different assumptions. |
2310.00922 | Hong Huy Nguyen | Huy H. Nguyen, Junichi Yamagishi, Isao Echizen | How Close are Other Computer Vision Tasks to Deepfake Detection? | Accepted to be Published in Proceedings of the IEEE International
Joint Conference on Biometrics (IJCB 2023) | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we challenge the conventional belief that supervised
ImageNet-trained models have strong generalizability and are suitable for use
as feature extractors in deepfake detection. We present a new measurement,
"model separability," for visually and quantitatively assessing a model's raw
capacity to separate data in an unsupervised manner. We also present a
systematic benchmark for determining the correlation between deepfake detection
and other computer vision tasks using pre-trained models. Our analysis shows
that pre-trained face recognition models are more closely related to deepfake
detection than other models. Additionally, models trained using self-supervised
methods are more effective in separation than those trained using supervised
methods. After fine-tuning all models on a small deepfake dataset, we found
that self-supervised models deliver the best results, but there is a risk of
overfitting. Our results provide valuable insights that should help researchers
and practitioners develop more effective deepfake detection models.
| [
{
"created": "Mon, 2 Oct 2023 06:32:35 GMT",
"version": "v1"
}
] | 2023-10-03 | [
[
"Nguyen",
"Huy H.",
""
],
[
"Yamagishi",
"Junichi",
""
],
[
"Echizen",
"Isao",
""
]
] | In this paper, we challenge the conventional belief that supervised ImageNet-trained models have strong generalizability and are suitable for use as feature extractors in deepfake detection. We present a new measurement, "model separability," for visually and quantitatively assessing a model's raw capacity to separate data in an unsupervised manner. We also present a systematic benchmark for determining the correlation between deepfake detection and other computer vision tasks using pre-trained models. Our analysis shows that pre-trained face recognition models are more closely related to deepfake detection than other models. Additionally, models trained using self-supervised methods are more effective in separation than those trained using supervised methods. After fine-tuning all models on a small deepfake dataset, we found that self-supervised models deliver the best results, but there is a risk of overfitting. Our results provide valuable insights that should help researchers and practitioners develop more effective deepfake detection models. |
2107.07983 | Zhi-Gang Liu | Zhi-Gang Liu, Paul N. Whatmough, Yuhao Zhu, Matthew Mattina | S2TA: Exploiting Structured Sparsity for Energy-Efficient Mobile CNN
Acceleration | Accepted by the HPCA 20222, the 28th IEEE International Symposium on
High-Performance Computer Architecture (HPCA-28) | null | null | null | cs.AR cs.LG | http://creativecommons.org/licenses/by/4.0/ | Exploiting sparsity is a key technique in accelerating quantized
convolutional neural network (CNN) inference on mobile devices. Prior sparse
CNN accelerators largely exploit un-structured sparsity and achieve significant
speedups. Due to the unbounded, largely unpredictable sparsity patterns,
however, exploiting unstructured sparsity requires complicated hardware design
with significant energy and area overhead, which is particularly detrimental to
mobile/IoT inference scenarios where energy and area efficiency are crucial. We
propose to exploit structured sparsity, more specifically, Density Bound Block
(DBB) sparsity for both weights and activations. DBB block tensors bound the
maximum number of non-zeros per block. DBB thus exposes statically predictable
sparsity patterns that enable lean sparsity-exploiting hardware. We propose new
hardware primitives to implement DBB sparsity for (static) weights and
(dynamic) activations, respectively, with very low overheads. Building on top
of the primitives, we describe S2TA, a systolic array-based CNN accelerator
that exploits joint weight and activation DBB sparsity and new dimensions of
data reuse unavailable on the traditional systolic array. S2TA in 16nm achieves
more than 2x speedup and energy reduction compared to a strong baseline of a
systolic array with zero-value clock gating, over five popular CNN benchmarks.
Compared to two recent non-systolic sparse accelerators, Eyeriss v2 (65nm) and
SparTen (45nm), S2TA in 65nm uses about 2.2x and 3.1x less energy per
inference, respectively.
| [
{
"created": "Fri, 16 Jul 2021 15:57:06 GMT",
"version": "v1"
},
{
"created": "Thu, 6 Jan 2022 16:23:55 GMT",
"version": "v2"
}
] | 2022-01-07 | [
[
"Liu",
"Zhi-Gang",
""
],
[
"Whatmough",
"Paul N.",
""
],
[
"Zhu",
"Yuhao",
""
],
[
"Mattina",
"Matthew",
""
]
] | Exploiting sparsity is a key technique in accelerating quantized convolutional neural network (CNN) inference on mobile devices. Prior sparse CNN accelerators largely exploit un-structured sparsity and achieve significant speedups. Due to the unbounded, largely unpredictable sparsity patterns, however, exploiting unstructured sparsity requires complicated hardware design with significant energy and area overhead, which is particularly detrimental to mobile/IoT inference scenarios where energy and area efficiency are crucial. We propose to exploit structured sparsity, more specifically, Density Bound Block (DBB) sparsity for both weights and activations. DBB block tensors bound the maximum number of non-zeros per block. DBB thus exposes statically predictable sparsity patterns that enable lean sparsity-exploiting hardware. We propose new hardware primitives to implement DBB sparsity for (static) weights and (dynamic) activations, respectively, with very low overheads. Building on top of the primitives, we describe S2TA, a systolic array-based CNN accelerator that exploits joint weight and activation DBB sparsity and new dimensions of data reuse unavailable on the traditional systolic array. S2TA in 16nm achieves more than 2x speedup and energy reduction compared to a strong baseline of a systolic array with zero-value clock gating, over five popular CNN benchmarks. Compared to two recent non-systolic sparse accelerators, Eyeriss v2 (65nm) and SparTen (45nm), S2TA in 65nm uses about 2.2x and 3.1x less energy per inference, respectively. |
1804.06682 | Mostafa Wahby | Mostafa Wahby, Mary Katherine Heinrich, Daniel Nicolas Hofstadler,
Payam Zahadat, Sebastian Risi, Phil Ayres, Thomas Schmickl and Heiko Hamann | A Robot to Shape your Natural Plant: The Machine Learning Approach to
Model and Control Bio-Hybrid Systems | null | null | 10.1145/3205455.3205516 | null | cs.NE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Bio-hybrid systems---close couplings of natural organisms with
technology---are high potential and still underexplored. In existing work,
robots have mostly influenced group behaviors of animals. We explore the
possibilities of mixing robots with natural plants, merging useful attributes.
Significant synergies arise by combining the plants' ability to efficiently
produce shaped material and the robots' ability to extend sensing and
decision-making behaviors. However, programming robots to control plant motion
and shape requires good knowledge of complex plant behaviors. Therefore, we use
machine learning to create a holistic plant model and evolve robot controllers.
As a benchmark task we choose obstacle avoidance. We use computer vision to
construct a model of plant stem stiffening and motion dynamics by training an
LSTM network. The LSTM network acts as a forward model predicting change in the
plant, driving the evolution of neural network robot controllers. The evolved
controllers augment the plants' natural light-finding and tissue-stiffening
behaviors to avoid obstacles and grow desired shapes. We successfully verify
the robot controllers and bio-hybrid behavior in reality, with a physical setup
and actual plants.
| [
{
"created": "Wed, 18 Apr 2018 12:30:18 GMT",
"version": "v1"
},
{
"created": "Thu, 19 Apr 2018 09:26:34 GMT",
"version": "v2"
}
] | 2018-04-20 | [
[
"Wahby",
"Mostafa",
""
],
[
"Heinrich",
"Mary Katherine",
""
],
[
"Hofstadler",
"Daniel Nicolas",
""
],
[
"Zahadat",
"Payam",
""
],
[
"Risi",
"Sebastian",
""
],
[
"Ayres",
"Phil",
""
],
[
"Schmickl",
"Thomas",
""
],
[
"Hamann",
"Heiko",
""
]
] | Bio-hybrid systems---close couplings of natural organisms with technology---are high potential and still underexplored. In existing work, robots have mostly influenced group behaviors of animals. We explore the possibilities of mixing robots with natural plants, merging useful attributes. Significant synergies arise by combining the plants' ability to efficiently produce shaped material and the robots' ability to extend sensing and decision-making behaviors. However, programming robots to control plant motion and shape requires good knowledge of complex plant behaviors. Therefore, we use machine learning to create a holistic plant model and evolve robot controllers. As a benchmark task we choose obstacle avoidance. We use computer vision to construct a model of plant stem stiffening and motion dynamics by training an LSTM network. The LSTM network acts as a forward model predicting change in the plant, driving the evolution of neural network robot controllers. The evolved controllers augment the plants' natural light-finding and tissue-stiffening behaviors to avoid obstacles and grow desired shapes. We successfully verify the robot controllers and bio-hybrid behavior in reality, with a physical setup and actual plants. |
2008.10715 | Binghui Wang | Binghui Wang, Jinyuan Jia, Xiaoyu Cao, Neil Zhenqiang Gong | Certified Robustness of Graph Neural Networks against Adversarial
Structural Perturbation | Accepted by ACM SIGKDD'21 | null | null | null | cs.CR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Graph neural networks (GNNs) have recently gained much attention for node and
graph classification tasks on graph-structured data. However, multiple recent
works showed that an attacker can easily make GNNs predict incorrectly via
perturbing the graph structure, i.e., adding or deleting edges in the graph. We
aim to defend against such attacks via developing certifiably robust GNNs.
Specifically, we prove the certified robustness guarantee of any GNN for both
node and graph classifications against structural perturbation. Moreover, we
show that our certified robustness guarantee is tight. Our results are based on
a recently proposed technique called randomized smoothing, which we extend to
graph data. We also empirically evaluate our method for both node and graph
classifications on multiple GNNs and multiple benchmark datasets. For instance,
on the Cora dataset, Graph Convolutional Network with our randomized smoothing
can achieve a certified accuracy of 0.49 when the attacker can arbitrarily
add/delete at most 15 edges in the graph.
| [
{
"created": "Mon, 24 Aug 2020 21:39:10 GMT",
"version": "v1"
},
{
"created": "Fri, 4 Jun 2021 02:34:29 GMT",
"version": "v2"
},
{
"created": "Fri, 16 Jul 2021 01:54:43 GMT",
"version": "v3"
}
] | 2021-07-19 | [
[
"Wang",
"Binghui",
""
],
[
"Jia",
"Jinyuan",
""
],
[
"Cao",
"Xiaoyu",
""
],
[
"Gong",
"Neil Zhenqiang",
""
]
] | Graph neural networks (GNNs) have recently gained much attention for node and graph classification tasks on graph-structured data. However, multiple recent works showed that an attacker can easily make GNNs predict incorrectly via perturbing the graph structure, i.e., adding or deleting edges in the graph. We aim to defend against such attacks via developing certifiably robust GNNs. Specifically, we prove the certified robustness guarantee of any GNN for both node and graph classifications against structural perturbation. Moreover, we show that our certified robustness guarantee is tight. Our results are based on a recently proposed technique called randomized smoothing, which we extend to graph data. We also empirically evaluate our method for both node and graph classifications on multiple GNNs and multiple benchmark datasets. For instance, on the Cora dataset, Graph Convolutional Network with our randomized smoothing can achieve a certified accuracy of 0.49 when the attacker can arbitrarily add/delete at most 15 edges in the graph. |
2307.16651 | Yu Wu | Yu Wu, Dimitris Spathis, Hong Jia, Ignacio Perez-Pozuelo, Tomas
Gonzales, Soren Brage, Nicholas Wareham, Cecilia Mascolo | UDAMA: Unsupervised Domain Adaptation through Multi-discriminator
Adversarial Training with Noisy Labels Improves Cardio-fitness Prediction | Accepted at Machine Learning for Healthcare (MLHC) 2023 | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Deep learning models have shown great promise in various healthcare
monitoring applications. However, most healthcare datasets with high-quality
(gold-standard) labels are small-scale, as directly collecting ground truth is
often costly and time-consuming. As a result, models developed and validated on
small-scale datasets often suffer from overfitting and do not generalize well
to unseen scenarios. At the same time, large amounts of imprecise
(silver-standard) labeled data, annotated by approximate methods with the help
of modern wearables and in the absence of ground truth validation, are starting
to emerge. However, due to measurement differences, this data displays
significant label distribution shifts, which motivates the use of domain
adaptation. To this end, we introduce UDAMA, a method with two key components:
Unsupervised Domain Adaptation and Multidiscriminator Adversarial Training,
where we pre-train on the silver-standard data and employ adversarial
adaptation with the gold-standard data along with two domain discriminators. In
particular, we showcase the practical potential of UDAMA by applying it to
Cardio-respiratory fitness (CRF) prediction. CRF is a crucial determinant of
metabolic disease and mortality, and it presents labels with various levels of
noise (goldand silver-standard), making it challenging to establish an accurate
prediction model. Our results show promising performance by alleviating
distribution shifts in various label shift settings. Additionally, by using
data from two free-living cohort studies (Fenland and BBVS), we show that UDAMA
consistently outperforms up to 12% compared to competitive transfer learning
and state-of-the-art domain adaptation models, paving the way for leveraging
noisy labeled data to improve fitness estimation at scale.
| [
{
"created": "Mon, 31 Jul 2023 13:31:53 GMT",
"version": "v1"
}
] | 2023-08-01 | [
[
"Wu",
"Yu",
""
],
[
"Spathis",
"Dimitris",
""
],
[
"Jia",
"Hong",
""
],
[
"Perez-Pozuelo",
"Ignacio",
""
],
[
"Gonzales",
"Tomas",
""
],
[
"Brage",
"Soren",
""
],
[
"Wareham",
"Nicholas",
""
],
[
"Mascolo",
"Cecilia",
""
]
] | Deep learning models have shown great promise in various healthcare monitoring applications. However, most healthcare datasets with high-quality (gold-standard) labels are small-scale, as directly collecting ground truth is often costly and time-consuming. As a result, models developed and validated on small-scale datasets often suffer from overfitting and do not generalize well to unseen scenarios. At the same time, large amounts of imprecise (silver-standard) labeled data, annotated by approximate methods with the help of modern wearables and in the absence of ground truth validation, are starting to emerge. However, due to measurement differences, this data displays significant label distribution shifts, which motivates the use of domain adaptation. To this end, we introduce UDAMA, a method with two key components: Unsupervised Domain Adaptation and Multidiscriminator Adversarial Training, where we pre-train on the silver-standard data and employ adversarial adaptation with the gold-standard data along with two domain discriminators. In particular, we showcase the practical potential of UDAMA by applying it to Cardio-respiratory fitness (CRF) prediction. CRF is a crucial determinant of metabolic disease and mortality, and it presents labels with various levels of noise (goldand silver-standard), making it challenging to establish an accurate prediction model. Our results show promising performance by alleviating distribution shifts in various label shift settings. Additionally, by using data from two free-living cohort studies (Fenland and BBVS), we show that UDAMA consistently outperforms up to 12% compared to competitive transfer learning and state-of-the-art domain adaptation models, paving the way for leveraging noisy labeled data to improve fitness estimation at scale. |
2403.14614 | Yuning Cui | Yuning Cui and Syed Waqas Zamir and Salman Khan and Alois Knoll and
Mubarak Shah and Fahad Shahbaz Khan | AdaIR: Adaptive All-in-One Image Restoration via Frequency Mining and
Modulation | 28 pages,15 figures | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In the image acquisition process, various forms of degradation, including
noise, haze, and rain, are frequently introduced. These degradations typically
arise from the inherent limitations of cameras or unfavorable ambient
conditions. To recover clean images from degraded versions, numerous
specialized restoration methods have been developed, each targeting a specific
type of degradation. Recently, all-in-one algorithms have garnered significant
attention by addressing different types of degradations within a single model
without requiring prior information of the input degradation type. However,
these methods purely operate in the spatial domain and do not delve into the
distinct frequency variations inherent to different degradation types. To
address this gap, we propose an adaptive all-in-one image restoration network
based on frequency mining and modulation. Our approach is motivated by the
observation that different degradation types impact the image content on
different frequency subbands, thereby requiring different treatments for each
restoration task. Specifically, we first mine low- and high-frequency
information from the input features, guided by the adaptively decoupled spectra
of the degraded image. The extracted features are then modulated by a
bidirectional operator to facilitate interactions between different frequency
components. Finally, the modulated features are merged into the original input
for a progressively guided restoration. With this approach, the model achieves
adaptive reconstruction by accentuating the informative frequency subbands
according to different input degradations. Extensive experiments demonstrate
that the proposed method achieves state-of-the-art performance on different
image restoration tasks, including denoising, dehazing, deraining, motion
deblurring, and low-light image enhancement. Our code is available at
https://github.com/c-yn/AdaIR.
| [
{
"created": "Thu, 21 Mar 2024 17:58:14 GMT",
"version": "v1"
}
] | 2024-03-22 | [
[
"Cui",
"Yuning",
""
],
[
"Zamir",
"Syed Waqas",
""
],
[
"Khan",
"Salman",
""
],
[
"Knoll",
"Alois",
""
],
[
"Shah",
"Mubarak",
""
],
[
"Khan",
"Fahad Shahbaz",
""
]
] | In the image acquisition process, various forms of degradation, including noise, haze, and rain, are frequently introduced. These degradations typically arise from the inherent limitations of cameras or unfavorable ambient conditions. To recover clean images from degraded versions, numerous specialized restoration methods have been developed, each targeting a specific type of degradation. Recently, all-in-one algorithms have garnered significant attention by addressing different types of degradations within a single model without requiring prior information of the input degradation type. However, these methods purely operate in the spatial domain and do not delve into the distinct frequency variations inherent to different degradation types. To address this gap, we propose an adaptive all-in-one image restoration network based on frequency mining and modulation. Our approach is motivated by the observation that different degradation types impact the image content on different frequency subbands, thereby requiring different treatments for each restoration task. Specifically, we first mine low- and high-frequency information from the input features, guided by the adaptively decoupled spectra of the degraded image. The extracted features are then modulated by a bidirectional operator to facilitate interactions between different frequency components. Finally, the modulated features are merged into the original input for a progressively guided restoration. With this approach, the model achieves adaptive reconstruction by accentuating the informative frequency subbands according to different input degradations. Extensive experiments demonstrate that the proposed method achieves state-of-the-art performance on different image restoration tasks, including denoising, dehazing, deraining, motion deblurring, and low-light image enhancement. Our code is available at https://github.com/c-yn/AdaIR. |
2202.02524 | Harichandana B S S | Harichandana B S S, Vibhav Agarwal, Sourav Ghosh, Gopi Ramena, Sumit
Kumar and Barath Raj Kandur Raja | PrivPAS: A real time Privacy-Preserving AI System and applied ethics | Accepted at 16th IEEE International Conference on Semantic Computing
(ICSC), January 26-28, 2022 [update: Best Paper candidate at ICSC 2022] | 2022 IEEE 16th International Conference on Semantic Computing
(ICSC), Laguna Hills, CA, USA, 2022, pp. 9-16 | 10.1109/ICSC52841.2022.00010 | null | cs.CV cs.CR | http://creativecommons.org/licenses/by-nc-nd/4.0/ | With 3.78 billion social media users worldwide in 2021 (48% of the human
population), almost 3 billion images are shared daily. At the same time, a
consistent evolution of smartphone cameras has led to a photography explosion
with 85% of all new pictures being captured using smartphones. However, lately,
there has been an increased discussion of privacy concerns when a person being
photographed is unaware of the picture being taken or has reservations about
the same being shared. These privacy violations are amplified for people with
disabilities, who may find it challenging to raise dissent even if they are
aware. Such unauthorized image captures may also be misused to gain sympathy by
third-party organizations, leading to a privacy breach. Privacy for people with
disabilities has so far received comparatively less attention from the AI
community. This motivates us to work towards a solution to generate
privacy-conscious cues for raising awareness in smartphone users of any
sensitivity in their viewfinder content. To this end, we introduce PrivPAS (A
real time Privacy-Preserving AI System) a novel framework to identify sensitive
content. Additionally, we curate and annotate a dataset to identify and
localize accessibility markers and classify whether an image is sensitive to a
featured subject with a disability. We demonstrate that the proposed
lightweight architecture, with a memory footprint of a mere 8.49MB, achieves a
high mAP of 89.52% on resource-constrained devices. Furthermore, our pipeline,
trained on face anonymized data, achieves an F1-score of 73.1%.
| [
{
"created": "Sat, 5 Feb 2022 09:52:54 GMT",
"version": "v1"
},
{
"created": "Tue, 8 Feb 2022 14:23:15 GMT",
"version": "v2"
}
] | 2022-04-05 | [
[
"S",
"Harichandana B S",
""
],
[
"Agarwal",
"Vibhav",
""
],
[
"Ghosh",
"Sourav",
""
],
[
"Ramena",
"Gopi",
""
],
[
"Kumar",
"Sumit",
""
],
[
"Raja",
"Barath Raj Kandur",
""
]
] | With 3.78 billion social media users worldwide in 2021 (48% of the human population), almost 3 billion images are shared daily. At the same time, a consistent evolution of smartphone cameras has led to a photography explosion with 85% of all new pictures being captured using smartphones. However, lately, there has been an increased discussion of privacy concerns when a person being photographed is unaware of the picture being taken or has reservations about the same being shared. These privacy violations are amplified for people with disabilities, who may find it challenging to raise dissent even if they are aware. Such unauthorized image captures may also be misused to gain sympathy by third-party organizations, leading to a privacy breach. Privacy for people with disabilities has so far received comparatively less attention from the AI community. This motivates us to work towards a solution to generate privacy-conscious cues for raising awareness in smartphone users of any sensitivity in their viewfinder content. To this end, we introduce PrivPAS (A real time Privacy-Preserving AI System) a novel framework to identify sensitive content. Additionally, we curate and annotate a dataset to identify and localize accessibility markers and classify whether an image is sensitive to a featured subject with a disability. We demonstrate that the proposed lightweight architecture, with a memory footprint of a mere 8.49MB, achieves a high mAP of 89.52% on resource-constrained devices. Furthermore, our pipeline, trained on face anonymized data, achieves an F1-score of 73.1%. |
1510.05860 | Ya-Feng Liu | Ya-Feng Liu | Dynamic Spectrum Management: A Complete Complexity Characterization | The paper has been accepted for publication in IEEE Transactions on
Information Theory | null | null | null | cs.IT cs.CC math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Consider a multi-user multi-carrier communication system where multiple users
share multiple discrete subcarriers. To achieve high spectrum efficiency, the
users in the system must choose their transmit power dynamically in response to
fast channel fluctuations. Assuming perfect channel state information, two
formulations for the spectrum management (power control) problem are considered
in this paper: the first is to minimize the total transmission power subject to
all users' transmission data rate constraints, and the second is to maximize
the min-rate utility subject to individual power constraints at each user. It
is known in the literature that both formulations of the problem are polynomial
time solvable when the number of subcarriers is one and strongly NP-hard when
the number of subcarriers are greater than or equal to three. However, the
complexity characterization of the problem when the number of subcarriers is
two has been missing for a long time. This paper answers this long-standing
open question: both formulations of the problem are strongly NP-hard when the
number of subcarriers is two.
| [
{
"created": "Tue, 20 Oct 2015 12:24:35 GMT",
"version": "v1"
},
{
"created": "Sat, 29 Oct 2016 00:26:26 GMT",
"version": "v2"
}
] | 2016-11-01 | [
[
"Liu",
"Ya-Feng",
""
]
] | Consider a multi-user multi-carrier communication system where multiple users share multiple discrete subcarriers. To achieve high spectrum efficiency, the users in the system must choose their transmit power dynamically in response to fast channel fluctuations. Assuming perfect channel state information, two formulations for the spectrum management (power control) problem are considered in this paper: the first is to minimize the total transmission power subject to all users' transmission data rate constraints, and the second is to maximize the min-rate utility subject to individual power constraints at each user. It is known in the literature that both formulations of the problem are polynomial time solvable when the number of subcarriers is one and strongly NP-hard when the number of subcarriers are greater than or equal to three. However, the complexity characterization of the problem when the number of subcarriers is two has been missing for a long time. This paper answers this long-standing open question: both formulations of the problem are strongly NP-hard when the number of subcarriers is two. |
2408.07191 | Jonas Linkerh\"agner | Jonas Linkerh\"agner, Cheng Shi, Ivan Dokmani\'c | Joint Graph Rewiring and Feature Denoising via Spectral Resonance | null | null | null | null | cs.LG cs.SI stat.ML | http://creativecommons.org/licenses/by/4.0/ | Graph neural networks (GNNs) take as input the graph structure and the
feature vectors associated with the nodes. Both contain noisy information about
the labels. Here we propose joint denoising and rewiring (JDR)--an algorithm to
jointly denoise the graph structure and features, which can improve the
performance of any downstream algorithm. We do this by defining and maximizing
the alignment between the leading eigenspaces of graph and feature matrices. To
approximately solve this computationally hard problem, we propose a heuristic
that efficiently handles real-world graph datasets with many classes and
different levels of homophily or heterophily. We experimentally verify the
effectiveness of our approach on synthetic data and real-world graph datasets.
The results show that JDR consistently outperforms existing rewiring methods on
node classification tasks using GNNs as downstream models.
| [
{
"created": "Tue, 13 Aug 2024 20:16:11 GMT",
"version": "v1"
}
] | 2024-08-15 | [
[
"Linkerhägner",
"Jonas",
""
],
[
"Shi",
"Cheng",
""
],
[
"Dokmanić",
"Ivan",
""
]
] | Graph neural networks (GNNs) take as input the graph structure and the feature vectors associated with the nodes. Both contain noisy information about the labels. Here we propose joint denoising and rewiring (JDR)--an algorithm to jointly denoise the graph structure and features, which can improve the performance of any downstream algorithm. We do this by defining and maximizing the alignment between the leading eigenspaces of graph and feature matrices. To approximately solve this computationally hard problem, we propose a heuristic that efficiently handles real-world graph datasets with many classes and different levels of homophily or heterophily. We experimentally verify the effectiveness of our approach on synthetic data and real-world graph datasets. The results show that JDR consistently outperforms existing rewiring methods on node classification tasks using GNNs as downstream models. |
2201.04402 | Ekrem \c{C}etinkaya | Ekrem \c{C}etinkaya and Minh Nguyen and Christian Timmerer | MoViDNN: A Mobile Platform for Evaluating Video Quality Enhancement with
Deep Neural Networks | 8 pages, 3 figures | MMM 2022: MultiMedia Modeling pp 465-472 | 10.1007/978-3-030-98355-0_40 | null | cs.CV cs.MM | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Deep neural network (DNN) based approaches have been intensively studied to
improve video quality thanks to their fast advancement in recent years. These
approaches are designed mainly for desktop devices due to their high
computational cost. However, with the increasing performance of mobile devices
in recent years, it became possible to execute DNN based approaches in mobile
devices. Despite having the required computational power, utilizing DNNs to
improve the video quality for mobile devices is still an active research area.
In this paper, we propose an open-source mobile platform, namely MoViDNN, to
evaluate DNN based video quality enhancement methods, such as super-resolution,
denoising, and deblocking. Our proposed platform can be used to evaluate the
DNN based approaches both objectively and subjectively. For objective
evaluation, we report common metrics such as execution time, PSNR, and SSIM.
For subjective evaluation, Mean Score Opinion (MOS) is reported. The proposed
platform is available publicly at https://github.com/cd-athena/MoViDNN
| [
{
"created": "Wed, 12 Jan 2022 10:38:04 GMT",
"version": "v1"
}
] | 2022-03-22 | [
[
"Çetinkaya",
"Ekrem",
""
],
[
"Nguyen",
"Minh",
""
],
[
"Timmerer",
"Christian",
""
]
] | Deep neural network (DNN) based approaches have been intensively studied to improve video quality thanks to their fast advancement in recent years. These approaches are designed mainly for desktop devices due to their high computational cost. However, with the increasing performance of mobile devices in recent years, it became possible to execute DNN based approaches in mobile devices. Despite having the required computational power, utilizing DNNs to improve the video quality for mobile devices is still an active research area. In this paper, we propose an open-source mobile platform, namely MoViDNN, to evaluate DNN based video quality enhancement methods, such as super-resolution, denoising, and deblocking. Our proposed platform can be used to evaluate the DNN based approaches both objectively and subjectively. For objective evaluation, we report common metrics such as execution time, PSNR, and SSIM. For subjective evaluation, Mean Score Opinion (MOS) is reported. The proposed platform is available publicly at https://github.com/cd-athena/MoViDNN |
1908.08332 | Luis Cruz | Luis Cruz, Rui Abreu, John Grundy, Li Li, Xin Xia | Do Energy-oriented Changes Hinder Maintainability? | International Conference on Software Maintenance and Evolution -
ICSME 2019 | null | null | null | cs.SE cs.PF | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Energy efficiency is a crucial quality requirement for mobile applications.
However, improving energy efficiency is far from trivial as developers lack the
knowledge and tools to aid in this activity. In this paper we study the impact
of changes to improve energy efficiency on the maintainability of Android
applications. Using a dataset containing 539 energy efficiency-oriented
commits, we measure maintainability -- as computed by the Software Improvement
Group's web-based source code analysis service Better Code Hub (BCH) -- before
and after energy efficiency-related code changes. Results show that in general
improving energy efficiency comes with a significant decrease in
maintainability. This is particularly evident in code changes to accommodate
the Power Save Mode and Wakelock Addition energy patterns. In addition, we
perform manual analysis to assess how real examples of energy-oriented changes
affect maintainability. Our results help mobile app developers to 1) avoid
common maintainability issues when improving the energy efficiency of their
apps; and 2) adopt development processes to build maintainable and
energy-efficient code. We also support researchers by identifying challenges in
mobile app development that still need to be addressed.
| [
{
"created": "Thu, 22 Aug 2019 12:21:08 GMT",
"version": "v1"
}
] | 2019-08-29 | [
[
"Cruz",
"Luis",
""
],
[
"Abreu",
"Rui",
""
],
[
"Grundy",
"John",
""
],
[
"Li",
"Li",
""
],
[
"Xia",
"Xin",
""
]
] | Energy efficiency is a crucial quality requirement for mobile applications. However, improving energy efficiency is far from trivial as developers lack the knowledge and tools to aid in this activity. In this paper we study the impact of changes to improve energy efficiency on the maintainability of Android applications. Using a dataset containing 539 energy efficiency-oriented commits, we measure maintainability -- as computed by the Software Improvement Group's web-based source code analysis service Better Code Hub (BCH) -- before and after energy efficiency-related code changes. Results show that in general improving energy efficiency comes with a significant decrease in maintainability. This is particularly evident in code changes to accommodate the Power Save Mode and Wakelock Addition energy patterns. In addition, we perform manual analysis to assess how real examples of energy-oriented changes affect maintainability. Our results help mobile app developers to 1) avoid common maintainability issues when improving the energy efficiency of their apps; and 2) adopt development processes to build maintainable and energy-efficient code. We also support researchers by identifying challenges in mobile app development that still need to be addressed. |
1403.2508 | Rajib Das | Sunirmal Khatua, Preetam K. Sur, Rajib K. Das and Nandini Mukherjee | Heuristic-based Optimal Resource Provisioning in Application-centric
Cloud | null | null | null | null | cs.DC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Cloud Service Providers (CSPs) adapt different pricing models for their
offered services. Some of the models are suitable for short term requirement
while others may be suitable for the Cloud Service User's (CSU) long term
requirement. In this paper, we look at the problem of finding the amount of
resources to be reserved to satisfy the CSU's long term demands with the aim of
minimizing the total cost. Finding the optimal resource requirement to satisfy
the the CSU's demand for resources needs sufficient research effort. Various
algorithms were discussed in the last couple of years for finding the optimal
resource requirement but most of them are based on IPP which is NP in nature.
In this paper, we derive some heuristic-based polynomial time algorithms to
find some near optimal solution to the problem. We show that the cost for CSU
using our approach is comparable to the solution obtained using optimal Integer
Programming Problem(IPP).
| [
{
"created": "Tue, 11 Mar 2014 09:07:16 GMT",
"version": "v1"
}
] | 2014-03-12 | [
[
"Khatua",
"Sunirmal",
""
],
[
"Sur",
"Preetam K.",
""
],
[
"Das",
"Rajib K.",
""
],
[
"Mukherjee",
"Nandini",
""
]
] | Cloud Service Providers (CSPs) adapt different pricing models for their offered services. Some of the models are suitable for short term requirement while others may be suitable for the Cloud Service User's (CSU) long term requirement. In this paper, we look at the problem of finding the amount of resources to be reserved to satisfy the CSU's long term demands with the aim of minimizing the total cost. Finding the optimal resource requirement to satisfy the the CSU's demand for resources needs sufficient research effort. Various algorithms were discussed in the last couple of years for finding the optimal resource requirement but most of them are based on IPP which is NP in nature. In this paper, we derive some heuristic-based polynomial time algorithms to find some near optimal solution to the problem. We show that the cost for CSU using our approach is comparable to the solution obtained using optimal Integer Programming Problem(IPP). |
1211.3719 | Athanasios Lioumpas S. | Athanasios S. Lioumpas, Petros S. Bithas, Angeliki Alexiou | Partitioning of Distributed MIMO Systems based on Overhead
Considerations | IEEE Wireless Communications Letters | null | 10.1109/WCL.2013.072913.130449 | null | cs.NI cs.IT math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Distributed-Multiple Input Multiple Output (DMIMO) networks is a promising
enabler to address the challenges of high traffic demand in future wireless
networks. A limiting factor that is directly related to the performance of
these systems is the overhead signaling required for distributing data and
control information among the network elements. In this paper, the concept of
orthogonal partitioning is extended to D-MIMO networks employing joint
multi-user beamforming, aiming to maximize the effective sum-rate, i.e., the
actual transmitted information data. Furthermore, in order to comply with
practical requirements, the overhead subframe size is considered to be
constrained. In this context, a novel formulation of constrained orthogonal
partitioning is introduced as an elegant Knapsack optimization problem, which
allows the derivation of quick and accurate solutions. Several numerical
results give insight into the capabilities of D-MIMO networks and the actual
sum-rate scaling under overhead constraints.
| [
{
"created": "Thu, 15 Nov 2012 20:21:29 GMT",
"version": "v1"
},
{
"created": "Fri, 16 Nov 2012 17:18:49 GMT",
"version": "v2"
},
{
"created": "Sun, 21 Jul 2013 19:49:23 GMT",
"version": "v3"
}
] | 2016-11-18 | [
[
"Lioumpas",
"Athanasios S.",
""
],
[
"Bithas",
"Petros S.",
""
],
[
"Alexiou",
"Angeliki",
""
]
] | Distributed-Multiple Input Multiple Output (DMIMO) networks is a promising enabler to address the challenges of high traffic demand in future wireless networks. A limiting factor that is directly related to the performance of these systems is the overhead signaling required for distributing data and control information among the network elements. In this paper, the concept of orthogonal partitioning is extended to D-MIMO networks employing joint multi-user beamforming, aiming to maximize the effective sum-rate, i.e., the actual transmitted information data. Furthermore, in order to comply with practical requirements, the overhead subframe size is considered to be constrained. In this context, a novel formulation of constrained orthogonal partitioning is introduced as an elegant Knapsack optimization problem, which allows the derivation of quick and accurate solutions. Several numerical results give insight into the capabilities of D-MIMO networks and the actual sum-rate scaling under overhead constraints. |
2006.12779 | Francesco Cicala | Francesco Cicala, Luca Bortolussi | Density-embedding layers: a general framework for adaptive receptive
fields | 13 pages, 2 figures, submitted to NeurIPS 2020 | null | null | null | cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The effectiveness and performance of artificial neural networks, particularly
for visual tasks, depends in crucial ways on the receptive field of neurons.
The receptive field itself depends on the interplay between several
architectural aspects, including sparsity, pooling, and activation functions.
In recent literature there are several ad hoc proposals trying to make
receptive fields more flexible and adaptive to data. For instance, different
parameterizations of convolutional and pooling layers have been proposed to
increase their adaptivity. In this paper, we propose the novel theoretical
framework of density-embedded layers, generalizing the transformation
represented by a neuron. Specifically, the affine transformation applied on the
input is replaced by a scalar product of the input, suitably represented as a
piecewise constant function, with a density function associated with the
neuron. This density is shown to describe directly the receptive field of the
neuron. Crucially, by suitably representing such a density as a linear
combination of a parametric family of functions, we can efficiently train the
densities by means of any automatic differentiation system, making it adaptable
to the problem at hand, and computationally efficient to evaluate. This
framework captures and generalizes recent methods, allowing a fine tuning of
the receptive field. In the paper, we define some novel layers and we
experimentally validate them on the classic MNIST dataset.
| [
{
"created": "Tue, 23 Jun 2020 06:09:16 GMT",
"version": "v1"
},
{
"created": "Mon, 6 Jul 2020 07:36:24 GMT",
"version": "v2"
}
] | 2020-07-07 | [
[
"Cicala",
"Francesco",
""
],
[
"Bortolussi",
"Luca",
""
]
] | The effectiveness and performance of artificial neural networks, particularly for visual tasks, depends in crucial ways on the receptive field of neurons. The receptive field itself depends on the interplay between several architectural aspects, including sparsity, pooling, and activation functions. In recent literature there are several ad hoc proposals trying to make receptive fields more flexible and adaptive to data. For instance, different parameterizations of convolutional and pooling layers have been proposed to increase their adaptivity. In this paper, we propose the novel theoretical framework of density-embedded layers, generalizing the transformation represented by a neuron. Specifically, the affine transformation applied on the input is replaced by a scalar product of the input, suitably represented as a piecewise constant function, with a density function associated with the neuron. This density is shown to describe directly the receptive field of the neuron. Crucially, by suitably representing such a density as a linear combination of a parametric family of functions, we can efficiently train the densities by means of any automatic differentiation system, making it adaptable to the problem at hand, and computationally efficient to evaluate. This framework captures and generalizes recent methods, allowing a fine tuning of the receptive field. In the paper, we define some novel layers and we experimentally validate them on the classic MNIST dataset. |
2303.14828 | Dina Bashkirova | Dina Bashkirova, Samarth Mishra, Diala Lteif, Piotr Teterwak, Donghyun
Kim, Fadi Alladkani, James Akl, Berk Calli, Sarah Adel Bargal, Kate Saenko,
Daehan Kim, Minseok Seo, YoungJin Jeon, Dong-Geol Choi, Shahaf Ettedgui, Raja
Giryes, Shady Abu-Hussein, Binhui Xie, Shuang Li | VisDA 2022 Challenge: Domain Adaptation for Industrial Waste Sorting | Proceedings of Machine Learning Research | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Label-efficient and reliable semantic segmentation is essential for many
real-life applications, especially for industrial settings with high visual
diversity, such as waste sorting. In industrial waste sorting, one of the
biggest challenges is the extreme diversity of the input stream depending on
factors like the location of the sorting facility, the equipment available in
the facility, and the time of year, all of which significantly impact the
composition and visual appearance of the waste stream. These changes in the
data are called ``visual domains'', and label-efficient adaptation of models to
such domains is needed for successful semantic segmentation of industrial
waste. To test the abilities of computer vision models on this task, we present
the VisDA 2022 Challenge on Domain Adaptation for Industrial Waste Sorting. Our
challenge incorporates a fully-annotated waste sorting dataset, ZeroWaste,
collected from two real material recovery facilities in different locations and
seasons, as well as a novel procedurally generated synthetic waste sorting
dataset, SynthWaste. In this competition, we aim to answer two questions: 1)
can we leverage domain adaptation techniques to minimize the domain gap? and 2)
can synthetic data augmentation improve performance on this task and help adapt
to changing data distributions? The results of the competition show that
industrial waste detection poses a real domain adaptation problem, that domain
generalization techniques such as augmentations, ensembling, etc., improve the
overall performance on the unlabeled target domain examples, and that
leveraging synthetic data effectively remains an open problem. See
https://ai.bu.edu/visda-2022/
| [
{
"created": "Sun, 26 Mar 2023 21:38:38 GMT",
"version": "v1"
}
] | 2023-03-28 | [
[
"Bashkirova",
"Dina",
""
],
[
"Mishra",
"Samarth",
""
],
[
"Lteif",
"Diala",
""
],
[
"Teterwak",
"Piotr",
""
],
[
"Kim",
"Donghyun",
""
],
[
"Alladkani",
"Fadi",
""
],
[
"Akl",
"James",
""
],
[
"Calli",
"Berk",
""
],
[
"Bargal",
"Sarah Adel",
""
],
[
"Saenko",
"Kate",
""
],
[
"Kim",
"Daehan",
""
],
[
"Seo",
"Minseok",
""
],
[
"Jeon",
"YoungJin",
""
],
[
"Choi",
"Dong-Geol",
""
],
[
"Ettedgui",
"Shahaf",
""
],
[
"Giryes",
"Raja",
""
],
[
"Abu-Hussein",
"Shady",
""
],
[
"Xie",
"Binhui",
""
],
[
"Li",
"Shuang",
""
]
] | Label-efficient and reliable semantic segmentation is essential for many real-life applications, especially for industrial settings with high visual diversity, such as waste sorting. In industrial waste sorting, one of the biggest challenges is the extreme diversity of the input stream depending on factors like the location of the sorting facility, the equipment available in the facility, and the time of year, all of which significantly impact the composition and visual appearance of the waste stream. These changes in the data are called ``visual domains'', and label-efficient adaptation of models to such domains is needed for successful semantic segmentation of industrial waste. To test the abilities of computer vision models on this task, we present the VisDA 2022 Challenge on Domain Adaptation for Industrial Waste Sorting. Our challenge incorporates a fully-annotated waste sorting dataset, ZeroWaste, collected from two real material recovery facilities in different locations and seasons, as well as a novel procedurally generated synthetic waste sorting dataset, SynthWaste. In this competition, we aim to answer two questions: 1) can we leverage domain adaptation techniques to minimize the domain gap? and 2) can synthetic data augmentation improve performance on this task and help adapt to changing data distributions? The results of the competition show that industrial waste detection poses a real domain adaptation problem, that domain generalization techniques such as augmentations, ensembling, etc., improve the overall performance on the unlabeled target domain examples, and that leveraging synthetic data effectively remains an open problem. See https://ai.bu.edu/visda-2022/ |
1405.2199 | Madhumangal Pal Dr. | Madhumangal Pal and Anita Pal | Scheduling algorithm to select $k$ optimal programme slots in television
channels: A graph theoretic approach | 25 pages | null | null | null | cs.DM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, it is shown that all programmes of all television channels can
be modelled as an interval graph. The programme slots are taken as the vertices
of the graph and if the time duration of two {programme slots} have non-empty
intersection, the corresponding vertices are considered to be connected by an
edge. The number of viewers of a programme is taken as the weight of the
vertex. A set of programmes that are mutually exclusive in respect of time
scheduling is called a session. We assume that a company sets the objective of
selecting the popular programmes in $k$ parallel sessions among different
channels so as to make its commercial advertisement reach the maximum number of
viewers, that is, a company selects $k$ suitable programme slots simultaneously
for advertisement. The aim of the paper is, therefore, to {help} the companies
to select the programme slots, which are mutually exclusive with respect to the
time schedule of telecasting time, in such a way that the total number of
viewers of the selected programme in $k$ parallel slots rises to the optimum
level. It is shown that the solution of this problem is obtained by solving the
maximum weight $k$-colouring problem on an interval {graph}. An algorithm is
designed to solve this just-in-time optimization problem using $O(kMn^2)$ time,
where $n$ and $M$ represent the total number of programmes of all channels and
the upper bound of the viewers of all programmes of all channels respectively.
The problem considered in this paper is a daily life problem which is modeled
by $k$-colouring problem on interval graph.
| [
{
"created": "Fri, 9 May 2014 10:29:10 GMT",
"version": "v1"
}
] | 2014-05-12 | [
[
"Pal",
"Madhumangal",
""
],
[
"Pal",
"Anita",
""
]
] | In this paper, it is shown that all programmes of all television channels can be modelled as an interval graph. The programme slots are taken as the vertices of the graph and if the time duration of two {programme slots} have non-empty intersection, the corresponding vertices are considered to be connected by an edge. The number of viewers of a programme is taken as the weight of the vertex. A set of programmes that are mutually exclusive in respect of time scheduling is called a session. We assume that a company sets the objective of selecting the popular programmes in $k$ parallel sessions among different channels so as to make its commercial advertisement reach the maximum number of viewers, that is, a company selects $k$ suitable programme slots simultaneously for advertisement. The aim of the paper is, therefore, to {help} the companies to select the programme slots, which are mutually exclusive with respect to the time schedule of telecasting time, in such a way that the total number of viewers of the selected programme in $k$ parallel slots rises to the optimum level. It is shown that the solution of this problem is obtained by solving the maximum weight $k$-colouring problem on an interval {graph}. An algorithm is designed to solve this just-in-time optimization problem using $O(kMn^2)$ time, where $n$ and $M$ represent the total number of programmes of all channels and the upper bound of the viewers of all programmes of all channels respectively. The problem considered in this paper is a daily life problem which is modeled by $k$-colouring problem on interval graph. |
2103.10107 | Luk\'a\v{s} Picek | Luk\'a\v{s} Picek, Milan \v{S}ulc, Ji\v{r}\'i Matas, Jacob
Heilmann-Clausen, Thomas S. Jeppesen, Thomas L{\ae}ss{\o}e, Tobias Fr{\o}slev | Danish Fungi 2020 -- Not Just Another Image Recognition Dataset | null | null | 10.1109/WACV51458.2022.00334 | null | cs.CV eess.IV | http://creativecommons.org/licenses/by/4.0/ | We introduce a novel fine-grained dataset and benchmark, the Danish Fungi
2020 (DF20). The dataset, constructed from observations submitted to the Atlas
of Danish Fungi, is unique in its taxonomy-accurate class labels, small number
of errors, highly unbalanced long-tailed class distribution, rich observation
metadata, and well-defined class hierarchy. DF20 has zero overlap with
ImageNet, allowing unbiased comparison of models fine-tuned from publicly
available ImageNet checkpoints. The proposed evaluation protocol enables
testing the ability to improve classification using metadata -- e.g. precise
geographic location, habitat, and substrate, facilitates classifier calibration
testing, and finally allows to study the impact of the device settings on the
classification performance. Experiments using Convolutional Neural Networks
(CNN) and the recent Vision Transformers (ViT) show that DF20 presents a
challenging task. Interestingly, ViT achieves results superior to CNN baselines
with 80.45% accuracy and 0.743 macro F1 score, reducing the CNN error by 9% and
12% respectively. A simple procedure for including metadata into the decision
process improves the classification accuracy by more than 2.95 percentage
points, reducing the error rate by 15%. The source code for all methods and
experiments is available at https://sites.google.com/view/danish-fungi-dataset.
| [
{
"created": "Thu, 18 Mar 2021 09:33:11 GMT",
"version": "v1"
},
{
"created": "Fri, 19 Mar 2021 12:15:47 GMT",
"version": "v2"
},
{
"created": "Mon, 22 Mar 2021 08:43:04 GMT",
"version": "v3"
},
{
"created": "Fri, 20 Aug 2021 14:35:44 GMT",
"version": "v4"
}
] | 2022-06-13 | [
[
"Picek",
"Lukáš",
""
],
[
"Šulc",
"Milan",
""
],
[
"Matas",
"Jiří",
""
],
[
"Heilmann-Clausen",
"Jacob",
""
],
[
"Jeppesen",
"Thomas S.",
""
],
[
"Læssøe",
"Thomas",
""
],
[
"Frøslev",
"Tobias",
""
]
] | We introduce a novel fine-grained dataset and benchmark, the Danish Fungi 2020 (DF20). The dataset, constructed from observations submitted to the Atlas of Danish Fungi, is unique in its taxonomy-accurate class labels, small number of errors, highly unbalanced long-tailed class distribution, rich observation metadata, and well-defined class hierarchy. DF20 has zero overlap with ImageNet, allowing unbiased comparison of models fine-tuned from publicly available ImageNet checkpoints. The proposed evaluation protocol enables testing the ability to improve classification using metadata -- e.g. precise geographic location, habitat, and substrate, facilitates classifier calibration testing, and finally allows to study the impact of the device settings on the classification performance. Experiments using Convolutional Neural Networks (CNN) and the recent Vision Transformers (ViT) show that DF20 presents a challenging task. Interestingly, ViT achieves results superior to CNN baselines with 80.45% accuracy and 0.743 macro F1 score, reducing the CNN error by 9% and 12% respectively. A simple procedure for including metadata into the decision process improves the classification accuracy by more than 2.95 percentage points, reducing the error rate by 15%. The source code for all methods and experiments is available at https://sites.google.com/view/danish-fungi-dataset. |
1605.00398 | Akshay Khatri | Akshay Khatri, Sankalp Kolhe, Nupur Giri | Dynamic Address Allocation Algorithm for Mobile Ad hoc Networks | null | null | null | null | cs.NI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A Mobile Ad hoc network (MANET) consists of nodes which use multi-hop
communication to establish connection between nodes. Traditional infrastructure
based systems use a centralized architecture for address allocation. However,
this is not possible in Ad hoc networks due to their dynamic structure. Many
schemes have been proposed to solve this problem, but most of them use
network-wide broadcasts to ensure the availability of a new address. This
becomes extremely difficult as network size grows. In this paper, we propose an
address allocation algorithm which avoids network-wide broadcasts to allocate
address to a new node. Moreover, the algorithm allocates addresses dynamically
such that the network maintains an "IP resembles topology" state. In such a
state, routing becomes easier and the overall overhead in communication is
reduced. This algorithm is particularly useful for routing protocols which use
topology information to route messages in the network. Our solution is designed
with scalability in mind such that the cost of address assignment to a new node
is independent of the number of nodes in the network.
| [
{
"created": "Mon, 2 May 2016 09:10:44 GMT",
"version": "v1"
}
] | 2016-05-03 | [
[
"Khatri",
"Akshay",
""
],
[
"Kolhe",
"Sankalp",
""
],
[
"Giri",
"Nupur",
""
]
] | A Mobile Ad hoc network (MANET) consists of nodes which use multi-hop communication to establish connection between nodes. Traditional infrastructure based systems use a centralized architecture for address allocation. However, this is not possible in Ad hoc networks due to their dynamic structure. Many schemes have been proposed to solve this problem, but most of them use network-wide broadcasts to ensure the availability of a new address. This becomes extremely difficult as network size grows. In this paper, we propose an address allocation algorithm which avoids network-wide broadcasts to allocate address to a new node. Moreover, the algorithm allocates addresses dynamically such that the network maintains an "IP resembles topology" state. In such a state, routing becomes easier and the overall overhead in communication is reduced. This algorithm is particularly useful for routing protocols which use topology information to route messages in the network. Our solution is designed with scalability in mind such that the cost of address assignment to a new node is independent of the number of nodes in the network. |
2404.14406 | Kartik Narayan | Kartik Narayan, Vishal M. Patel | Hyp-OC: Hyperbolic One Class Classification for Face Anti-Spoofing | Accepted in FG2024, Project Page -
https://kartik-3004.github.io/hyp-oc/ | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Face recognition technology has become an integral part of modern security
systems and user authentication processes. However, these systems are
vulnerable to spoofing attacks and can easily be circumvented. Most prior
research in face anti-spoofing (FAS) approaches it as a two-class
classification task where models are trained on real samples and known spoof
attacks and tested for detection performance on unknown spoof attacks. However,
in practice, FAS should be treated as a one-class classification task where,
while training, one cannot assume any knowledge regarding the spoof samples a
priori. In this paper, we reformulate the face anti-spoofing task from a
one-class perspective and propose a novel hyperbolic one-class classification
framework. To train our network, we use a pseudo-negative class sampled from
the Gaussian distribution with a weighted running mean and propose two novel
loss functions: (1) Hyp-PC: Hyperbolic Pairwise Confusion loss, and (2) Hyp-CE:
Hyperbolic Cross Entropy loss, which operate in the hyperbolic space.
Additionally, we employ Euclidean feature clipping and gradient clipping to
stabilize the training in the hyperbolic space. To the best of our knowledge,
this is the first work extending hyperbolic embeddings for face anti-spoofing
in a one-class manner. With extensive experiments on five benchmark datasets:
Rose-Youtu, MSU-MFSD, CASIA-MFSD, Idiap Replay-Attack, and OULU-NPU, we
demonstrate that our method significantly outperforms the state-of-the-art,
achieving better spoof detection performance.
| [
{
"created": "Mon, 22 Apr 2024 17:59:18 GMT",
"version": "v1"
}
] | 2024-04-23 | [
[
"Narayan",
"Kartik",
""
],
[
"Patel",
"Vishal M.",
""
]
] | Face recognition technology has become an integral part of modern security systems and user authentication processes. However, these systems are vulnerable to spoofing attacks and can easily be circumvented. Most prior research in face anti-spoofing (FAS) approaches it as a two-class classification task where models are trained on real samples and known spoof attacks and tested for detection performance on unknown spoof attacks. However, in practice, FAS should be treated as a one-class classification task where, while training, one cannot assume any knowledge regarding the spoof samples a priori. In this paper, we reformulate the face anti-spoofing task from a one-class perspective and propose a novel hyperbolic one-class classification framework. To train our network, we use a pseudo-negative class sampled from the Gaussian distribution with a weighted running mean and propose two novel loss functions: (1) Hyp-PC: Hyperbolic Pairwise Confusion loss, and (2) Hyp-CE: Hyperbolic Cross Entropy loss, which operate in the hyperbolic space. Additionally, we employ Euclidean feature clipping and gradient clipping to stabilize the training in the hyperbolic space. To the best of our knowledge, this is the first work extending hyperbolic embeddings for face anti-spoofing in a one-class manner. With extensive experiments on five benchmark datasets: Rose-Youtu, MSU-MFSD, CASIA-MFSD, Idiap Replay-Attack, and OULU-NPU, we demonstrate that our method significantly outperforms the state-of-the-art, achieving better spoof detection performance. |
2212.07618 | Mengnan Shi | Bohao Li, Chang Liu, Mengnan Shi, Xiaozhong Chen, Xiangyang Ji,
Qixiang Ye | Proposal Distribution Calibration for Few-Shot Object Detection | This paper is under review in IEEE TNNLS | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Adapting object detectors learned with sufficient supervision to novel
classes under low data regimes is charming yet challenging. In few-shot object
detection (FSOD), the two-step training paradigm is widely adopted to mitigate
the severe sample imbalance, i.e., holistic pre-training on base classes, then
partial fine-tuning in a balanced setting with all classes. Since unlabeled
instances are suppressed as backgrounds in the base training phase, the learned
RPN is prone to produce biased proposals for novel instances, resulting in
dramatic performance degradation. Unfortunately, the extreme data scarcity
aggravates the proposal distribution bias, hindering the RoI head from evolving
toward novel classes. In this paper, we introduce a simple yet effective
proposal distribution calibration (PDC) approach to neatly enhance the
localization and classification abilities of the RoI head by recycling its
localization ability endowed in base training and enriching high-quality
positive samples for semantic fine-tuning. Specifically, we sample proposals
based on the base proposal statistics to calibrate the distribution bias and
impose additional localization and classification losses upon the sampled
proposals for fast expanding the base detector to novel classes. Experiments on
the commonly used Pascal VOC and MS COCO datasets with explicit
state-of-the-art performances justify the efficacy of our PDC for FSOD. Code is
available at github.com/Bohao-Lee/PDC.
| [
{
"created": "Thu, 15 Dec 2022 05:09:11 GMT",
"version": "v1"
}
] | 2022-12-16 | [
[
"Li",
"Bohao",
""
],
[
"Liu",
"Chang",
""
],
[
"Shi",
"Mengnan",
""
],
[
"Chen",
"Xiaozhong",
""
],
[
"Ji",
"Xiangyang",
""
],
[
"Ye",
"Qixiang",
""
]
] | Adapting object detectors learned with sufficient supervision to novel classes under low data regimes is charming yet challenging. In few-shot object detection (FSOD), the two-step training paradigm is widely adopted to mitigate the severe sample imbalance, i.e., holistic pre-training on base classes, then partial fine-tuning in a balanced setting with all classes. Since unlabeled instances are suppressed as backgrounds in the base training phase, the learned RPN is prone to produce biased proposals for novel instances, resulting in dramatic performance degradation. Unfortunately, the extreme data scarcity aggravates the proposal distribution bias, hindering the RoI head from evolving toward novel classes. In this paper, we introduce a simple yet effective proposal distribution calibration (PDC) approach to neatly enhance the localization and classification abilities of the RoI head by recycling its localization ability endowed in base training and enriching high-quality positive samples for semantic fine-tuning. Specifically, we sample proposals based on the base proposal statistics to calibrate the distribution bias and impose additional localization and classification losses upon the sampled proposals for fast expanding the base detector to novel classes. Experiments on the commonly used Pascal VOC and MS COCO datasets with explicit state-of-the-art performances justify the efficacy of our PDC for FSOD. Code is available at github.com/Bohao-Lee/PDC. |
1612.08845 | Toni Heidenreich | Toni Heidenreich | The formal-logical characterisation of lies, deception, and associated
notions | Literature review prepared as a student at King's College London | null | null | null | cs.LO cs.AI cs.MA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Defining various dishonest notions in a formal way is a key step to enable
intelligent agents to act in untrustworthy environments. This review evaluates
the literature for this topic by looking at formal definitions based on modal
logic as well as other formal approaches. Criteria from philosophical
groundwork is used to assess the definitions for correctness and completeness.
The key contribution of this review is to show that only a few definitions
fully comply with this gold standard and to point out the missing steps towards
a successful application of these definitions in an actual agent environment.
| [
{
"created": "Wed, 28 Dec 2016 10:35:05 GMT",
"version": "v1"
}
] | 2016-12-30 | [
[
"Heidenreich",
"Toni",
""
]
] | Defining various dishonest notions in a formal way is a key step to enable intelligent agents to act in untrustworthy environments. This review evaluates the literature for this topic by looking at formal definitions based on modal logic as well as other formal approaches. Criteria from philosophical groundwork is used to assess the definitions for correctness and completeness. The key contribution of this review is to show that only a few definitions fully comply with this gold standard and to point out the missing steps towards a successful application of these definitions in an actual agent environment. |
2006.11456 | Abiola Osho | Abiola Osho and Ethan Tucker and George Amariucai | Implicit Crowdsourcing for Identifying Abusive Behavior in Online Social
Networks | null | null | null | null | cs.SI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The increased use of online social networks for the dissemination of
information comes with the misuse of the internet for cyberbullying,
cybercrime, spam, vandalism, amongst other things. To proactively identify
abuse in the networks, we propose a model to identify abusive posts by
crowdsourcing. The crowdsourcing part of the detection mechanism is implemented
implicitly, by simply observing the natural interaction between users
encountering the messages. We explore the node-to-node spread of information on
Twitter and propose a model that predicts the abuse level (abusive, hate, spam,
normal) associated with the tweet by observing the attributes of the message,
along with those of the users interacting with it. We demonstrate that the
difference in users' interactions with abusive posts can be leveraged in
identifying posts of varying abuse levels.
| [
{
"created": "Sat, 20 Jun 2020 01:14:30 GMT",
"version": "v1"
}
] | 2020-06-23 | [
[
"Osho",
"Abiola",
""
],
[
"Tucker",
"Ethan",
""
],
[
"Amariucai",
"George",
""
]
] | The increased use of online social networks for the dissemination of information comes with the misuse of the internet for cyberbullying, cybercrime, spam, vandalism, amongst other things. To proactively identify abuse in the networks, we propose a model to identify abusive posts by crowdsourcing. The crowdsourcing part of the detection mechanism is implemented implicitly, by simply observing the natural interaction between users encountering the messages. We explore the node-to-node spread of information on Twitter and propose a model that predicts the abuse level (abusive, hate, spam, normal) associated with the tweet by observing the attributes of the message, along with those of the users interacting with it. We demonstrate that the difference in users' interactions with abusive posts can be leveraged in identifying posts of varying abuse levels. |
2001.09046 | Bart Smets | Bart Smets, Jim Portegies, Erik Bekkers, Remco Duits | PDE-based Group Equivariant Convolutional Neural Networks | 27 pages, 18 figures. v2 changes: - mentioned KerCNNs - added section
Generalization of G-CNNs - clarification that the experiments utilized
automatic differentiation and SGD. v3 changes: - streamlined theoretical
framework - formulation and proof Thm.1 & 2 - expanded experiments. v4
changes: typos in Prop.5 and (20) v5/6 changes: minor revision | null | 10.1007/s10851-022-01114-x | null | cs.LG cs.CV math.DG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present a PDE-based framework that generalizes Group equivariant
Convolutional Neural Networks (G-CNNs). In this framework, a network layer is
seen as a set of PDE-solvers where geometrically meaningful PDE-coefficients
become the layer's trainable weights. Formulating our PDEs on homogeneous
spaces allows these networks to be designed with built-in symmetries such as
rotation in addition to the standard translation equivariance of CNNs.
Having all the desired symmetries included in the design obviates the need to
include them by means of costly techniques such as data augmentation. We will
discuss our PDE-based G-CNNs (PDE-G-CNNs) in a general homogeneous space
setting while also going into the specifics of our primary case of interest:
roto-translation equivariance.
We solve the PDE of interest by a combination of linear group convolutions
and non-linear morphological group convolutions with analytic kernel
approximations that we underpin with formal theorems. Our kernel approximations
allow for fast GPU-implementation of the PDE-solvers, we release our
implementation with this article in the form of the LieTorch extension to
PyTorch, available at https://gitlab.com/bsmetsjr/lietorch . Just like for
linear convolution a morphological convolution is specified by a kernel that we
train in our PDE-G-CNNs. In PDE-G-CNNs we do not use non-linearities such as
max/min-pooling and ReLUs as they are already subsumed by morphological
convolutions.
We present a set of experiments to demonstrate the strength of the proposed
PDE-G-CNNs in increasing the performance of deep learning based imaging
applications with far fewer parameters than traditional CNNs.
| [
{
"created": "Fri, 24 Jan 2020 15:00:46 GMT",
"version": "v1"
},
{
"created": "Mon, 9 Mar 2020 14:16:16 GMT",
"version": "v2"
},
{
"created": "Mon, 12 Jul 2021 07:56:22 GMT",
"version": "v3"
},
{
"created": "Sat, 24 Jul 2021 11:14:06 GMT",
"version": "v4"
},
{
"created": "Tue, 26 Apr 2022 10:17:22 GMT",
"version": "v5"
},
{
"created": "Mon, 30 May 2022 19:05:29 GMT",
"version": "v6"
}
] | 2022-08-24 | [
[
"Smets",
"Bart",
""
],
[
"Portegies",
"Jim",
""
],
[
"Bekkers",
"Erik",
""
],
[
"Duits",
"Remco",
""
]
] | We present a PDE-based framework that generalizes Group equivariant Convolutional Neural Networks (G-CNNs). In this framework, a network layer is seen as a set of PDE-solvers where geometrically meaningful PDE-coefficients become the layer's trainable weights. Formulating our PDEs on homogeneous spaces allows these networks to be designed with built-in symmetries such as rotation in addition to the standard translation equivariance of CNNs. Having all the desired symmetries included in the design obviates the need to include them by means of costly techniques such as data augmentation. We will discuss our PDE-based G-CNNs (PDE-G-CNNs) in a general homogeneous space setting while also going into the specifics of our primary case of interest: roto-translation equivariance. We solve the PDE of interest by a combination of linear group convolutions and non-linear morphological group convolutions with analytic kernel approximations that we underpin with formal theorems. Our kernel approximations allow for fast GPU-implementation of the PDE-solvers, we release our implementation with this article in the form of the LieTorch extension to PyTorch, available at https://gitlab.com/bsmetsjr/lietorch . Just like for linear convolution a morphological convolution is specified by a kernel that we train in our PDE-G-CNNs. In PDE-G-CNNs we do not use non-linearities such as max/min-pooling and ReLUs as they are already subsumed by morphological convolutions. We present a set of experiments to demonstrate the strength of the proposed PDE-G-CNNs in increasing the performance of deep learning based imaging applications with far fewer parameters than traditional CNNs. |
1304.0954 | Marko Horvat | Marko Horvat, Anton Grbin, Gordan Gledec | Labeling and Retrieval of Emotionally-Annotated Images using WordNet | 16 pages, 4 figures. arXiv admin note: substantial text overlap with
arXiv:1302.2223 | International Journal of Knowledge-Based and Intelligent
Engineering Systems, Vol. 17, No. 2, pp. 157-166, 2013 | null | null | cs.IR cs.HC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Repositories of images with semantic and emotion content descriptions are
valuable tools in many areas such as Affective Computing and Human-Computer
Interaction, but they are also important in the development of multimodal
searchable online databases. Ever growing number of image documents available
on the Internet continuously motivates research of better annotation models and
more efficient retrieval methods which use mash-up of available data on
semantics, scenes, objects, events, context and emotion. Formal knowledge
representation of such high-level semantics requires rich, explicit, human but
also machine-processable information. To achieve these goals we present an
online ontology-based image annotation tool WNtags and demonstrate its
usefulness in knowledge representation and image retrieval using the
International Affective Picture System database. The WNtags uses WordNet as
image tagging glossary but considers Suggested Upper Merged Ontology as the
preferred upper labeling formalism. The retrieval is performed using node
distance metrics to establish semantic relatedness between a query and the
collaboratively weighted tags describing high-level image semantics, after
which the result is ranked according to the derived importance. We also
elaborate plans to improve the WNtags to create a collaborative Web-based
multimedia repository for research in human emotion and attention.
| [
{
"created": "Wed, 3 Apr 2013 13:58:56 GMT",
"version": "v1"
},
{
"created": "Fri, 10 Jan 2014 23:27:00 GMT",
"version": "v2"
}
] | 2017-12-06 | [
[
"Horvat",
"Marko",
""
],
[
"Grbin",
"Anton",
""
],
[
"Gledec",
"Gordan",
""
]
] | Repositories of images with semantic and emotion content descriptions are valuable tools in many areas such as Affective Computing and Human-Computer Interaction, but they are also important in the development of multimodal searchable online databases. Ever growing number of image documents available on the Internet continuously motivates research of better annotation models and more efficient retrieval methods which use mash-up of available data on semantics, scenes, objects, events, context and emotion. Formal knowledge representation of such high-level semantics requires rich, explicit, human but also machine-processable information. To achieve these goals we present an online ontology-based image annotation tool WNtags and demonstrate its usefulness in knowledge representation and image retrieval using the International Affective Picture System database. The WNtags uses WordNet as image tagging glossary but considers Suggested Upper Merged Ontology as the preferred upper labeling formalism. The retrieval is performed using node distance metrics to establish semantic relatedness between a query and the collaboratively weighted tags describing high-level image semantics, after which the result is ranked according to the derived importance. We also elaborate plans to improve the WNtags to create a collaborative Web-based multimedia repository for research in human emotion and attention. |
2006.14683 | Itzik Malkiel | Itzik Malkiel, Lior Wolf | MTAdam: Automatic Balancing of Multiple Training Loss Terms | null | null | null | null | cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | When training neural models, it is common to combine multiple loss terms. The
balancing of these terms requires considerable human effort and is
computationally demanding. Moreover, the optimal trade-off between the loss
term can change as training progresses, especially for adversarial terms. In
this work, we generalize the Adam optimization algorithm to handle multiple
loss terms. The guiding principle is that for every layer, the gradient
magnitude of the terms should be balanced. To this end, the Multi-Term Adam
(MTAdam) computes the derivative of each loss term separately, infers the first
and second moments per parameter and loss term, and calculates a first moment
for the magnitude per layer of the gradients arising from each loss. This
magnitude is used to continuously balance the gradients across all layers, in a
manner that both varies from one layer to the next and dynamically changes over
time. Our results show that training with the new method leads to fast recovery
from suboptimal initial loss weighting and to training outcomes that match
conventional training with the prescribed hyperparameters of each method.
| [
{
"created": "Thu, 25 Jun 2020 20:27:27 GMT",
"version": "v1"
}
] | 2020-06-29 | [
[
"Malkiel",
"Itzik",
""
],
[
"Wolf",
"Lior",
""
]
] | When training neural models, it is common to combine multiple loss terms. The balancing of these terms requires considerable human effort and is computationally demanding. Moreover, the optimal trade-off between the loss term can change as training progresses, especially for adversarial terms. In this work, we generalize the Adam optimization algorithm to handle multiple loss terms. The guiding principle is that for every layer, the gradient magnitude of the terms should be balanced. To this end, the Multi-Term Adam (MTAdam) computes the derivative of each loss term separately, infers the first and second moments per parameter and loss term, and calculates a first moment for the magnitude per layer of the gradients arising from each loss. This magnitude is used to continuously balance the gradients across all layers, in a manner that both varies from one layer to the next and dynamically changes over time. Our results show that training with the new method leads to fast recovery from suboptimal initial loss weighting and to training outcomes that match conventional training with the prescribed hyperparameters of each method. |
1907.02841 | Li Qiang | Wenxiang Zuo, Qiang Li, Xianming Liu | Depth Restoration: A fast low-rank matrix completion via dual-graph
regularization | The paper will be added more experiments. The main idea of the paper
needs to be revamped. Please withdraw the paper | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | As a real scenes sensing approach, depth information obtains the widespread
applications. However, resulting from the restriction of depth sensing
technology, the depth map captured in practice usually suffers terrible noise
and missing values at plenty of pixels. In this paper, we propose a fast
low-rank matrix completion via dual-graph regularization for depth restoration.
Specifically, the depth restoration can be transformed into a low-rank matrix
completion problem. In order to complete the low-rank matrix and restore it to
the depth map, the proposed dual-graph method containing the local and
non-local graph regularizations exploits the local similarity of depth maps and
the gradient consistency of depth-color counterparts respectively. In addition,
the proposed approach achieves the high speed depth restoration due to
closed-form solution. Experimental results demonstrate that the proposed method
outperforms the state-of-the-art methods with respect to both objective and
subjective quality evaluations, especially for serious depth degeneration.
| [
{
"created": "Fri, 5 Jul 2019 14:09:31 GMT",
"version": "v1"
},
{
"created": "Mon, 28 Oct 2019 11:06:38 GMT",
"version": "v2"
},
{
"created": "Thu, 31 Oct 2019 13:14:36 GMT",
"version": "v3"
},
{
"created": "Wed, 8 Jan 2020 09:29:44 GMT",
"version": "v4"
}
] | 2020-01-09 | [
[
"Zuo",
"Wenxiang",
""
],
[
"Li",
"Qiang",
""
],
[
"Liu",
"Xianming",
""
]
] | As a real scenes sensing approach, depth information obtains the widespread applications. However, resulting from the restriction of depth sensing technology, the depth map captured in practice usually suffers terrible noise and missing values at plenty of pixels. In this paper, we propose a fast low-rank matrix completion via dual-graph regularization for depth restoration. Specifically, the depth restoration can be transformed into a low-rank matrix completion problem. In order to complete the low-rank matrix and restore it to the depth map, the proposed dual-graph method containing the local and non-local graph regularizations exploits the local similarity of depth maps and the gradient consistency of depth-color counterparts respectively. In addition, the proposed approach achieves the high speed depth restoration due to closed-form solution. Experimental results demonstrate that the proposed method outperforms the state-of-the-art methods with respect to both objective and subjective quality evaluations, especially for serious depth degeneration. |
2004.10495 | Dong Wang | Dong Wang, Xiaoqian Qin, Fengyi Song, Li Cheng | Stabilizing Training of Generative Adversarial Nets via Langevin Stein
Variational Gradient Descent | null | null | null | null | cs.LG cs.CV stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Generative adversarial networks (GANs), famous for the capability of learning
complex underlying data distribution, are however known to be tricky in the
training process, which would probably result in mode collapse or performance
deterioration. Current approaches of dealing with GANs' issues almost utilize
some practical training techniques for the purpose of regularization, which on
the other hand undermines the convergence and theoretical soundness of GAN. In
this paper, we propose to stabilize GAN training via a novel particle-based
variational inference -- Langevin Stein variational gradient descent (LSVGD),
which not only inherits the flexibility and efficiency of original SVGD but
aims to address its instability issues by incorporating an extra disturbance
into the update dynamics. We further demonstrate that by properly adjusting the
noise variance, LSVGD simulates a Langevin process whose stationary
distribution is exactly the target distribution. We also show that LSVGD
dynamics has an implicit regularization which is able to enhance particles'
spread-out and diversity. At last we present an efficient way of applying
particle-based variational inference on a general GAN training procedure no
matter what loss function is adopted. Experimental results on one synthetic
dataset and three popular benchmark datasets -- Cifar-10, Tiny-ImageNet and
CelebA validate that LSVGD can remarkably improve the performance and stability
of various GAN models.
| [
{
"created": "Wed, 22 Apr 2020 11:20:04 GMT",
"version": "v1"
}
] | 2020-04-23 | [
[
"Wang",
"Dong",
""
],
[
"Qin",
"Xiaoqian",
""
],
[
"Song",
"Fengyi",
""
],
[
"Cheng",
"Li",
""
]
] | Generative adversarial networks (GANs), famous for the capability of learning complex underlying data distribution, are however known to be tricky in the training process, which would probably result in mode collapse or performance deterioration. Current approaches of dealing with GANs' issues almost utilize some practical training techniques for the purpose of regularization, which on the other hand undermines the convergence and theoretical soundness of GAN. In this paper, we propose to stabilize GAN training via a novel particle-based variational inference -- Langevin Stein variational gradient descent (LSVGD), which not only inherits the flexibility and efficiency of original SVGD but aims to address its instability issues by incorporating an extra disturbance into the update dynamics. We further demonstrate that by properly adjusting the noise variance, LSVGD simulates a Langevin process whose stationary distribution is exactly the target distribution. We also show that LSVGD dynamics has an implicit regularization which is able to enhance particles' spread-out and diversity. At last we present an efficient way of applying particle-based variational inference on a general GAN training procedure no matter what loss function is adopted. Experimental results on one synthetic dataset and three popular benchmark datasets -- Cifar-10, Tiny-ImageNet and CelebA validate that LSVGD can remarkably improve the performance and stability of various GAN models. |
2112.13050 | Susmit Agrawal | K. Ram Prabhakar, Susmit Agrawal, R. Venkatesh Babu | Self-Gated Memory Recurrent Network for Efficient Scalable HDR
Deghosting | 12 pages | IEEE Transactions on Computational Imaging (Volume 7, 2021)
1228-1239 | 10.1109/TCI.2021.3112920 | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | We propose a novel recurrent network-based HDR deghosting method for fusing
arbitrary length dynamic sequences. The proposed method uses convolutional and
recurrent architectures to generate visually pleasing, ghosting-free HDR
images. We introduce a new recurrent cell architecture, namely Self-Gated
Memory (SGM) cell, that outperforms the standard LSTM cell while containing
fewer parameters and having faster running times. In the SGM cell, the
information flow through a gate is controlled by multiplying the gate's output
by a function of itself. Additionally, we use two SGM cells in a bidirectional
setting to improve output quality. The proposed approach achieves
state-of-the-art performance compared to existing HDR deghosting methods
quantitatively across three publicly available datasets while simultaneously
achieving scalability to fuse variable-length input sequence without
necessitating re-training. Through extensive ablations, we demonstrate the
importance of individual components in our proposed approach. The code is
available at https://val.cds.iisc.ac.in/HDR/HDRRNN/index.html.
| [
{
"created": "Fri, 24 Dec 2021 12:36:33 GMT",
"version": "v1"
}
] | 2021-12-28 | [
[
"Prabhakar",
"K. Ram",
""
],
[
"Agrawal",
"Susmit",
""
],
[
"Babu",
"R. Venkatesh",
""
]
] | We propose a novel recurrent network-based HDR deghosting method for fusing arbitrary length dynamic sequences. The proposed method uses convolutional and recurrent architectures to generate visually pleasing, ghosting-free HDR images. We introduce a new recurrent cell architecture, namely Self-Gated Memory (SGM) cell, that outperforms the standard LSTM cell while containing fewer parameters and having faster running times. In the SGM cell, the information flow through a gate is controlled by multiplying the gate's output by a function of itself. Additionally, we use two SGM cells in a bidirectional setting to improve output quality. The proposed approach achieves state-of-the-art performance compared to existing HDR deghosting methods quantitatively across three publicly available datasets while simultaneously achieving scalability to fuse variable-length input sequence without necessitating re-training. Through extensive ablations, we demonstrate the importance of individual components in our proposed approach. The code is available at https://val.cds.iisc.ac.in/HDR/HDRRNN/index.html. |
1904.10522 | Hyunsu Cho | Theodore Vasiloudis, Hyunsu Cho, Henrik Bostr\"om | Block-distributed Gradient Boosted Trees | SIGIR 2019 | null | null | null | cs.LG cs.IR stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The Gradient Boosted Tree (GBT) algorithm is one of the most popular machine
learning algorithms used in production, for tasks that include Click-Through
Rate (CTR) prediction and learning-to-rank. To deal with the massive datasets
available today, many distributed GBT methods have been proposed. However, they
all assume a row-distributed dataset, addressing scalability only with respect
to the number of data points and not the number of features, and increasing
communication cost for high-dimensional data. In order to allow for scalability
across both the data point and feature dimensions, and reduce communication
cost, we propose block-distributed GBTs. We achieve communication efficiency by
making full use of the data sparsity and adapting the Quickscorer algorithm to
the block-distributed setting. We evaluate our approach using datasets with
millions of features, and demonstrate that we are able to achieve multiple
orders of magnitude reduction in communication cost for sparse data, with no
loss in accuracy, while providing a more scalable design. As a result, we are
able to reduce the training time for high-dimensional data, and allow more
cost-effective scale-out without the need for expensive network communication.
| [
{
"created": "Tue, 23 Apr 2019 20:10:36 GMT",
"version": "v1"
},
{
"created": "Tue, 28 May 2019 19:32:35 GMT",
"version": "v2"
}
] | 2019-05-30 | [
[
"Vasiloudis",
"Theodore",
""
],
[
"Cho",
"Hyunsu",
""
],
[
"Boström",
"Henrik",
""
]
] | The Gradient Boosted Tree (GBT) algorithm is one of the most popular machine learning algorithms used in production, for tasks that include Click-Through Rate (CTR) prediction and learning-to-rank. To deal with the massive datasets available today, many distributed GBT methods have been proposed. However, they all assume a row-distributed dataset, addressing scalability only with respect to the number of data points and not the number of features, and increasing communication cost for high-dimensional data. In order to allow for scalability across both the data point and feature dimensions, and reduce communication cost, we propose block-distributed GBTs. We achieve communication efficiency by making full use of the data sparsity and adapting the Quickscorer algorithm to the block-distributed setting. We evaluate our approach using datasets with millions of features, and demonstrate that we are able to achieve multiple orders of magnitude reduction in communication cost for sparse data, with no loss in accuracy, while providing a more scalable design. As a result, we are able to reduce the training time for high-dimensional data, and allow more cost-effective scale-out without the need for expensive network communication. |
2306.09547 | Daria Reshetova | Daria Reshetova, Wei-Ning Chen, Ayfer \"Ozg\"ur | Training generative models from privatized data | null | null | null | null | cs.LG cs.CR cs.IT math.IT | http://creativecommons.org/licenses/by/4.0/ | Local differential privacy is a powerful method for privacy-preserving data
collection. In this paper, we develop a framework for training Generative
Adversarial Networks (GANs) on differentially privatized data. We show that
entropic regularization of optimal transport - a popular regularization method
in the literature that has often been leveraged for its computational benefits
- enables the generator to learn the raw (unprivatized) data distribution even
though it only has access to privatized samples. We prove that at the same time
this leads to fast statistical convergence at the parametric rate. This shows
that entropic regularization of optimal transport uniquely enables the
mitigation of both the effects of privatization noise and the curse of
dimensionality in statistical convergence. We provide experimental evidence to
support the efficacy of our framework in practice.
| [
{
"created": "Thu, 15 Jun 2023 23:28:45 GMT",
"version": "v1"
},
{
"created": "Fri, 1 Mar 2024 01:54:15 GMT",
"version": "v2"
}
] | 2024-03-04 | [
[
"Reshetova",
"Daria",
""
],
[
"Chen",
"Wei-Ning",
""
],
[
"Özgür",
"Ayfer",
""
]
] | Local differential privacy is a powerful method for privacy-preserving data collection. In this paper, we develop a framework for training Generative Adversarial Networks (GANs) on differentially privatized data. We show that entropic regularization of optimal transport - a popular regularization method in the literature that has often been leveraged for its computational benefits - enables the generator to learn the raw (unprivatized) data distribution even though it only has access to privatized samples. We prove that at the same time this leads to fast statistical convergence at the parametric rate. This shows that entropic regularization of optimal transport uniquely enables the mitigation of both the effects of privatization noise and the curse of dimensionality in statistical convergence. We provide experimental evidence to support the efficacy of our framework in practice. |
1809.01301 | Pamela Shapiro | Pamela Shapiro and Kevin Duh | BPE and CharCNNs for Translation of Morphology: A Cross-Lingual
Comparison and Analysis | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Neural Machine Translation (NMT) in low-resource settings and of
morphologically rich languages is made difficult in part by data sparsity of
vocabulary words. Several methods have been used to help reduce this sparsity,
notably Byte-Pair Encoding (BPE) and a character-based CNN layer (charCNN).
However, the charCNN has largely been neglected, possibly because it has only
been compared to BPE rather than combined with it. We argue for a
reconsideration of the charCNN, based on cross-lingual improvements on
low-resource data. We translate from 8 languages into English, using a
multi-way parallel collection of TED transcripts. We find that in most cases,
using both BPE and a charCNN performs best, while in Hebrew, using a charCNN
over words is best.
| [
{
"created": "Wed, 5 Sep 2018 02:26:09 GMT",
"version": "v1"
},
{
"created": "Sat, 8 Sep 2018 23:36:53 GMT",
"version": "v2"
}
] | 2018-09-11 | [
[
"Shapiro",
"Pamela",
""
],
[
"Duh",
"Kevin",
""
]
] | Neural Machine Translation (NMT) in low-resource settings and of morphologically rich languages is made difficult in part by data sparsity of vocabulary words. Several methods have been used to help reduce this sparsity, notably Byte-Pair Encoding (BPE) and a character-based CNN layer (charCNN). However, the charCNN has largely been neglected, possibly because it has only been compared to BPE rather than combined with it. We argue for a reconsideration of the charCNN, based on cross-lingual improvements on low-resource data. We translate from 8 languages into English, using a multi-way parallel collection of TED transcripts. We find that in most cases, using both BPE and a charCNN performs best, while in Hebrew, using a charCNN over words is best. |
cs/0109012 | Michael Geist | Michael Geist | Is There a There There: Towards Greater Certainty for Internet
Jurisdiction | 29th TPRC Conference, 2001 | 16 (3) Berkeley Tech. LJ (forthcoming 2001) | null | TPRC-2001-017 | cs.CY | null | The unique challenge presented by the Internet is that compliance with local
laws is rarely sufficient to assure a business that it has limited its exposure
to legal risk. The paper identifies why the challenge of adequately accounting
for the legal risk arising from Internet jurisdiction has been aggravated in
recent years by the adoption of the Zippo legal framework, commonly referred to
as the passive versus active test. The test provides parties with only limited
guidance and often results in detrimental judicial decisions from a policy
perspective. Given the inadequacies of the Zippo passive versus active test,
the paper argues that it is now fitting to identify a more effective standard
for determining when it is appropriate to assert jurisdiction in cases
involving predominantly Internet-based contacts. The solution submitted in the
paper is to move toward a targeting-based analysis. Unlike the Zippo approach,
a targeting analysis would seek to identify the intentions of the parties and
to assess the steps taken to either enter or avoid a particular jurisdiction.
Targeting would also lessen the reliance on effects-based analysis, the source
of considerable uncertainty since Internet-based activity can ordinarily be
said to create some effects in most jurisdictions. To identify the appropriate
criteria for a targeting test, the paper recommends returning to the core
jurisdictional principle -- foreseeability. Foreseeability in the targeting
context depends on three factors -- contracts, technology, and actual or
implied knowledge.
| [
{
"created": "Tue, 11 Sep 2001 03:22:25 GMT",
"version": "v1"
}
] | 2007-05-23 | [
[
"Geist",
"Michael",
""
]
] | The unique challenge presented by the Internet is that compliance with local laws is rarely sufficient to assure a business that it has limited its exposure to legal risk. The paper identifies why the challenge of adequately accounting for the legal risk arising from Internet jurisdiction has been aggravated in recent years by the adoption of the Zippo legal framework, commonly referred to as the passive versus active test. The test provides parties with only limited guidance and often results in detrimental judicial decisions from a policy perspective. Given the inadequacies of the Zippo passive versus active test, the paper argues that it is now fitting to identify a more effective standard for determining when it is appropriate to assert jurisdiction in cases involving predominantly Internet-based contacts. The solution submitted in the paper is to move toward a targeting-based analysis. Unlike the Zippo approach, a targeting analysis would seek to identify the intentions of the parties and to assess the steps taken to either enter or avoid a particular jurisdiction. Targeting would also lessen the reliance on effects-based analysis, the source of considerable uncertainty since Internet-based activity can ordinarily be said to create some effects in most jurisdictions. To identify the appropriate criteria for a targeting test, the paper recommends returning to the core jurisdictional principle -- foreseeability. Foreseeability in the targeting context depends on three factors -- contracts, technology, and actual or implied knowledge. |
2303.06611 | Weilin Lin | Weilin Lin, Xiangyu Zhao, Yejing Wang, Yuanshao Zhu, Wanyu Wang | AutoDenoise: Automatic Data Instance Denoising for Recommendations | 9 pages, 4 figures, 5 tables, conference | null | 10.1145/3543507.3583339 | null | cs.IR | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Historical user-item interaction datasets are essential in training modern
recommender systems for predicting user preferences. However, the arbitrary
user behaviors in most recommendation scenarios lead to a large volume of noisy
data instances being recorded, which cannot fully represent their true
interests. While a large number of denoising studies are emerging in the
recommender system community, all of them suffer from highly dynamic data
distributions. In this paper, we propose a Deep Reinforcement Learning (DRL)
based framework, AutoDenoise, with an Instance Denoising Policy Network, for
denoising data instances with an instance selection manner in deep recommender
systems. To be specific, AutoDenoise serves as an agent in DRL to adaptively
select noise-free and predictive data instances, which can then be utilized
directly in training representative recommendation models. In addition, we
design an alternate two-phase optimization strategy to train and validate the
AutoDenoise properly. In the searching phase, we aim to train the policy
network with the capacity of instance denoising; in the validation phase, we
find out and evaluate the denoised subset of data instances selected by the
trained policy network, so as to validate its denoising ability. We conduct
extensive experiments to validate the effectiveness of AutoDenoise combined
with multiple representative recommender system models.
| [
{
"created": "Sun, 12 Mar 2023 08:36:15 GMT",
"version": "v1"
}
] | 2023-03-14 | [
[
"Lin",
"Weilin",
""
],
[
"Zhao",
"Xiangyu",
""
],
[
"Wang",
"Yejing",
""
],
[
"Zhu",
"Yuanshao",
""
],
[
"Wang",
"Wanyu",
""
]
] | Historical user-item interaction datasets are essential in training modern recommender systems for predicting user preferences. However, the arbitrary user behaviors in most recommendation scenarios lead to a large volume of noisy data instances being recorded, which cannot fully represent their true interests. While a large number of denoising studies are emerging in the recommender system community, all of them suffer from highly dynamic data distributions. In this paper, we propose a Deep Reinforcement Learning (DRL) based framework, AutoDenoise, with an Instance Denoising Policy Network, for denoising data instances with an instance selection manner in deep recommender systems. To be specific, AutoDenoise serves as an agent in DRL to adaptively select noise-free and predictive data instances, which can then be utilized directly in training representative recommendation models. In addition, we design an alternate two-phase optimization strategy to train and validate the AutoDenoise properly. In the searching phase, we aim to train the policy network with the capacity of instance denoising; in the validation phase, we find out and evaluate the denoised subset of data instances selected by the trained policy network, so as to validate its denoising ability. We conduct extensive experiments to validate the effectiveness of AutoDenoise combined with multiple representative recommender system models. |
2301.07849 | Giovanni Viglietta | Giuseppe A. Di Luna and Giovanni Viglietta | Efficient Computation in Congested Anonymous Dynamic Networks | 26 pages, 2 figures | null | null | null | cs.DC cs.DM | http://creativecommons.org/licenses/by/4.0/ | An anonymous dynamic network is a network of indistinguishable processes
whose communication links may appear or disappear unpredictably over time.
Previous research has shown that deterministically computing an arbitrary
function of a multiset of input values given to these processes takes only a
linear number of communication rounds (Di Luna-Viglietta, FOCS 2022).
However, fast algorithms for anonymous dynamic networks rely on the
construction and transmission of large data structures called "history trees",
whose size is polynomial in the number of processes. This approach is
unfeasible if the network is congested, and only messages of logarithmic size
can be sent through its links. Observe that sending a large message piece by
piece over several rounds is not in itself a solution, due to the anonymity of
the processes combined with the dynamic nature of the network. Moreover, it is
known that certain basic tasks such as all-to-all token dissemination (by means
of single-token forwarding) require $\Omega(n^2/\log n)$ rounds in congested
networks (Dutta et al., SODA 2013).
In this work, we develop a series of practical and efficient techniques that
make it possible to use history trees in congested anonymous dynamic networks.
Among other applications, we show how to compute arbitrary functions in such
networks in $O(n^3)$ communication rounds, greatly improving upon previous
state-of-the-art algorithms for congested networks.
| [
{
"created": "Thu, 19 Jan 2023 02:11:47 GMT",
"version": "v1"
},
{
"created": "Sat, 6 May 2023 15:22:15 GMT",
"version": "v2"
},
{
"created": "Tue, 5 Sep 2023 03:03:07 GMT",
"version": "v3"
},
{
"created": "Sat, 29 Jun 2024 12:53:12 GMT",
"version": "v4"
}
] | 2024-07-02 | [
[
"Di Luna",
"Giuseppe A.",
""
],
[
"Viglietta",
"Giovanni",
""
]
] | An anonymous dynamic network is a network of indistinguishable processes whose communication links may appear or disappear unpredictably over time. Previous research has shown that deterministically computing an arbitrary function of a multiset of input values given to these processes takes only a linear number of communication rounds (Di Luna-Viglietta, FOCS 2022). However, fast algorithms for anonymous dynamic networks rely on the construction and transmission of large data structures called "history trees", whose size is polynomial in the number of processes. This approach is unfeasible if the network is congested, and only messages of logarithmic size can be sent through its links. Observe that sending a large message piece by piece over several rounds is not in itself a solution, due to the anonymity of the processes combined with the dynamic nature of the network. Moreover, it is known that certain basic tasks such as all-to-all token dissemination (by means of single-token forwarding) require $\Omega(n^2/\log n)$ rounds in congested networks (Dutta et al., SODA 2013). In this work, we develop a series of practical and efficient techniques that make it possible to use history trees in congested anonymous dynamic networks. Among other applications, we show how to compute arbitrary functions in such networks in $O(n^3)$ communication rounds, greatly improving upon previous state-of-the-art algorithms for congested networks. |
2407.18813 | Chenming Wu | Zhe Xin and Yufeng Yue and Liangjun Zhang and Chenming Wu | HERO-SLAM: Hybrid Enhanced Robust Optimization of Neural SLAM | Accepted to ICRA 2024 | null | null | null | cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Simultaneous Localization and Mapping (SLAM) is a fundamental task in
robotics, driving numerous applications such as autonomous driving and virtual
reality. Recent progress on neural implicit SLAM has shown encouraging and
impressive results. However, the robustness of neural SLAM, particularly in
challenging or data-limited situations, remains an unresolved issue. This paper
presents HERO-SLAM, a Hybrid Enhanced Robust Optimization method for neural
SLAM, which combines the benefits of neural implicit field and feature-metric
optimization. This hybrid method optimizes a multi-resolution implicit field
and enhances robustness in challenging environments with sudden viewpoint
changes or sparse data collection. Our comprehensive experimental results on
benchmarking datasets validate the effectiveness of our hybrid approach,
demonstrating its superior performance over existing implicit field-based
methods in challenging scenarios. HERO-SLAM provides a new pathway to enhance
the stability, performance, and applicability of neural SLAM in real-world
scenarios. Code is available on the project page: https://hero-slam.github.io.
| [
{
"created": "Fri, 26 Jul 2024 15:22:14 GMT",
"version": "v1"
}
] | 2024-07-29 | [
[
"Xin",
"Zhe",
""
],
[
"Yue",
"Yufeng",
""
],
[
"Zhang",
"Liangjun",
""
],
[
"Wu",
"Chenming",
""
]
] | Simultaneous Localization and Mapping (SLAM) is a fundamental task in robotics, driving numerous applications such as autonomous driving and virtual reality. Recent progress on neural implicit SLAM has shown encouraging and impressive results. However, the robustness of neural SLAM, particularly in challenging or data-limited situations, remains an unresolved issue. This paper presents HERO-SLAM, a Hybrid Enhanced Robust Optimization method for neural SLAM, which combines the benefits of neural implicit field and feature-metric optimization. This hybrid method optimizes a multi-resolution implicit field and enhances robustness in challenging environments with sudden viewpoint changes or sparse data collection. Our comprehensive experimental results on benchmarking datasets validate the effectiveness of our hybrid approach, demonstrating its superior performance over existing implicit field-based methods in challenging scenarios. HERO-SLAM provides a new pathway to enhance the stability, performance, and applicability of neural SLAM in real-world scenarios. Code is available on the project page: https://hero-slam.github.io. |
cs/0702083 | Serebrenik Alexander | Alexander Serebrenik, Tom Schrijvers, Bart Demoen | Improving Prolog programs: Refactoring for Prolog | To appear in Theory and Practice of Logic Programming (TPLP) | null | null | 2006-1 | cs.SE | null | Refactoring is an established technique from the object-oriented (OO)
programming community to restructure code: it aims at improving software
readability, maintainability and extensibility. Although refactoring is not
tied to the OO-paradigm in particular, its ideas have not been applied to Logic
Programming until now.
This paper applies the ideas of refactoring to Prolog programs. A catalogue
is presented listing refactorings classified according to scope. Some of the
refactorings have been adapted from the OO-paradigm, while others have been
specifically designed for Prolog. The discrepancy between intended and
operational semantics in Prolog is also addressed by some of the refactorings.
In addition, ViPReSS, a semi-automatic refactoring browser, is discussed and
the experience with applying ViPReSS to a large Prolog legacy system is
reported. The main conclusion is that refactoring is both a viable technique in
Prolog and a rather desirable one.
| [
{
"created": "Wed, 14 Feb 2007 09:53:37 GMT",
"version": "v1"
}
] | 2007-05-23 | [
[
"Serebrenik",
"Alexander",
""
],
[
"Schrijvers",
"Tom",
""
],
[
"Demoen",
"Bart",
""
]
] | Refactoring is an established technique from the object-oriented (OO) programming community to restructure code: it aims at improving software readability, maintainability and extensibility. Although refactoring is not tied to the OO-paradigm in particular, its ideas have not been applied to Logic Programming until now. This paper applies the ideas of refactoring to Prolog programs. A catalogue is presented listing refactorings classified according to scope. Some of the refactorings have been adapted from the OO-paradigm, while others have been specifically designed for Prolog. The discrepancy between intended and operational semantics in Prolog is also addressed by some of the refactorings. In addition, ViPReSS, a semi-automatic refactoring browser, is discussed and the experience with applying ViPReSS to a large Prolog legacy system is reported. The main conclusion is that refactoring is both a viable technique in Prolog and a rather desirable one. |
2306.01116 | Julien Launay | Guilherme Penedo, Quentin Malartic, Daniel Hesslow, Ruxandra Cojocaru,
Alessandro Cappelli, Hamza Alobeidli, Baptiste Pannier, Ebtesam Almazrouei,
Julien Launay | The RefinedWeb Dataset for Falcon LLM: Outperforming Curated Corpora
with Web Data, and Web Data Only | null | null | null | null | cs.CL cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large language models are commonly trained on a mixture of filtered web data
and curated high-quality corpora, such as social media conversations, books, or
technical papers. This curation process is believed to be necessary to produce
performant models with broad zero-shot generalization abilities. However, as
larger models requiring pretraining on trillions of tokens are considered, it
is unclear how scalable is curation and whether we will run out of unique
high-quality data soon. At variance with previous beliefs, we show that
properly filtered and deduplicated web data alone can lead to powerful models;
even significantly outperforming models from the state-of-the-art trained on
The Pile. Despite extensive filtering, the high-quality data we extract from
the web is still plentiful, and we are able to obtain five trillion tokens from
CommonCrawl. We publicly release an extract of 600 billion tokens from our
RefinedWeb dataset, and 1.3/7.5B parameters language models trained on it.
| [
{
"created": "Thu, 1 Jun 2023 20:03:56 GMT",
"version": "v1"
}
] | 2023-06-05 | [
[
"Penedo",
"Guilherme",
""
],
[
"Malartic",
"Quentin",
""
],
[
"Hesslow",
"Daniel",
""
],
[
"Cojocaru",
"Ruxandra",
""
],
[
"Cappelli",
"Alessandro",
""
],
[
"Alobeidli",
"Hamza",
""
],
[
"Pannier",
"Baptiste",
""
],
[
"Almazrouei",
"Ebtesam",
""
],
[
"Launay",
"Julien",
""
]
] | Large language models are commonly trained on a mixture of filtered web data and curated high-quality corpora, such as social media conversations, books, or technical papers. This curation process is believed to be necessary to produce performant models with broad zero-shot generalization abilities. However, as larger models requiring pretraining on trillions of tokens are considered, it is unclear how scalable is curation and whether we will run out of unique high-quality data soon. At variance with previous beliefs, we show that properly filtered and deduplicated web data alone can lead to powerful models; even significantly outperforming models from the state-of-the-art trained on The Pile. Despite extensive filtering, the high-quality data we extract from the web is still plentiful, and we are able to obtain five trillion tokens from CommonCrawl. We publicly release an extract of 600 billion tokens from our RefinedWeb dataset, and 1.3/7.5B parameters language models trained on it. |
1701.00416 | Tom Mens | Alexandre Decan, Mathieu Goeminne, Tom Mens | On the Interaction of Relational Database Access Technologies in Open
Source Java Projects | Postproceeding of the SATTOSE 2015 Research Seminar on Advanced Tools
and Techniques for Software Evolution. To be published in CEUR.WS workshop
proceedings (2017) | null | null | null | cs.SE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This article presents an empirical study of how the use of relational
database access technologies in open source Java projects evolves over time.
Our observations may be useful to project managers to make more informed
decisions on which technologies to introduce into an existing project and when.
We selected 2,457 Java projects on GitHub using the low-level JDBC technology
and higher-level object relational mappings such as Hibernate XML configuration
files and JPA annotations. At a coarse-grained level, we analysed the
probability of introducing such technologies over time, as well as the
likelihood that multiple technologies co-occur within the same project. At a
fine-grained level, we analysed to which extent these different technologies
are used within the same set of project files. We also explored how the
introduction of a new database technology in a Java project impacts the use of
existing ones. We observed that, contrary to what could have been expected,
object-relational mapping technologies do not tend to replace existing ones but
rather complement them.
| [
{
"created": "Mon, 2 Jan 2017 15:07:36 GMT",
"version": "v1"
}
] | 2017-01-03 | [
[
"Decan",
"Alexandre",
""
],
[
"Goeminne",
"Mathieu",
""
],
[
"Mens",
"Tom",
""
]
] | This article presents an empirical study of how the use of relational database access technologies in open source Java projects evolves over time. Our observations may be useful to project managers to make more informed decisions on which technologies to introduce into an existing project and when. We selected 2,457 Java projects on GitHub using the low-level JDBC technology and higher-level object relational mappings such as Hibernate XML configuration files and JPA annotations. At a coarse-grained level, we analysed the probability of introducing such technologies over time, as well as the likelihood that multiple technologies co-occur within the same project. At a fine-grained level, we analysed to which extent these different technologies are used within the same set of project files. We also explored how the introduction of a new database technology in a Java project impacts the use of existing ones. We observed that, contrary to what could have been expected, object-relational mapping technologies do not tend to replace existing ones but rather complement them. |
1401.3556 | Sergiy Vorobyov A. | Alex E. Geyer, Reza Nikjah, Sergiy A. Vorobyov, and Norman C. Beaulieu | Equivalent Codes, Optimality, and Performance Analysis of OSTBC:
Textbook Study | 33 pages, 12 figures, 5 tables, full size journal paper, Finished in
Oct. 2009, Unpublished | IEEE Trans. Communications, vol. 63, no. 8, pp. 2912-2923, Aug.
2015 | null | null | cs.IT math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | An equivalent model for a multi-input multi-output (MIMO) communication
system with orthogonal space-time block codes (OSTBCs) is proposed based on a
newly revealed connection between OSTBCs and Euclidean codes. Examples of
distance spectra, signal constellations, and signal coordinate diagrams of
Euclidean codes equivalent to simplest OSTBCs are given. A new asymptotic upper
bound for the symbol error rate (SER) of OSTBCs, based on the distance spectra
of the introduced equivalent Euclidean codes is derived, and new general design
criteria for signal constellations of the optimal OSTBC are proposed. Some
bounds relating distance properties, dimensionality, and cardinality of OSTBCs
with constituent signals of equal energy are given, and new optimal signal
constellations with cardinalities M = 8 and M = 16 for Alamouti's code are
designed. Using the new model for MIMO communication systems with OSTBCs, a
general methodology for performance analysis of OSTBCs is developed. As an
example of the application of this methodology, an exact evaluation of the SER
of any OSTBC is given. Namely, a new expression for the SER of Alamouti's OSTBC
with binary phase shift keying (BPSK) signals is derived.
| [
{
"created": "Wed, 15 Jan 2014 12:07:56 GMT",
"version": "v1"
}
] | 2016-03-03 | [
[
"Geyer",
"Alex E.",
""
],
[
"Nikjah",
"Reza",
""
],
[
"Vorobyov",
"Sergiy A.",
""
],
[
"Beaulieu",
"Norman C.",
""
]
] | An equivalent model for a multi-input multi-output (MIMO) communication system with orthogonal space-time block codes (OSTBCs) is proposed based on a newly revealed connection between OSTBCs and Euclidean codes. Examples of distance spectra, signal constellations, and signal coordinate diagrams of Euclidean codes equivalent to simplest OSTBCs are given. A new asymptotic upper bound for the symbol error rate (SER) of OSTBCs, based on the distance spectra of the introduced equivalent Euclidean codes is derived, and new general design criteria for signal constellations of the optimal OSTBC are proposed. Some bounds relating distance properties, dimensionality, and cardinality of OSTBCs with constituent signals of equal energy are given, and new optimal signal constellations with cardinalities M = 8 and M = 16 for Alamouti's code are designed. Using the new model for MIMO communication systems with OSTBCs, a general methodology for performance analysis of OSTBCs is developed. As an example of the application of this methodology, an exact evaluation of the SER of any OSTBC is given. Namely, a new expression for the SER of Alamouti's OSTBC with binary phase shift keying (BPSK) signals is derived. |
1910.05268 | Asier Mujika | Florian Meier and Asier Mujika and Marcelo Matheus Gauy and Angelika
Steger | Improving Gradient Estimation in Evolutionary Strategies With Past
Descent Directions | null | null | null | null | cs.NE cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Evolutionary Strategies (ES) are known to be an effective black-box
optimization technique for deep neural networks when the true gradients cannot
be computed, such as in Reinforcement Learning. We continue a recent line of
research that uses surrogate gradients to improve the gradient estimation of
ES. We propose a novel method to optimally incorporate surrogate gradient
information. Our approach, unlike previous work, needs no information about the
quality of the surrogate gradients and is always guaranteed to find a descent
direction that is better than the surrogate gradient. This allows to
iteratively use the previous gradient estimate as surrogate gradient for the
current search point. We theoretically prove that this yields fast convergence
to the true gradient for linear functions and show under simplifying
assumptions that it significantly improves gradient estimates for general
functions. Finally, we evaluate our approach empirically on MNIST and
reinforcement learning tasks and show that it considerably improves the
gradient estimation of ES at no extra computational cost.
| [
{
"created": "Fri, 11 Oct 2019 16:00:39 GMT",
"version": "v1"
}
] | 2019-10-14 | [
[
"Meier",
"Florian",
""
],
[
"Mujika",
"Asier",
""
],
[
"Gauy",
"Marcelo Matheus",
""
],
[
"Steger",
"Angelika",
""
]
] | Evolutionary Strategies (ES) are known to be an effective black-box optimization technique for deep neural networks when the true gradients cannot be computed, such as in Reinforcement Learning. We continue a recent line of research that uses surrogate gradients to improve the gradient estimation of ES. We propose a novel method to optimally incorporate surrogate gradient information. Our approach, unlike previous work, needs no information about the quality of the surrogate gradients and is always guaranteed to find a descent direction that is better than the surrogate gradient. This allows to iteratively use the previous gradient estimate as surrogate gradient for the current search point. We theoretically prove that this yields fast convergence to the true gradient for linear functions and show under simplifying assumptions that it significantly improves gradient estimates for general functions. Finally, we evaluate our approach empirically on MNIST and reinforcement learning tasks and show that it considerably improves the gradient estimation of ES at no extra computational cost. |
2204.03479 | Zuzana Jel\v{c}icov\'a | Zuzana Jel\v{c}icov\'a and Marian Verhelst | Delta Keyword Transformer: Bringing Transformers to the Edge through
Dynamically Pruned Multi-Head Self-Attention | null | null | null | null | cs.CL cs.LG | http://creativecommons.org/licenses/by-sa/4.0/ | Multi-head self-attention forms the core of Transformer networks. However,
their quadratically growing complexity with respect to the input sequence
length impedes their deployment on resource-constrained edge devices. We
address this challenge by proposing a dynamic pruning method, which exploits
the temporal stability of data across tokens to reduce inference cost. The
threshold-based method only retains significant differences between the
subsequent tokens, effectively reducing the number of multiply-accumulates, as
well as the internal tensor data sizes. The approach is evaluated on the Google
Speech Commands Dataset for keyword spotting, and the performance is compared
against the baseline Keyword Transformer. Our experiments show that we can
reduce ~80% of operations while maintaining the original 98.4% accuracy.
Moreover, a reduction of ~87-94% operations can be achieved when only degrading
the accuracy by 1-4%, speeding up the multi-head self-attention inference by a
factor of ~7.5-16.
| [
{
"created": "Sun, 20 Mar 2022 20:59:13 GMT",
"version": "v1"
}
] | 2022-04-08 | [
[
"Jelčicová",
"Zuzana",
""
],
[
"Verhelst",
"Marian",
""
]
] | Multi-head self-attention forms the core of Transformer networks. However, their quadratically growing complexity with respect to the input sequence length impedes their deployment on resource-constrained edge devices. We address this challenge by proposing a dynamic pruning method, which exploits the temporal stability of data across tokens to reduce inference cost. The threshold-based method only retains significant differences between the subsequent tokens, effectively reducing the number of multiply-accumulates, as well as the internal tensor data sizes. The approach is evaluated on the Google Speech Commands Dataset for keyword spotting, and the performance is compared against the baseline Keyword Transformer. Our experiments show that we can reduce ~80% of operations while maintaining the original 98.4% accuracy. Moreover, a reduction of ~87-94% operations can be achieved when only degrading the accuracy by 1-4%, speeding up the multi-head self-attention inference by a factor of ~7.5-16. |
2404.03081 | Hans De Sterck | Yifan Qu, Oliver Krzysik, Hans De Sterck, Omer Ege Kara | First-order PDES for Graph Neural Networks: Advection And Burgers
Equation Models | null | null | null | null | cs.LG cs.NA math.NA | http://creativecommons.org/licenses/by/4.0/ | Graph Neural Networks (GNNs) have established themselves as the preferred
methodology in a multitude of domains, ranging from computer vision to
computational biology, especially in contexts where data inherently conform to
graph structures. While many existing methods have endeavored to model GNNs
using various techniques, a prevalent challenge they grapple with is the issue
of over-smoothing. This paper presents new Graph Neural Network models that
incorporate two first-order Partial Differential Equations (PDEs). These models
do not increase complexity but effectively mitigate the over-smoothing problem.
Our experimental findings highlight the capacity of our new PDE model to
achieve comparable results with higher-order PDE models and fix the
over-smoothing problem up to 64 layers. These results underscore the
adaptability and versatility of GNNs, indicating that unconventional approaches
can yield outcomes on par with established techniques.
| [
{
"created": "Wed, 3 Apr 2024 21:47:02 GMT",
"version": "v1"
}
] | 2024-04-05 | [
[
"Qu",
"Yifan",
""
],
[
"Krzysik",
"Oliver",
""
],
[
"De Sterck",
"Hans",
""
],
[
"Kara",
"Omer Ege",
""
]
] | Graph Neural Networks (GNNs) have established themselves as the preferred methodology in a multitude of domains, ranging from computer vision to computational biology, especially in contexts where data inherently conform to graph structures. While many existing methods have endeavored to model GNNs using various techniques, a prevalent challenge they grapple with is the issue of over-smoothing. This paper presents new Graph Neural Network models that incorporate two first-order Partial Differential Equations (PDEs). These models do not increase complexity but effectively mitigate the over-smoothing problem. Our experimental findings highlight the capacity of our new PDE model to achieve comparable results with higher-order PDE models and fix the over-smoothing problem up to 64 layers. These results underscore the adaptability and versatility of GNNs, indicating that unconventional approaches can yield outcomes on par with established techniques. |
1705.02038 | Jiazi Zhang | Jiazi Zhang and Zhigang Chu and Lalitha Sankar and Oliver Kosut | False Data Injection Attacks on Phasor Measurements That Bypass Low-rank
Decomposition | 6 pages, 4 figures, submitted to 2017 IEEE International Conference
on Smart Grid Communications (SmartGridComm) | null | null | null | cs.SY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper studies the vulnerability of phasor measurement units (PMUs) to
false data injection (FDI) attacks. Prior work demonstrated that unobservable
FDI attacks that can bypass traditional bad data detectors based on measurement
residuals can be identified by detector based on low-rank decomposition (LD).
In this work, a class of more sophisticated FDI attacks that captures the
temporal correlation of PMU data is introduced. Such attacks are designed with
a convex optimization problem and can always bypass the LD detector. The
vulnerability of this attack model is illustrated on both the IEEE 24-bus RTS
and the IEEE 118-bus systems.
| [
{
"created": "Thu, 4 May 2017 22:33:04 GMT",
"version": "v1"
}
] | 2017-05-08 | [
[
"Zhang",
"Jiazi",
""
],
[
"Chu",
"Zhigang",
""
],
[
"Sankar",
"Lalitha",
""
],
[
"Kosut",
"Oliver",
""
]
] | This paper studies the vulnerability of phasor measurement units (PMUs) to false data injection (FDI) attacks. Prior work demonstrated that unobservable FDI attacks that can bypass traditional bad data detectors based on measurement residuals can be identified by detector based on low-rank decomposition (LD). In this work, a class of more sophisticated FDI attacks that captures the temporal correlation of PMU data is introduced. Such attacks are designed with a convex optimization problem and can always bypass the LD detector. The vulnerability of this attack model is illustrated on both the IEEE 24-bus RTS and the IEEE 118-bus systems. |
1907.12430 | Alexander V Terekhov | Alexander V. Terekhov and J. Kevin O'Regan | Learning abstract perceptual notions: the example of space | arXiv admin note: text overlap with arXiv:1308.2124 | null | null | null | cs.AI q-bio.NC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Humans are extremely swift learners. We are able to grasp highly abstract
notions, whether they come from art perception or pure mathematics. Current
machine learning techniques demonstrate astonishing results in extracting
patterns in information. Yet the abstract notions we possess are more than just
statistical patterns in the incoming information. Sensorimotor theory suggests
that they represent functions, laws, describing how the information can be
transformed, or, in other words, they represent the statistics of sensorimotor
changes rather than sensory inputs themselves. The aim of our work is to
suggest a way for machine learning and sensorimotor theory to benefit from each
other so as to pave the way toward new horizons in learning. We show in this
study that a highly abstract notion, that of space, can be seen as a collection
of laws of transformations of sensory information and that these laws could in
theory be learned by a naive agent. As an illustration we do a one-dimensional
simulation in which an agent extracts spatial knowledge in the form of
internalized ("sensible") rigid displacements. The agent uses them to encode
its own displacements in a way which is isometrically related to external
space. Though the algorithm allowing acquisition of rigid displacements is
designed \emph{ad hoc}, we believe it can stimulate the development of
unsupervised learning techniques leading to similar results.
| [
{
"created": "Wed, 24 Jul 2019 17:57:54 GMT",
"version": "v1"
}
] | 2019-07-30 | [
[
"Terekhov",
"Alexander V.",
""
],
[
"O'Regan",
"J. Kevin",
""
]
] | Humans are extremely swift learners. We are able to grasp highly abstract notions, whether they come from art perception or pure mathematics. Current machine learning techniques demonstrate astonishing results in extracting patterns in information. Yet the abstract notions we possess are more than just statistical patterns in the incoming information. Sensorimotor theory suggests that they represent functions, laws, describing how the information can be transformed, or, in other words, they represent the statistics of sensorimotor changes rather than sensory inputs themselves. The aim of our work is to suggest a way for machine learning and sensorimotor theory to benefit from each other so as to pave the way toward new horizons in learning. We show in this study that a highly abstract notion, that of space, can be seen as a collection of laws of transformations of sensory information and that these laws could in theory be learned by a naive agent. As an illustration we do a one-dimensional simulation in which an agent extracts spatial knowledge in the form of internalized ("sensible") rigid displacements. The agent uses them to encode its own displacements in a way which is isometrically related to external space. Though the algorithm allowing acquisition of rigid displacements is designed \emph{ad hoc}, we believe it can stimulate the development of unsupervised learning techniques leading to similar results. |
2203.05735 | Quoc Nguyen | Huu-Quoc Nguyen, Tien-Dung Nguyen, Van-Nam Pham, Xuan-Qui Pham,
Quang-Thai Ngo, Eui-Nam Huh | An Efficient Video Streaming Architecture with QoS Control for Virtual
Desktop Infrastructure in Cloud Computing | 26 pages, Multimedia Tools and Applications Journal | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | In virtual desktop infrastructure (VDI) environments, the remote display
protocol has a big responsibility to transmit video data from a data
center-hosted desktop to the endpoint. The protocol must ensure a high level of
client perceived end-to-end quality of service (QoS) under heavy work load
conditions. Each remote display protocol works differently depending on the
network and which applications are being delivered. In healthcare applications,
doctors and nurses can use mobile devices directly to monitor patients.
Moreover, the ability to implement tasks requiring high consumption of CPU and
other resources is applicable to a variety of applications including research
and cloud gaming. Such computer games and complex processes will run on
powerful cloud servers and the screen contents will be transmitted to the
client. TO enable such applications, remote display technology requires further
enhancements to meet more stringent requirements on bandwidth and QoS, an to
allow realtime operation. In this paper, we present an architecture including
flexible QoS control to improve the user quality of experience (QoE). The QoS
control is developed based on linear regression modeling using historical
network data. Additionally, the architecture includes a novel compression
algorithm of 2D images, designed to guarantee the best image quality and to
reduce video delay; this algorithm is based on k-means clustering and can
satisfy the requirements of realtime onboard processing. Through simulations
with a real work dataset collected by the MIT Computer Science and Artificial
Lab, we present experimental as well as explain the performance of the QoS
system.
| [
{
"created": "Fri, 11 Mar 2022 03:22:11 GMT",
"version": "v1"
}
] | 2022-03-14 | [
[
"Nguyen",
"Huu-Quoc",
""
],
[
"Nguyen",
"Tien-Dung",
""
],
[
"Pham",
"Van-Nam",
""
],
[
"Pham",
"Xuan-Qui",
""
],
[
"Ngo",
"Quang-Thai",
""
],
[
"Huh",
"Eui-Nam",
""
]
] | In virtual desktop infrastructure (VDI) environments, the remote display protocol has a big responsibility to transmit video data from a data center-hosted desktop to the endpoint. The protocol must ensure a high level of client perceived end-to-end quality of service (QoS) under heavy work load conditions. Each remote display protocol works differently depending on the network and which applications are being delivered. In healthcare applications, doctors and nurses can use mobile devices directly to monitor patients. Moreover, the ability to implement tasks requiring high consumption of CPU and other resources is applicable to a variety of applications including research and cloud gaming. Such computer games and complex processes will run on powerful cloud servers and the screen contents will be transmitted to the client. TO enable such applications, remote display technology requires further enhancements to meet more stringent requirements on bandwidth and QoS, an to allow realtime operation. In this paper, we present an architecture including flexible QoS control to improve the user quality of experience (QoE). The QoS control is developed based on linear regression modeling using historical network data. Additionally, the architecture includes a novel compression algorithm of 2D images, designed to guarantee the best image quality and to reduce video delay; this algorithm is based on k-means clustering and can satisfy the requirements of realtime onboard processing. Through simulations with a real work dataset collected by the MIT Computer Science and Artificial Lab, we present experimental as well as explain the performance of the QoS system. |
2008.07689 | Yaorui Zhang | Yitong Deng, Yaorui Zhang, Xingzhe He, Shuqi Yang, Yunjin Tong,
Michael Zhang, Daniel DiPietro, Bo Zhu | Soft Multicopter Control using Neural Dynamics Identification | null | null | null | null | cs.RO cs.GR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Dynamic control of a soft-body robot to deliver complex behaviors with
low-dimensional actuation inputs is challenging. In this paper, we present a
computational approach to automatically generate versatile, underactuated
control policies that drives soft-bodied machines with complicated structures
and nonlinear dynamics. Our target application is focused on the autonomous
control of a soft multicopter, featured by its elastic material components,
non-conventional shapes, and asymmetric rotor layouts, to precisely deliver
compliant deformation and agile locomotion. The central piece of our approach
lies in a lightweight neural surrogate model to identify and predict the
temporal evolution of a set of geometric variables characterizing an elastic
soft body. This physics-based learning model is further integrated into a
Linear Quadratic Regulator (LQR) control loop enhanced by a novel online
fixed-point relinearization scheme to accommodate the dynamic body balance,
allowing an aggressive reduction of the computational overhead caused by the
conventional full-scale sensing-simulation-control workflow. We demonstrate the
efficacy of our approach by generating controllers for a broad spectrum of
customized soft multicopter designs and testing them in a high-fidelity physics
simulation environment. The control algorithm enables the multicopters to
perform a variety of tasks, including hovering, trajectory tracking, cruising
and active deforming.
| [
{
"created": "Tue, 18 Aug 2020 01:38:18 GMT",
"version": "v1"
},
{
"created": "Mon, 31 Aug 2020 19:37:18 GMT",
"version": "v2"
},
{
"created": "Wed, 2 Sep 2020 09:44:15 GMT",
"version": "v3"
},
{
"created": "Tue, 1 Dec 2020 09:11:02 GMT",
"version": "v4"
}
] | 2020-12-02 | [
[
"Deng",
"Yitong",
""
],
[
"Zhang",
"Yaorui",
""
],
[
"He",
"Xingzhe",
""
],
[
"Yang",
"Shuqi",
""
],
[
"Tong",
"Yunjin",
""
],
[
"Zhang",
"Michael",
""
],
[
"DiPietro",
"Daniel",
""
],
[
"Zhu",
"Bo",
""
]
] | Dynamic control of a soft-body robot to deliver complex behaviors with low-dimensional actuation inputs is challenging. In this paper, we present a computational approach to automatically generate versatile, underactuated control policies that drives soft-bodied machines with complicated structures and nonlinear dynamics. Our target application is focused on the autonomous control of a soft multicopter, featured by its elastic material components, non-conventional shapes, and asymmetric rotor layouts, to precisely deliver compliant deformation and agile locomotion. The central piece of our approach lies in a lightweight neural surrogate model to identify and predict the temporal evolution of a set of geometric variables characterizing an elastic soft body. This physics-based learning model is further integrated into a Linear Quadratic Regulator (LQR) control loop enhanced by a novel online fixed-point relinearization scheme to accommodate the dynamic body balance, allowing an aggressive reduction of the computational overhead caused by the conventional full-scale sensing-simulation-control workflow. We demonstrate the efficacy of our approach by generating controllers for a broad spectrum of customized soft multicopter designs and testing them in a high-fidelity physics simulation environment. The control algorithm enables the multicopters to perform a variety of tasks, including hovering, trajectory tracking, cruising and active deforming. |
1206.4126 | Yousuf Ibrahim Khan | Yousuf Ibrahim Khan | Image based Cryptography from a distance | null | null | null | null | cs.CR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | An information is a message which is received and understood. Information can
be sent one person to another over a long range but the process of sending
information must be done in a secure way especially in case of a private
message. Mathematicians and Engineers have historically relied on different
algorithmic techniques to secure messages and signals. Cryptography, to most
people, is concerned with keeping communications private. Indeed, the
protection of sensitive communications has been the emphasis of cryptography
throughout much of its history. Sometimes it is safer to send a message using
an image and thus cryptography can also be done using images during an
emergency. The need to extract information from images and interpret their
contents has been one of the driving factors in the development of image
processing and cryptography during the past decades. In this paper, a simple
cryptographic method was used to decode a message which was in an image and it
was done using a popular computational software.
| [
{
"created": "Tue, 19 Jun 2012 06:02:32 GMT",
"version": "v1"
}
] | 2012-06-20 | [
[
"Khan",
"Yousuf Ibrahim",
""
]
] | An information is a message which is received and understood. Information can be sent one person to another over a long range but the process of sending information must be done in a secure way especially in case of a private message. Mathematicians and Engineers have historically relied on different algorithmic techniques to secure messages and signals. Cryptography, to most people, is concerned with keeping communications private. Indeed, the protection of sensitive communications has been the emphasis of cryptography throughout much of its history. Sometimes it is safer to send a message using an image and thus cryptography can also be done using images during an emergency. The need to extract information from images and interpret their contents has been one of the driving factors in the development of image processing and cryptography during the past decades. In this paper, a simple cryptographic method was used to decode a message which was in an image and it was done using a popular computational software. |
1908.01650 | Chunming Tang | Sihem Mesnager, Yanfeng Qi, Hongming Ru, Chunming Tang | Minimal linear codes from characteristic functions | null | null | null | null | cs.IT math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Minimal linear codes have interesting applications in secret sharing schemes
and secure two-party computation. This paper uses characteristic functions of
some subsets of $\mathbb{F}_q$ to construct minimal linear codes. By properties
of characteristic functions, we can obtain more minimal binary linear codes
from known minimal binary linear codes, which generalizes results of Ding et
al. [IEEE Trans. Inf. Theory, vol. 64, no. 10, pp. 6536-6545, 2018]. By
characteristic functions corresponding to some subspaces of $\mathbb{F}_q$, we
obtain many minimal linear codes, which generalizes results of [IEEE Trans.
Inf. Theory, vol. 64, no. 10, pp. 6536-6545, 2018] and [IEEE Trans. Inf.
Theory, vol. 65, no. 11, pp. 7067-7078, 2019]. Finally, we use characteristic
functions to present a characterization of minimal linear codes from the
defining set method and present a class of minimal linear codes.
| [
{
"created": "Mon, 5 Aug 2019 14:40:23 GMT",
"version": "v1"
},
{
"created": "Wed, 20 Nov 2019 11:45:55 GMT",
"version": "v2"
}
] | 2019-11-21 | [
[
"Mesnager",
"Sihem",
""
],
[
"Qi",
"Yanfeng",
""
],
[
"Ru",
"Hongming",
""
],
[
"Tang",
"Chunming",
""
]
] | Minimal linear codes have interesting applications in secret sharing schemes and secure two-party computation. This paper uses characteristic functions of some subsets of $\mathbb{F}_q$ to construct minimal linear codes. By properties of characteristic functions, we can obtain more minimal binary linear codes from known minimal binary linear codes, which generalizes results of Ding et al. [IEEE Trans. Inf. Theory, vol. 64, no. 10, pp. 6536-6545, 2018]. By characteristic functions corresponding to some subspaces of $\mathbb{F}_q$, we obtain many minimal linear codes, which generalizes results of [IEEE Trans. Inf. Theory, vol. 64, no. 10, pp. 6536-6545, 2018] and [IEEE Trans. Inf. Theory, vol. 65, no. 11, pp. 7067-7078, 2019]. Finally, we use characteristic functions to present a characterization of minimal linear codes from the defining set method and present a class of minimal linear codes. |
2202.04076 | Kun Wang | Kun Wang, Jingyi Wang, Christopher M. Poskitt, Xiangxiang Chen, Jun
Sun, and Peng Cheng | K-ST: A Formal Executable Semantics of the Structured Text Language for
PLCs | Accepted by IEEE Transactions on Software Engineering | IEEE Trans. Software Eng., 2023 | 10.1109/TSE.2023.3315292 | null | cs.PL cs.SE | http://creativecommons.org/licenses/by/4.0/ | Programmable Logic Controllers (PLCs) are responsible for automating process
control in many industrial systems (e.g. in manufacturing and public
infrastructure), and thus it is critical to ensure that they operate correctly
and safely. The majority of PLCs are programmed in languages such as Structured
Text (ST). However, a lack of formal semantics makes it difficult to ascertain
the correctness of their translators and compilers, which vary from
vendor-to-vendor. In this work, we develop K-ST, a formal executable semantics
for ST in the K framework. Defined with respect to the IEC 61131-3 standard and
PLC vendor manuals, K-ST is a high-level reference semantics that can be used
to evaluate the correctness and consistency of different ST implementations. We
validate K-ST by executing 509 ST programs extracted from Github and comparing
the results against existing commercial compilers (i.e., CODESYS,
CX-Programmer, and GX Works2). We then apply K-ST to validate the
implementation of the open source OpenPLC platform, comparing the executions of
several test programs to uncover five bugs and nine functional defects in the
compiler.
| [
{
"created": "Tue, 8 Feb 2022 17:34:08 GMT",
"version": "v1"
},
{
"created": "Tue, 12 Sep 2023 02:05:17 GMT",
"version": "v2"
}
] | 2023-09-19 | [
[
"Wang",
"Kun",
""
],
[
"Wang",
"Jingyi",
""
],
[
"Poskitt",
"Christopher M.",
""
],
[
"Chen",
"Xiangxiang",
""
],
[
"Sun",
"Jun",
""
],
[
"Cheng",
"Peng",
""
]
] | Programmable Logic Controllers (PLCs) are responsible for automating process control in many industrial systems (e.g. in manufacturing and public infrastructure), and thus it is critical to ensure that they operate correctly and safely. The majority of PLCs are programmed in languages such as Structured Text (ST). However, a lack of formal semantics makes it difficult to ascertain the correctness of their translators and compilers, which vary from vendor-to-vendor. In this work, we develop K-ST, a formal executable semantics for ST in the K framework. Defined with respect to the IEC 61131-3 standard and PLC vendor manuals, K-ST is a high-level reference semantics that can be used to evaluate the correctness and consistency of different ST implementations. We validate K-ST by executing 509 ST programs extracted from Github and comparing the results against existing commercial compilers (i.e., CODESYS, CX-Programmer, and GX Works2). We then apply K-ST to validate the implementation of the open source OpenPLC platform, comparing the executions of several test programs to uncover five bugs and nine functional defects in the compiler. |
2306.17804 | Darren Strash | Anthony Hevia, Benjamin Kallus, Summer McClintic, Samantha Reisner,
Darren Strash, and Johnathan Wilson | Solving Edge Clique Cover Exactly via Synergistic Data Reduction | 22 pages, 5 figures, 6 tables, accepted at the 31st Annual European
Symposium on Algorithms (ESA 2023) | null | null | null | cs.DS | http://creativecommons.org/licenses/by/4.0/ | The edge clique cover (ECC) problem -- where the goal is to find a minimum
cardinality set of cliques that cover all the edges of a graph -- is a classic
NP-hard problem that has received much attention from both the theoretical and
experimental algorithms communities. While small sparse graphs can be solved
exactly via the branch-and-reduce algorithm of Gramm et al. [JEA 2009], larger
instances can currently only be solved inexactly using heuristics with unknown
overall solution quality. We revisit computing minimum ECCs exactly in practice
by combining data reduction for both the ECC \emph{and} vertex clique cover
(VCC) problems. We do so by modifying the polynomial-time reduction of Kou et
al. [Commun. ACM 1978] to transform a reduced ECC instance to a VCC instance;
alternatively, we show it is possible to ``lift'' some VCC reductions to the
ECC problem. Our experiments show that combining data reduction for both
problems (which we call \emph{synergistic data reduction}) enables finding
exact minimum ECCs orders of magnitude faster than the technique of Gramm et
al., and allows solving large sparse graphs on up to millions of vertices and
edges that have never before been solved. With these new exact solutions, we
evaluate the quality of recent heuristic algorithms on large instances for the
first time. The most recent of these, \textsf{EO-ECC} by Abdullah et al. [ICCS
2022], solves 8 of the 27 instances for which we have exact solutions. It is
our hope that our strategy rallies researchers to seek improved algorithms for
the ECC problem.
| [
{
"created": "Fri, 30 Jun 2023 17:06:04 GMT",
"version": "v1"
},
{
"created": "Tue, 4 Jul 2023 18:04:39 GMT",
"version": "v2"
}
] | 2023-07-06 | [
[
"Hevia",
"Anthony",
""
],
[
"Kallus",
"Benjamin",
""
],
[
"McClintic",
"Summer",
""
],
[
"Reisner",
"Samantha",
""
],
[
"Strash",
"Darren",
""
],
[
"Wilson",
"Johnathan",
""
]
] | The edge clique cover (ECC) problem -- where the goal is to find a minimum cardinality set of cliques that cover all the edges of a graph -- is a classic NP-hard problem that has received much attention from both the theoretical and experimental algorithms communities. While small sparse graphs can be solved exactly via the branch-and-reduce algorithm of Gramm et al. [JEA 2009], larger instances can currently only be solved inexactly using heuristics with unknown overall solution quality. We revisit computing minimum ECCs exactly in practice by combining data reduction for both the ECC \emph{and} vertex clique cover (VCC) problems. We do so by modifying the polynomial-time reduction of Kou et al. [Commun. ACM 1978] to transform a reduced ECC instance to a VCC instance; alternatively, we show it is possible to ``lift'' some VCC reductions to the ECC problem. Our experiments show that combining data reduction for both problems (which we call \emph{synergistic data reduction}) enables finding exact minimum ECCs orders of magnitude faster than the technique of Gramm et al., and allows solving large sparse graphs on up to millions of vertices and edges that have never before been solved. With these new exact solutions, we evaluate the quality of recent heuristic algorithms on large instances for the first time. The most recent of these, \textsf{EO-ECC} by Abdullah et al. [ICCS 2022], solves 8 of the 27 instances for which we have exact solutions. It is our hope that our strategy rallies researchers to seek improved algorithms for the ECC problem. |
2209.14468 | Yiheng Shen | Kamesh Munagala, Yiheng Shen, Kangning Wang | Auditing for Core Stability in Participatory Budgeting | accepted by the 18th Conference on Web and Internet Economics (WINE
2022) | null | null | null | cs.GT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We consider the participatory budgeting problem where each of $n$ voters
specifies additive utilities over $m$ candidate projects with given sizes, and
the goal is to choose a subset of projects (i.e., a committee) with total size
at most $k$. Participatory budgeting mathematically generalizes multiwinner
elections, and both have received great attention in computational social
choice recently. A well-studied notion of group fairness in this setting is
core stability: Each voter is assigned an "entitlement" of $\frac{k}{n}$, so
that a subset $S$ of voters can pay for a committee of size at most $|S| \cdot
\frac{k}{n}$. A given committee is in the core if no subset of voters can pay
for another committee that provides each of them strictly larger utility. This
provides proportional representation to all voters in a strong sense.
In this paper, we study the following auditing question: Given a committee
computed by some preference aggregation method, how close is it to the core?
Concretely, how much does the entitlement of each voter need to be scaled down
by, so that the core property subsequently holds? As our main contribution, we
present computational hardness results for this problem, as well as a
logarithmic approximation algorithm via linear program rounding. We show that
our analysis is tight against the linear programming bound. Additionally, we
consider two related notions of group fairness that have similar audit
properties. The first is Lindahl priceability, which audits the closeness of a
committee to a market clearing solution. We show that this is related to the
linear programming relaxation of auditing the core, leading to efficient exact
and approximation algorithms for auditing. The second is a novel weakening of
the core that we term the sub-core, and we present computational results for
auditing this notion as well.
| [
{
"created": "Wed, 28 Sep 2022 23:13:06 GMT",
"version": "v1"
}
] | 2022-09-30 | [
[
"Munagala",
"Kamesh",
""
],
[
"Shen",
"Yiheng",
""
],
[
"Wang",
"Kangning",
""
]
] | We consider the participatory budgeting problem where each of $n$ voters specifies additive utilities over $m$ candidate projects with given sizes, and the goal is to choose a subset of projects (i.e., a committee) with total size at most $k$. Participatory budgeting mathematically generalizes multiwinner elections, and both have received great attention in computational social choice recently. A well-studied notion of group fairness in this setting is core stability: Each voter is assigned an "entitlement" of $\frac{k}{n}$, so that a subset $S$ of voters can pay for a committee of size at most $|S| \cdot \frac{k}{n}$. A given committee is in the core if no subset of voters can pay for another committee that provides each of them strictly larger utility. This provides proportional representation to all voters in a strong sense. In this paper, we study the following auditing question: Given a committee computed by some preference aggregation method, how close is it to the core? Concretely, how much does the entitlement of each voter need to be scaled down by, so that the core property subsequently holds? As our main contribution, we present computational hardness results for this problem, as well as a logarithmic approximation algorithm via linear program rounding. We show that our analysis is tight against the linear programming bound. Additionally, we consider two related notions of group fairness that have similar audit properties. The first is Lindahl priceability, which audits the closeness of a committee to a market clearing solution. We show that this is related to the linear programming relaxation of auditing the core, leading to efficient exact and approximation algorithms for auditing. The second is a novel weakening of the core that we term the sub-core, and we present computational results for auditing this notion as well. |
2408.01196 | Zhang Shanfan | Shanfan Zhang, Xiaoting Shen, Zhan Bu | Game Theory Based Community-Aware Opinion Dynamics | 36 pages, 15figures | null | null | null | cs.GT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Examining the mechanisms underlying the formation and evolution of opinions
within real-world social systems, which consist of numerous individuals, can
provide valuable insights for effective social functioning and informed
business decision making. The focus of our study is on the dynamics of opinions
inside a networked multi-agent system. We provide a novel approach called the
Game Theory Based Community-Aware Opinion Formation Process (GCAOFP) to
accurately represent the co-evolutionary dynamics of communities and opinions
in real-world social systems. The GCAOFP algorithm comprises two distinct steps
in each iteration. 1) The Community Dynamics Process conceptualizes the process
of community formation as a non-cooperative game involving a finite number of
agents. Each individual agent aims to maximize their own utility by adopting a
response that leads to the most favorable update of the community label. 2) The
Opinion Formation Process involves the updating of an individual agent's
opinion within a community-aware framework that incorporates bounded
confidence. This process takes into account the updated matrix of community
members and ensures that an agent's opinion aligns with the opinions of others
within their community, within certain defined limits. The present study
provides a theoretical proof that under any initial conditions, the
aforementioned co-evolutionary dynamics process will ultimately reach an
equilibrium state. In this state, both the opinion vector and community member
matrix will stabilize after a finite number of iterations. In contrast to
conventional opinion dynamics models, the guaranteed convergence of agent
opinion within the same community ensures that the convergence of opinions
takes place exclusively inside a given community.
| [
{
"created": "Fri, 2 Aug 2024 11:24:56 GMT",
"version": "v1"
}
] | 2024-08-05 | [
[
"Zhang",
"Shanfan",
""
],
[
"Shen",
"Xiaoting",
""
],
[
"Bu",
"Zhan",
""
]
] | Examining the mechanisms underlying the formation and evolution of opinions within real-world social systems, which consist of numerous individuals, can provide valuable insights for effective social functioning and informed business decision making. The focus of our study is on the dynamics of opinions inside a networked multi-agent system. We provide a novel approach called the Game Theory Based Community-Aware Opinion Formation Process (GCAOFP) to accurately represent the co-evolutionary dynamics of communities and opinions in real-world social systems. The GCAOFP algorithm comprises two distinct steps in each iteration. 1) The Community Dynamics Process conceptualizes the process of community formation as a non-cooperative game involving a finite number of agents. Each individual agent aims to maximize their own utility by adopting a response that leads to the most favorable update of the community label. 2) The Opinion Formation Process involves the updating of an individual agent's opinion within a community-aware framework that incorporates bounded confidence. This process takes into account the updated matrix of community members and ensures that an agent's opinion aligns with the opinions of others within their community, within certain defined limits. The present study provides a theoretical proof that under any initial conditions, the aforementioned co-evolutionary dynamics process will ultimately reach an equilibrium state. In this state, both the opinion vector and community member matrix will stabilize after a finite number of iterations. In contrast to conventional opinion dynamics models, the guaranteed convergence of agent opinion within the same community ensures that the convergence of opinions takes place exclusively inside a given community. |
1912.12204 | Boyi Liu | Boyi Liu, Lujia Wang, Ming Liu, Cheng-Zhong Xu | Federated Imitation Learning: A Novel Framework for Cloud Robotic
Systems with Heterogeneous Sensor Data | arXiv admin note: substantial text overlap with arXiv:1909.00895 | null | null | null | cs.RO cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Humans are capable of learning a new behavior by observing others to perform
the skill. Similarly, robots can also implement this by imitation learning.
Furthermore, if with external guidance, humans can master the new behavior more
efficiently. So, how can robots achieve this? To address the issue, we present
a novel framework named FIL. It provides a heterogeneous knowledge fusion
mechanism for cloud robotic systems. Then, a knowledge fusion algorithm in FIL
is proposed. It enables the cloud to fuse heterogeneous knowledge from local
robots and generate guide models for robots with service requests. After that,
we introduce a knowledge transfer scheme to facilitate local robots acquiring
knowledge from the cloud. With FIL, a robot is capable of utilizing knowledge
from other robots to increase its imitation learning in accuracy and
efficiency. Compared with transfer learning and meta-learning, FIL is more
suitable to be deployed in cloud robotic systems. Finally, we conduct
experiments of a self-driving task for robots (cars). The experimental results
demonstrate that the shared model generated by FIL increases imitation learning
efficiency of local robots in cloud robotic systems.
| [
{
"created": "Tue, 24 Dec 2019 11:23:23 GMT",
"version": "v1"
}
] | 2019-12-30 | [
[
"Liu",
"Boyi",
""
],
[
"Wang",
"Lujia",
""
],
[
"Liu",
"Ming",
""
],
[
"Xu",
"Cheng-Zhong",
""
]
] | Humans are capable of learning a new behavior by observing others to perform the skill. Similarly, robots can also implement this by imitation learning. Furthermore, if with external guidance, humans can master the new behavior more efficiently. So, how can robots achieve this? To address the issue, we present a novel framework named FIL. It provides a heterogeneous knowledge fusion mechanism for cloud robotic systems. Then, a knowledge fusion algorithm in FIL is proposed. It enables the cloud to fuse heterogeneous knowledge from local robots and generate guide models for robots with service requests. After that, we introduce a knowledge transfer scheme to facilitate local robots acquiring knowledge from the cloud. With FIL, a robot is capable of utilizing knowledge from other robots to increase its imitation learning in accuracy and efficiency. Compared with transfer learning and meta-learning, FIL is more suitable to be deployed in cloud robotic systems. Finally, we conduct experiments of a self-driving task for robots (cars). The experimental results demonstrate that the shared model generated by FIL increases imitation learning efficiency of local robots in cloud robotic systems. |
2311.15512 | Dong Yonghao | Yonghao Dong, Le Wang, Sanpin Zhou, Gang Hua, and Changyin Sun | Sparse Pedestrian Character Learning for Trajectory Prediction | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Pedestrian trajectory prediction in a first-person view has recently
attracted much attention due to its importance in autonomous driving. Recent
work utilizes pedestrian character information, \textit{i.e.}, action and
appearance, to improve the learned trajectory embedding and achieves
state-of-the-art performance. However, it neglects the invalid and negative
pedestrian character information, which is harmful to trajectory representation
and thus leads to performance degradation. To address this issue, we present a
two-stream sparse-character-based network~(TSNet) for pedestrian trajectory
prediction. Specifically, TSNet learns the negative-removed characters in the
sparse character representation stream to improve the trajectory embedding
obtained in the trajectory representation stream. Moreover, to model the
negative-removed characters, we propose a novel sparse character graph,
including the sparse category and sparse temporal character graphs, to learn
the different effects of various characters in category and temporal
dimensions, respectively. Extensive experiments on two first-person view
datasets, PIE and JAAD, show that our method outperforms existing
state-of-the-art methods. In addition, ablation studies demonstrate different
effects of various characters and prove that TSNet outperforms approaches
without eliminating negative characters.
| [
{
"created": "Mon, 27 Nov 2023 03:15:48 GMT",
"version": "v1"
}
] | 2023-11-28 | [
[
"Dong",
"Yonghao",
""
],
[
"Wang",
"Le",
""
],
[
"Zhou",
"Sanpin",
""
],
[
"Hua",
"Gang",
""
],
[
"Sun",
"Changyin",
""
]
] | Pedestrian trajectory prediction in a first-person view has recently attracted much attention due to its importance in autonomous driving. Recent work utilizes pedestrian character information, \textit{i.e.}, action and appearance, to improve the learned trajectory embedding and achieves state-of-the-art performance. However, it neglects the invalid and negative pedestrian character information, which is harmful to trajectory representation and thus leads to performance degradation. To address this issue, we present a two-stream sparse-character-based network~(TSNet) for pedestrian trajectory prediction. Specifically, TSNet learns the negative-removed characters in the sparse character representation stream to improve the trajectory embedding obtained in the trajectory representation stream. Moreover, to model the negative-removed characters, we propose a novel sparse character graph, including the sparse category and sparse temporal character graphs, to learn the different effects of various characters in category and temporal dimensions, respectively. Extensive experiments on two first-person view datasets, PIE and JAAD, show that our method outperforms existing state-of-the-art methods. In addition, ablation studies demonstrate different effects of various characters and prove that TSNet outperforms approaches without eliminating negative characters. |
2209.09653 | Tonio Ball | Maryna Kapitonova, Philipp Kellmeyer, Simon Vogt and Tonio Ball | A Framework for Preserving Privacy and Cybersecurity in Brain-Computer
Interfacing Applications | null | null | null | null | cs.CR | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Brain-Computer Interfaces (BCIs) comprise a rapidly evolving field of
technology with the potential of far-reaching impact in domains ranging from
medical over industrial to artistic, gaming, and military. Today, these
emerging BCI applications are typically still at early technology readiness
levels, but because BCIs create novel, technical communication channels for the
human brain, they have raised privacy and security concerns. To mitigate such
risks, a large body of countermeasures has been proposed in the literature, but
a general framework is lacking which would describe how privacy and security of
BCI applications can be protected by design, i.e., already as an integral part
of the early BCI design process, in a systematic manner, and allowing suitable
depth of analysis for different contexts such as commercial BCI product
development vs. academic research and lab prototypes. Here we propose the
adoption of recent systems-engineering methodologies for privacy threat
modeling, risk assessment, and privacy engineering to the BCI field. These
methodologies address privacy and security concerns in a more systematic and
holistic way than previous approaches, and provide reusable patterns on how to
move from principles to actions. We apply these methodologies to BCI and data
flows and derive a generic, extensible, and actionable framework for
brain-privacy-preserving cybersecurity in BCI applications. This framework is
designed for flexible application to the wide range of current and future BCI
applications. We also propose a range of novel privacy-by-design features for
BCIs, with an emphasis on features promoting BCI transparency as a prerequisite
for informational self-determination of BCI users, as well as design features
for ensuring BCI user autonomy. We anticipate that our framework will
contribute to the development of privacy-respecting, trustworthy BCI
technologies.
| [
{
"created": "Mon, 19 Sep 2022 15:45:13 GMT",
"version": "v1"
}
] | 2022-09-21 | [
[
"Kapitonova",
"Maryna",
""
],
[
"Kellmeyer",
"Philipp",
""
],
[
"Vogt",
"Simon",
""
],
[
"Ball",
"Tonio",
""
]
] | Brain-Computer Interfaces (BCIs) comprise a rapidly evolving field of technology with the potential of far-reaching impact in domains ranging from medical over industrial to artistic, gaming, and military. Today, these emerging BCI applications are typically still at early technology readiness levels, but because BCIs create novel, technical communication channels for the human brain, they have raised privacy and security concerns. To mitigate such risks, a large body of countermeasures has been proposed in the literature, but a general framework is lacking which would describe how privacy and security of BCI applications can be protected by design, i.e., already as an integral part of the early BCI design process, in a systematic manner, and allowing suitable depth of analysis for different contexts such as commercial BCI product development vs. academic research and lab prototypes. Here we propose the adoption of recent systems-engineering methodologies for privacy threat modeling, risk assessment, and privacy engineering to the BCI field. These methodologies address privacy and security concerns in a more systematic and holistic way than previous approaches, and provide reusable patterns on how to move from principles to actions. We apply these methodologies to BCI and data flows and derive a generic, extensible, and actionable framework for brain-privacy-preserving cybersecurity in BCI applications. This framework is designed for flexible application to the wide range of current and future BCI applications. We also propose a range of novel privacy-by-design features for BCIs, with an emphasis on features promoting BCI transparency as a prerequisite for informational self-determination of BCI users, as well as design features for ensuring BCI user autonomy. We anticipate that our framework will contribute to the development of privacy-respecting, trustworthy BCI technologies. |
2112.14890 | Ke Wang | Jiayi Wang, Ke Wang, Boxing Chen, Yu Zhao, Weihua Luo, and Yuqi Zhang | QEMind: Alibaba's Submission to the WMT21 Quality Estimation Shared Task | Winner of WMT 2021 QE shared task 1 | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Quality Estimation, as a crucial step of quality control for machine
translation, has been explored for years. The goal is to investigate automatic
methods for estimating the quality of machine translation results without
reference translations. In this year's WMT QE shared task, we utilize the
large-scale XLM-Roberta pre-trained model and additionally propose several
useful features to evaluate the uncertainty of the translations to build our QE
system, named \textit{QEMind}. The system has been applied to the
sentence-level scoring task of Direct Assessment and the binary score
prediction task of Critical Error Detection. In this paper, we present our
submissions to the WMT 2021 QE shared task and an extensive set of experimental
results have shown us that our multilingual systems outperform the best system
in the Direct Assessment QE task of WMT 2020.
| [
{
"created": "Thu, 30 Dec 2021 02:27:29 GMT",
"version": "v1"
}
] | 2022-01-03 | [
[
"Wang",
"Jiayi",
""
],
[
"Wang",
"Ke",
""
],
[
"Chen",
"Boxing",
""
],
[
"Zhao",
"Yu",
""
],
[
"Luo",
"Weihua",
""
],
[
"Zhang",
"Yuqi",
""
]
] | Quality Estimation, as a crucial step of quality control for machine translation, has been explored for years. The goal is to investigate automatic methods for estimating the quality of machine translation results without reference translations. In this year's WMT QE shared task, we utilize the large-scale XLM-Roberta pre-trained model and additionally propose several useful features to evaluate the uncertainty of the translations to build our QE system, named \textit{QEMind}. The system has been applied to the sentence-level scoring task of Direct Assessment and the binary score prediction task of Critical Error Detection. In this paper, we present our submissions to the WMT 2021 QE shared task and an extensive set of experimental results have shown us that our multilingual systems outperform the best system in the Direct Assessment QE task of WMT 2020. |
2008.05297 | Umberto Straccia | Franco Alberto Cardillo and Umberto Straccia | Fuzzy OWL-BOOST: Learning Fuzzy Concept Inclusions via Real-Valued
Boosting | null | null | 10.1016/j.fss.2021.07.002 | null | cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | OWL ontologies are nowadays a quite popular way to describe structured
knowledge in terms of classes, relations among classes and class instances. In
this paper, given a target class T of an OWL ontology, we address the problem
of learning fuzzy concept inclusion axioms that describe sufficient conditions
for being an individual instance of T. To do so, we present Fuzzy OWL-BOOST
that relies on the Real AdaBoost boosting algorithm adapted to the (fuzzy) OWL
case. We illustrate its effectiveness by means of an experimentation. An
interesting feature is that the learned rules can be represented directly into
Fuzzy OWL 2. As a consequence, any Fuzzy OWL 2 reasoner can then be used to
automatically determine/classify (and to which degree) whether an individual
belongs to the target class T.
| [
{
"created": "Mon, 3 Aug 2020 15:19:31 GMT",
"version": "v1"
},
{
"created": "Fri, 26 Mar 2021 07:10:04 GMT",
"version": "v2"
}
] | 2022-03-10 | [
[
"Cardillo",
"Franco Alberto",
""
],
[
"Straccia",
"Umberto",
""
]
] | OWL ontologies are nowadays a quite popular way to describe structured knowledge in terms of classes, relations among classes and class instances. In this paper, given a target class T of an OWL ontology, we address the problem of learning fuzzy concept inclusion axioms that describe sufficient conditions for being an individual instance of T. To do so, we present Fuzzy OWL-BOOST that relies on the Real AdaBoost boosting algorithm adapted to the (fuzzy) OWL case. We illustrate its effectiveness by means of an experimentation. An interesting feature is that the learned rules can be represented directly into Fuzzy OWL 2. As a consequence, any Fuzzy OWL 2 reasoner can then be used to automatically determine/classify (and to which degree) whether an individual belongs to the target class T. |
2305.09281 | Fatma Elsafoury | Fatma Elsafoury, Gavin Abercrombie | On the Origins of Bias in NLP through the Lens of the Jim Code | 10 pages | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | In this paper, we trace the biases in current natural language processing
(NLP) models back to their origins in racism, sexism, and homophobia over the
last 500 years. We review literature from critical race theory, gender studies,
data ethics, and digital humanities studies, and summarize the origins of bias
in NLP models from these social science perspective. We show how the causes of
the biases in the NLP pipeline are rooted in social issues. Finally, we argue
that the only way to fix the bias and unfairness in NLP is by addressing the
social problems that caused them in the first place and by incorporating social
sciences and social scientists in efforts to mitigate bias in NLP models. We
provide actionable recommendations for the NLP research community to do so.
| [
{
"created": "Tue, 16 May 2023 08:37:13 GMT",
"version": "v1"
}
] | 2023-05-17 | [
[
"Elsafoury",
"Fatma",
""
],
[
"Abercrombie",
"Gavin",
""
]
] | In this paper, we trace the biases in current natural language processing (NLP) models back to their origins in racism, sexism, and homophobia over the last 500 years. We review literature from critical race theory, gender studies, data ethics, and digital humanities studies, and summarize the origins of bias in NLP models from these social science perspective. We show how the causes of the biases in the NLP pipeline are rooted in social issues. Finally, we argue that the only way to fix the bias and unfairness in NLP is by addressing the social problems that caused them in the first place and by incorporating social sciences and social scientists in efforts to mitigate bias in NLP models. We provide actionable recommendations for the NLP research community to do so. |
2305.11059 | Muhammad Husnain Mubarik | Ramakrishna Kanungo, Swamynathan Siva, Nathaniel Bleier, Muhammad
Husnain Mubarik, Lav Varshney and Rakesh Kumar | Understanding Interactions Between Chip Architecture and Uncertainties
in Semiconductor Supply and Demand | null | null | null | null | cs.AR cs.CE | http://creativecommons.org/licenses/by/4.0/ | Mitigating losses from supply and demand volatility in the semiconductor
supply chain and market has traditionally been cast as a logistics and
forecasting problem. We investigate how the architecture of a family of chips
influences how it is affected by supply and demand uncertainties. We observe
that semiconductor supply chains become fragile, in part, due to single demand
paths, where one chip can satisfy only one demand. Chip architects can enable
multiple paths to satisfy a chip demand, which improves supply chain
resilience. Based on this observation, we study composition and adaptation as
architectural strategies to improve resilience to volatility and also introduce
a third strategy of dispersion. These strategies allow multiple paths to
satisfy a given chip demand. We develop a model to analyze the impact of these
architectural techniques on supply chain costs under different regimes of
uncertainties and evaluate what happens when they are combined. We present
several interesting and even counterintuitive observations about the product
configurations and market conditions where these interventions are impactful
and where they are not. In all, we show that product redesign supported by
architectural changes can mitigate nearly half of the losses caused by supply
and demand volatility. As far as we know, this is the first such investigation
concerning chip architecture.
| [
{
"created": "Wed, 10 May 2023 18:07:34 GMT",
"version": "v1"
}
] | 2023-05-19 | [
[
"Kanungo",
"Ramakrishna",
""
],
[
"Siva",
"Swamynathan",
""
],
[
"Bleier",
"Nathaniel",
""
],
[
"Mubarik",
"Muhammad Husnain",
""
],
[
"Varshney",
"Lav",
""
],
[
"Kumar",
"Rakesh",
""
]
] | Mitigating losses from supply and demand volatility in the semiconductor supply chain and market has traditionally been cast as a logistics and forecasting problem. We investigate how the architecture of a family of chips influences how it is affected by supply and demand uncertainties. We observe that semiconductor supply chains become fragile, in part, due to single demand paths, where one chip can satisfy only one demand. Chip architects can enable multiple paths to satisfy a chip demand, which improves supply chain resilience. Based on this observation, we study composition and adaptation as architectural strategies to improve resilience to volatility and also introduce a third strategy of dispersion. These strategies allow multiple paths to satisfy a given chip demand. We develop a model to analyze the impact of these architectural techniques on supply chain costs under different regimes of uncertainties and evaluate what happens when they are combined. We present several interesting and even counterintuitive observations about the product configurations and market conditions where these interventions are impactful and where they are not. In all, we show that product redesign supported by architectural changes can mitigate nearly half of the losses caused by supply and demand volatility. As far as we know, this is the first such investigation concerning chip architecture. |
2204.02464 | Stefan Bosse | Stefan Bosse | BeeTS: Smart Distributed Sensor Tuple Spaces combined with Agents using
Bluetooth and IP Broadcasting | null | null | null | null | cs.NI cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Most Internet-of-Things (IoT) devices and smart sensors are connected via the
Internet using IP communication driectly accessed by a server that collect
sensor information periodically or event-based. Although, Internet access is
widely available, there are places that are not covered and WLAN and mobile
cell communication requires a descent amount of power not always available.
Finally, the spatial context (the environment in which the sensor or devices is
situated) is not considered (or lost) by Internet connectivity. In this work,
smart devices communicate connectionless and ad-hoc by using low-energy
Bluetooth broadcasting available in any smartphone and in most embedded
computers, e.g., the Raspberry PI devices. Bi-directional connectionless
communication is established via the advertisements and scanning modes. The
communication nodes can exchange data via functional tuples using a tuple space
service on each node. Tuple space access is performed by simple evenat-based
agents. Mobile devices act as tuple carriers that can carry tuples between
different locations. Additionally, UDP-based Intranet communication can be used
to access tuple spaces on a wider spatial range. The Bluetooth Low Energy Tuple
Space (BeeTS) service enables opportunistic, ad-hoc and loosely coupled device
communication with a spatial context.
| [
{
"created": "Tue, 5 Apr 2022 19:47:21 GMT",
"version": "v1"
}
] | 2022-04-07 | [
[
"Bosse",
"Stefan",
""
]
] | Most Internet-of-Things (IoT) devices and smart sensors are connected via the Internet using IP communication driectly accessed by a server that collect sensor information periodically or event-based. Although, Internet access is widely available, there are places that are not covered and WLAN and mobile cell communication requires a descent amount of power not always available. Finally, the spatial context (the environment in which the sensor or devices is situated) is not considered (or lost) by Internet connectivity. In this work, smart devices communicate connectionless and ad-hoc by using low-energy Bluetooth broadcasting available in any smartphone and in most embedded computers, e.g., the Raspberry PI devices. Bi-directional connectionless communication is established via the advertisements and scanning modes. The communication nodes can exchange data via functional tuples using a tuple space service on each node. Tuple space access is performed by simple evenat-based agents. Mobile devices act as tuple carriers that can carry tuples between different locations. Additionally, UDP-based Intranet communication can be used to access tuple spaces on a wider spatial range. The Bluetooth Low Energy Tuple Space (BeeTS) service enables opportunistic, ad-hoc and loosely coupled device communication with a spatial context. |
1901.06212 | Dmitry Kangin | Dmitry Kangin and Nicolas Pugeault | On-Policy Trust Region Policy Optimisation with Replay Buffers | null | null | null | null | cs.LG cs.AI stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Building upon the recent success of deep reinforcement learning methods, we
investigate the possibility of on-policy reinforcement learning improvement by
reusing the data from several consecutive policies. On-policy methods bring
many benefits, such as ability to evaluate each resulting policy. However, they
usually discard all the information about the policies which existed before. In
this work, we propose adaptation of the replay buffer concept, borrowed from
the off-policy learning setting, to create the method, combining advantages of
on- and off-policy learning. To achieve this, the proposed algorithm
generalises the $Q$-, value and advantage functions for data from multiple
policies. The method uses trust region optimisation, while avoiding some of the
common problems of the algorithms such as TRPO or ACKTR: it uses
hyperparameters to replace the trust region selection heuristics, as well as
the trainable covariance matrix instead of the fixed one. In many cases, the
method not only improves the results comparing to the state-of-the-art trust
region on-policy learning algorithms such as PPO, ACKTR and TRPO, but also with
respect to their off-policy counterpart DDPG.
| [
{
"created": "Fri, 18 Jan 2019 13:09:18 GMT",
"version": "v1"
}
] | 2019-01-21 | [
[
"Kangin",
"Dmitry",
""
],
[
"Pugeault",
"Nicolas",
""
]
] | Building upon the recent success of deep reinforcement learning methods, we investigate the possibility of on-policy reinforcement learning improvement by reusing the data from several consecutive policies. On-policy methods bring many benefits, such as ability to evaluate each resulting policy. However, they usually discard all the information about the policies which existed before. In this work, we propose adaptation of the replay buffer concept, borrowed from the off-policy learning setting, to create the method, combining advantages of on- and off-policy learning. To achieve this, the proposed algorithm generalises the $Q$-, value and advantage functions for data from multiple policies. The method uses trust region optimisation, while avoiding some of the common problems of the algorithms such as TRPO or ACKTR: it uses hyperparameters to replace the trust region selection heuristics, as well as the trainable covariance matrix instead of the fixed one. In many cases, the method not only improves the results comparing to the state-of-the-art trust region on-policy learning algorithms such as PPO, ACKTR and TRPO, but also with respect to their off-policy counterpart DDPG. |
2306.06791 | Shugang Hao | Shugang Hao and Lingjie Duan | To Save Mobile Crowdsourcing from Cheap-talk: A Game Theoretic Learning
Approach | null | null | null | null | cs.GT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Today mobile crowdsourcing platforms invite users to provide anonymous
reviews about service experiences, yet many reviews are found biased to be
extremely positive or negative. The existing methods find it difficult to learn
from biased reviews to infer the actual service state, as the state can also be
extreme and the platform cannot verify the truthfulness of reviews immediately.
Further, reviewers can hide their (positive or negative) bias types and
proactively adjust their anonymous reviews against the platform's inference. To
our best knowledge, we are the first to study how to save mobile crowdsourcing
from cheap-talk and strategically learn from biased users' reviews. We
formulate the problem as a dynamic Bayesian game, including users' service-type
messaging and the platform's follow-up rating/inference. Our closed-form PBE
shows that an extremely-biased user may still honestly message to convince the
platform of listening to his review. Such Bayesian game-theoretic learning
obviously outperforms the latest common schemes especially when there are
multiple diversely-biased users to compete. For the challenging single-user
case, we further propose a time-evolving mechanism with the platform's
commitment inferences to ensure the biased user's truthful messaging all the
time, whose performance improves with more time periods to learn from more
historical data.
| [
{
"created": "Sun, 11 Jun 2023 22:07:18 GMT",
"version": "v1"
},
{
"created": "Fri, 29 Dec 2023 05:10:44 GMT",
"version": "v2"
}
] | 2024-01-01 | [
[
"Hao",
"Shugang",
""
],
[
"Duan",
"Lingjie",
""
]
] | Today mobile crowdsourcing platforms invite users to provide anonymous reviews about service experiences, yet many reviews are found biased to be extremely positive or negative. The existing methods find it difficult to learn from biased reviews to infer the actual service state, as the state can also be extreme and the platform cannot verify the truthfulness of reviews immediately. Further, reviewers can hide their (positive or negative) bias types and proactively adjust their anonymous reviews against the platform's inference. To our best knowledge, we are the first to study how to save mobile crowdsourcing from cheap-talk and strategically learn from biased users' reviews. We formulate the problem as a dynamic Bayesian game, including users' service-type messaging and the platform's follow-up rating/inference. Our closed-form PBE shows that an extremely-biased user may still honestly message to convince the platform of listening to his review. Such Bayesian game-theoretic learning obviously outperforms the latest common schemes especially when there are multiple diversely-biased users to compete. For the challenging single-user case, we further propose a time-evolving mechanism with the platform's commitment inferences to ensure the biased user's truthful messaging all the time, whose performance improves with more time periods to learn from more historical data. |
2210.14638 | Marcin Pilipczuk | Daniel Lokshtanov and Marcin Pilipczuk and Micha{\l} Pilipczuk and
Saket Saurabh | Fixed-parameter tractability of Graph Isomorphism in graphs with an
excluded minor | Part I of a full version of a paper accepted at STOC 2022 | null | null | null | cs.DS cs.DM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We prove that Graph Isomorphism and Canonization in graphs excluding a fixed
graph $H$ as a minor can be solved by an algorithm working in time $f(H)\cdot
n^{O(1)}$, where $f$ is some function. In other words, we show that these
problems are fixed-parameter tractable when parameterized by the size of the
excluded minor, with the caveat that the bound on the running time is not
necessarily computable. The underlying approach is based on decomposing the
graph in a canonical way into unbreakable (intuitively, well-connected) parts,
which essentially provides a reduction to the case where the given
$H$-minor-free graph is unbreakable itself. This is complemented by an analysis
of unbreakable $H$-minor-free graphs, performed in a second subordinate
manuscript, which reveals that every such graph can be canonically decomposed
into a part that admits few automorphisms and a part that has bounded
treewidth.
| [
{
"created": "Wed, 26 Oct 2022 11:32:55 GMT",
"version": "v1"
}
] | 2022-10-27 | [
[
"Lokshtanov",
"Daniel",
""
],
[
"Pilipczuk",
"Marcin",
""
],
[
"Pilipczuk",
"Michał",
""
],
[
"Saurabh",
"Saket",
""
]
] | We prove that Graph Isomorphism and Canonization in graphs excluding a fixed graph $H$ as a minor can be solved by an algorithm working in time $f(H)\cdot n^{O(1)}$, where $f$ is some function. In other words, we show that these problems are fixed-parameter tractable when parameterized by the size of the excluded minor, with the caveat that the bound on the running time is not necessarily computable. The underlying approach is based on decomposing the graph in a canonical way into unbreakable (intuitively, well-connected) parts, which essentially provides a reduction to the case where the given $H$-minor-free graph is unbreakable itself. This is complemented by an analysis of unbreakable $H$-minor-free graphs, performed in a second subordinate manuscript, which reveals that every such graph can be canonically decomposed into a part that admits few automorphisms and a part that has bounded treewidth. |
1911.02423 | Dorjan Hitaj | Fabio De Gaspari, Dorjan Hitaj, Giulio Pagnotta, Lorenzo De Carli,
Luigi V. Mancini | The Naked Sun: Malicious Cooperation Between Benign-Looking Processes | 15 pages, 6 figures, 4 tables | null | null | null | cs.CR cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent progress in machine learning has generated promising results in
behavioral malware detection. Behavioral modeling identifies malicious
processes via features derived by their runtime behavior. Behavioral features
hold great promise as they are intrinsically related to the functioning of each
malware, and are therefore considered difficult to evade. Indeed, while a
significant amount of results exists on evasion of static malware features,
evasion of dynamic features has seen limited work. This paper thoroughly
examines the robustness of behavioral malware detectors to evasion, focusing
particularly on anti-ransomware evasion. We choose ransomware as its behavior
tends to differ significantly from that of benign processes, making it a
low-hanging fruit for behavioral detection (and a difficult candidate for
evasion). Our analysis identifies a set of novel attacks that distribute the
overall malware workload across a small set of cooperating processes to avoid
the generation of significant behavioral features. Our most effective attack
decreases the accuracy of a state-of-the-art classifier from 98.6% to 0% using
only 18 cooperating processes. Furthermore, we show our attacks to be effective
against commercial ransomware detectors even in a black-box setting.
| [
{
"created": "Wed, 6 Nov 2019 15:04:07 GMT",
"version": "v1"
}
] | 2019-11-07 | [
[
"De Gaspari",
"Fabio",
""
],
[
"Hitaj",
"Dorjan",
""
],
[
"Pagnotta",
"Giulio",
""
],
[
"De Carli",
"Lorenzo",
""
],
[
"Mancini",
"Luigi V.",
""
]
] | Recent progress in machine learning has generated promising results in behavioral malware detection. Behavioral modeling identifies malicious processes via features derived by their runtime behavior. Behavioral features hold great promise as they are intrinsically related to the functioning of each malware, and are therefore considered difficult to evade. Indeed, while a significant amount of results exists on evasion of static malware features, evasion of dynamic features has seen limited work. This paper thoroughly examines the robustness of behavioral malware detectors to evasion, focusing particularly on anti-ransomware evasion. We choose ransomware as its behavior tends to differ significantly from that of benign processes, making it a low-hanging fruit for behavioral detection (and a difficult candidate for evasion). Our analysis identifies a set of novel attacks that distribute the overall malware workload across a small set of cooperating processes to avoid the generation of significant behavioral features. Our most effective attack decreases the accuracy of a state-of-the-art classifier from 98.6% to 0% using only 18 cooperating processes. Furthermore, we show our attacks to be effective against commercial ransomware detectors even in a black-box setting. |
cs/0609009 | Virginia Vassilevska | Virginia Vassilevska, Ryan Williams and Raphael Yuster | Finding heaviest H-subgraphs in real weighted graphs, with applications | 23 pages | null | null | null | cs.DS cs.DM | null | For a graph G with real weights assigned to the vertices (edges), the MAX
H-SUBGRAPH problem is to find an H-subgraph of G with maximum total weight, if
one exists. The all-pairs MAX H-SUBGRAPH problem is to find for every pair of
vertices u,v, a maximum H-subgraph containing both u and v, if one exists. Our
main results are new strongly polynomial algorithms for the all-pairs MAX
H-SUBGRAPH problem for vertex weighted graphs. We also give improved algorithms
for the MAX-H SUBGRAPH problem for edge weighted graphs, and various related
problems, including computing the first k most significant bits of the distance
product of two matrices. Some of our algorithms are based, in part, on fast
matrix multiplication.
| [
{
"created": "Mon, 4 Sep 2006 08:08:00 GMT",
"version": "v1"
}
] | 2007-05-23 | [
[
"Vassilevska",
"Virginia",
""
],
[
"Williams",
"Ryan",
""
],
[
"Yuster",
"Raphael",
""
]
] | For a graph G with real weights assigned to the vertices (edges), the MAX H-SUBGRAPH problem is to find an H-subgraph of G with maximum total weight, if one exists. The all-pairs MAX H-SUBGRAPH problem is to find for every pair of vertices u,v, a maximum H-subgraph containing both u and v, if one exists. Our main results are new strongly polynomial algorithms for the all-pairs MAX H-SUBGRAPH problem for vertex weighted graphs. We also give improved algorithms for the MAX-H SUBGRAPH problem for edge weighted graphs, and various related problems, including computing the first k most significant bits of the distance product of two matrices. Some of our algorithms are based, in part, on fast matrix multiplication. |
0912.1216 | Ying Cui | Ying Cui, Vincent K.N.Lau and Rui Wang | Distributive Subband Allocation, Power and Rate Control for
Relay-Assisted OFDMA Cellular System with Imperfect System State Knowledge | 11 pages, 8 figures | null | null | null | cs.NI cs.IT math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we consider distributive subband, power and rate allocation
for a two-hop transmission in an orthogonal frequency-division multiple-access
(OFDMA) cellular system with fixed relays which operate in decode-and-forward
strategy. We take into account of system fairness by considering weighted sum
goodput as our optimization objective. Based on the cluster-based architecture,
we obtain a fast-converging distributive solution with only local imperfect
CSIT by using decomposition of the optimization problem. To further reduce the
signaling overhead and computational complexity, we propose a reduced feedback
distributive solution, which can achieve asymptotically optimal performance for
large number of users with arbitrarily small feedback overhead per user. We
also derive asymptotic average system throughput for the relay-assisted OFDMA
system so as to obtain useful design insights.
| [
{
"created": "Mon, 7 Dec 2009 12:30:57 GMT",
"version": "v1"
}
] | 2009-12-08 | [
[
"Cui",
"Ying",
""
],
[
"Lau",
"Vincent K. N.",
""
],
[
"Wang",
"Rui",
""
]
] | In this paper, we consider distributive subband, power and rate allocation for a two-hop transmission in an orthogonal frequency-division multiple-access (OFDMA) cellular system with fixed relays which operate in decode-and-forward strategy. We take into account of system fairness by considering weighted sum goodput as our optimization objective. Based on the cluster-based architecture, we obtain a fast-converging distributive solution with only local imperfect CSIT by using decomposition of the optimization problem. To further reduce the signaling overhead and computational complexity, we propose a reduced feedback distributive solution, which can achieve asymptotically optimal performance for large number of users with arbitrarily small feedback overhead per user. We also derive asymptotic average system throughput for the relay-assisted OFDMA system so as to obtain useful design insights. |
1810.03717 | Judy Hanwen Shen | Judy Hanwen Shen, Matthias Hofer, Bjarke Felbo, Roger Levy | Comparing Models of Associative Meaning: An Empirical Investigation of
Reference in Simple Language Games | Conference on Computational Natural Language Learning (CoNLL) 2018 | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Simple reference games are of central theoretical and empirical importance in
the study of situated language use. Although language provides rich,
compositional truth-conditional semantics to facilitate reference, speakers and
listeners may sometimes lack the overall lexical and cognitive resources to
guarantee successful reference through these means alone. However, language
also has rich associational structures that can serve as a further resource for
achieving successful reference. Here we investigate this use of associational
information in a setting where only associational information is available: a
simplified version of the popular game Codenames. Using optimal experiment
design techniques, we compare a range of models varying in the type of
associative information deployed and in level of pragmatic sophistication
against human behavior. In this setting, we find that listeners' behavior
reflects direct bigram collocational associations more strongly than
word-embedding or semantic knowledge graph-based associations and that there is
little evidence for pragmatically sophisticated behavior by either speakers or
listeners of the type that might be predicted by recursive-reasoning models
such as the Rational Speech Acts theory. These results shed light on the nature
of the lexical resources that speakers and listeners can bring to bear in
achieving reference through associative meaning alone.
| [
{
"created": "Mon, 8 Oct 2018 21:51:44 GMT",
"version": "v1"
}
] | 2018-10-10 | [
[
"Shen",
"Judy Hanwen",
""
],
[
"Hofer",
"Matthias",
""
],
[
"Felbo",
"Bjarke",
""
],
[
"Levy",
"Roger",
""
]
] | Simple reference games are of central theoretical and empirical importance in the study of situated language use. Although language provides rich, compositional truth-conditional semantics to facilitate reference, speakers and listeners may sometimes lack the overall lexical and cognitive resources to guarantee successful reference through these means alone. However, language also has rich associational structures that can serve as a further resource for achieving successful reference. Here we investigate this use of associational information in a setting where only associational information is available: a simplified version of the popular game Codenames. Using optimal experiment design techniques, we compare a range of models varying in the type of associative information deployed and in level of pragmatic sophistication against human behavior. In this setting, we find that listeners' behavior reflects direct bigram collocational associations more strongly than word-embedding or semantic knowledge graph-based associations and that there is little evidence for pragmatically sophisticated behavior by either speakers or listeners of the type that might be predicted by recursive-reasoning models such as the Rational Speech Acts theory. These results shed light on the nature of the lexical resources that speakers and listeners can bring to bear in achieving reference through associative meaning alone. |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.