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
2111.03278
Eric Neyman
Rafael Frongillo, Eric Neyman, Bo Waggoner
Agreement Implies Accuracy for Substitutable Signals
31 pages, 1 figure
null
null
null
cs.GT
http://creativecommons.org/licenses/by/4.0/
Inspired by Aumann's agreement theorem, Scott Aaronson studied the amount of communication necessary for two Bayesian experts to approximately agree on the expectation of a random variable. Aaronson showed that, remarkably, the number of bits does not depend on the amount of information available to each expert. However, in general the agreed-upon estimate may be inaccurate: far from the estimate they would settle on if they were to share all of their information. We show that if the experts' signals are \emph{substitutes} -- meaning the experts' information has diminishing marginal returns -- then it is the case that if the experts are close to agreement then they are close to the truth. We prove this result for a broad class of agreement and accuracy measures that includes squared distance and KL divergence. Additionally, we show that although these measures capture fundamentally different kinds of agreement, Aaronson's agreement result generalizes to them as well.
[ { "created": "Fri, 5 Nov 2021 05:47:03 GMT", "version": "v1" } ]
2021-11-08
[ [ "Frongillo", "Rafael", "" ], [ "Neyman", "Eric", "" ], [ "Waggoner", "Bo", "" ] ]
Inspired by Aumann's agreement theorem, Scott Aaronson studied the amount of communication necessary for two Bayesian experts to approximately agree on the expectation of a random variable. Aaronson showed that, remarkably, the number of bits does not depend on the amount of information available to each expert. However, in general the agreed-upon estimate may be inaccurate: far from the estimate they would settle on if they were to share all of their information. We show that if the experts' signals are \emph{substitutes} -- meaning the experts' information has diminishing marginal returns -- then it is the case that if the experts are close to agreement then they are close to the truth. We prove this result for a broad class of agreement and accuracy measures that includes squared distance and KL divergence. Additionally, we show that although these measures capture fundamentally different kinds of agreement, Aaronson's agreement result generalizes to them as well.
2404.00466
Heqiang Wang
Heqiang Wang, Jieming Bian, Lei Wang
Computation and Communication Efficient Lightweighting Vertical Federated Learning
null
null
null
null
cs.LG cs.DC
http://creativecommons.org/licenses/by/4.0/
The exploration of computational and communication efficiency within Federated Learning (FL) has emerged as a prominent and crucial field of study. While most existing efforts to enhance these efficiencies have focused on Horizontal FL, the distinct processes and model structures of Vertical FL preclude the direct application of Horizontal FL-based techniques. In response, we introduce the concept of Lightweight Vertical Federated Learning (LVFL), targeting both computational and communication efficiencies. This approach involves separate lightweighting strategies for the feature model, to improve computational efficiency, and for feature embedding, to enhance communication efficiency. Moreover, we establish a convergence bound for our LVFL algorithm, which accounts for both communication and computational lightweighting ratios. Our evaluation of the algorithm on a image classification dataset reveals that LVFL significantly alleviates computational and communication demands while preserving robust learning performance. This work effectively addresses the gaps in communication and computational efficiency within Vertical FL.
[ { "created": "Sat, 30 Mar 2024 20:19:28 GMT", "version": "v1" } ]
2024-04-02
[ [ "Wang", "Heqiang", "" ], [ "Bian", "Jieming", "" ], [ "Wang", "Lei", "" ] ]
The exploration of computational and communication efficiency within Federated Learning (FL) has emerged as a prominent and crucial field of study. While most existing efforts to enhance these efficiencies have focused on Horizontal FL, the distinct processes and model structures of Vertical FL preclude the direct application of Horizontal FL-based techniques. In response, we introduce the concept of Lightweight Vertical Federated Learning (LVFL), targeting both computational and communication efficiencies. This approach involves separate lightweighting strategies for the feature model, to improve computational efficiency, and for feature embedding, to enhance communication efficiency. Moreover, we establish a convergence bound for our LVFL algorithm, which accounts for both communication and computational lightweighting ratios. Our evaluation of the algorithm on a image classification dataset reveals that LVFL significantly alleviates computational and communication demands while preserving robust learning performance. This work effectively addresses the gaps in communication and computational efficiency within Vertical FL.
1106.1910
Vahid Majid Nezhad
Vahid Majid Nezhad, Habib Motee Gader and Evgueni Efimov
A Novel Hybrid Algorithm for Task Graph Scheduling
null
IJCSI, Vol 8, Issue 2, March 2011, p32-38
null
null
cs.CC
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
One of the important problems in multiprocessor systems is Task Graph Scheduling. Task Graph Scheduling is an NP-Hard problem. Both learning automata and genetic algorithms are search tools which are used for solving many NP-Hard problems. In this paper a new hybrid method based on Genetic Algorithm and Learning Automata is proposed. The proposed algorithm begins with an initial population of randomly generated chromosomes and after some stages, each chromosome maps to an automaton. Experimental results show that superiority of the proposed algorithm over the current approaches.
[ { "created": "Thu, 9 Jun 2011 20:26:22 GMT", "version": "v1" } ]
2011-06-13
[ [ "Nezhad", "Vahid Majid", "" ], [ "Gader", "Habib Motee", "" ], [ "Efimov", "Evgueni", "" ] ]
One of the important problems in multiprocessor systems is Task Graph Scheduling. Task Graph Scheduling is an NP-Hard problem. Both learning automata and genetic algorithms are search tools which are used for solving many NP-Hard problems. In this paper a new hybrid method based on Genetic Algorithm and Learning Automata is proposed. The proposed algorithm begins with an initial population of randomly generated chromosomes and after some stages, each chromosome maps to an automaton. Experimental results show that superiority of the proposed algorithm over the current approaches.
1909.02273
Chengyi Wang
Chengyi Wang, Shuangzhi Wu, Shujie Liu
Source Dependency-Aware Transformer with Supervised Self-Attention
6 pages
null
null
null
cs.CL
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Recently, Transformer has achieved the state-of-the-art performance on many machine translation tasks. However, without syntax knowledge explicitly considered in the encoder, incorrect context information that violates the syntax structure may be integrated into source hidden states, leading to erroneous translations. In this paper, we propose a novel method to incorporate source dependencies into the Transformer. Specifically, we adopt the source dependency tree and define two matrices to represent the dependency relations. Based on the matrices, two heads in the multi-head self-attention module are trained in a supervised manner and two extra cross entropy losses are introduced into the training objective function. Under this training objective, the model is trained to learn the source dependency relations directly. Without requiring pre-parsed input during inference, our model can generate better translations with the dependency-aware context information. Experiments on bi-directional Chinese-to-English, English-to-Japanese and English-to-German translation tasks show that our proposed method can significantly improve the Transformer baseline.
[ { "created": "Thu, 5 Sep 2019 09:17:37 GMT", "version": "v1" } ]
2019-09-06
[ [ "Wang", "Chengyi", "" ], [ "Wu", "Shuangzhi", "" ], [ "Liu", "Shujie", "" ] ]
Recently, Transformer has achieved the state-of-the-art performance on many machine translation tasks. However, without syntax knowledge explicitly considered in the encoder, incorrect context information that violates the syntax structure may be integrated into source hidden states, leading to erroneous translations. In this paper, we propose a novel method to incorporate source dependencies into the Transformer. Specifically, we adopt the source dependency tree and define two matrices to represent the dependency relations. Based on the matrices, two heads in the multi-head self-attention module are trained in a supervised manner and two extra cross entropy losses are introduced into the training objective function. Under this training objective, the model is trained to learn the source dependency relations directly. Without requiring pre-parsed input during inference, our model can generate better translations with the dependency-aware context information. Experiments on bi-directional Chinese-to-English, English-to-Japanese and English-to-German translation tasks show that our proposed method can significantly improve the Transformer baseline.
1607.01284
Ruifeng Duan
Ruifeng Duan, Riku J\"antti, H\"useyin Yi\u{g}itler, and Kalle Ruttik
On the Achievable Rate of Bi-Static Modulated Re-Scatter Systems
5 pages, 3 figures, accepted
null
null
null
cs.IT math.IT
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
In ambient re-scatter communications, devices convey information by modulating and re-scattering the radio frequency signals impinging on their antennas. In this correspondence, we consider a system consisting of a legacy modulated continuous carrier multiple-input-multiple-output (MIMO) link and a multi-antenna modulated re-scatter (MRS) node, where the MRS node modulates and re-scatters the signal generated by the legacy transmitter. The receiver seeks to decode both the original message and the information added by the MRS. We show that the achievable sum rate of this system exceeds that which the legacy system could achieve alone. We further consider the impact of channel estimation errors under the least squares channel estimation and study the achievable rate of the legacy and MRS systems, where a linear minimum mean square error receiver with successive interference cancellation is utilized for joint decoding.
[ { "created": "Tue, 5 Jul 2016 14:54:14 GMT", "version": "v1" }, { "created": "Mon, 12 Jun 2017 13:19:10 GMT", "version": "v2" } ]
2017-06-13
[ [ "Duan", "Ruifeng", "" ], [ "Jäntti", "Riku", "" ], [ "Yiğitler", "Hüseyin", "" ], [ "Ruttik", "Kalle", "" ] ]
In ambient re-scatter communications, devices convey information by modulating and re-scattering the radio frequency signals impinging on their antennas. In this correspondence, we consider a system consisting of a legacy modulated continuous carrier multiple-input-multiple-output (MIMO) link and a multi-antenna modulated re-scatter (MRS) node, where the MRS node modulates and re-scatters the signal generated by the legacy transmitter. The receiver seeks to decode both the original message and the information added by the MRS. We show that the achievable sum rate of this system exceeds that which the legacy system could achieve alone. We further consider the impact of channel estimation errors under the least squares channel estimation and study the achievable rate of the legacy and MRS systems, where a linear minimum mean square error receiver with successive interference cancellation is utilized for joint decoding.
1405.5202
Altaf Rahman
Altaf Rahman, Vincent Ng
Narrowing the Modeling Gap: A Cluster-Ranking Approach to Coreference Resolution
null
Journal Of Artificial Intelligence Research, Volume 40, pages 469-521, 2011
10.1613/jair.3120
null
cs.CL
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Traditional learning-based coreference resolvers operate by training the mention-pair model for determining whether two mentions are coreferent or not. Though conceptually simple and easy to understand, the mention-pair model is linguistically rather unappealing and lags far behind the heuristic-based coreference models proposed in the pre-statistical NLP era in terms of sophistication. Two independent lines of recent research have attempted to improve the mention-pair model, one by acquiring the mention-ranking model to rank preceding mentions for a given anaphor, and the other by training the entity-mention model to determine whether a preceding cluster is coreferent with a given mention. We propose a cluster-ranking approach to coreference resolution, which combines the strengths of the mention-ranking model and the entity-mention model, and is therefore theoretically more appealing than both of these models. In addition, we seek to improve cluster rankers via two extensions: (1) lexicalization and (2) incorporating knowledge of anaphoricity by jointly modeling anaphoricity determination and coreference resolution. Experimental results on the ACE data sets demonstrate the superior performance of cluster rankers to competing approaches as well as the effectiveness of our two extensions.
[ { "created": "Thu, 16 Jan 2014 05:06:09 GMT", "version": "v1" } ]
2014-05-21
[ [ "Rahman", "Altaf", "" ], [ "Ng", "Vincent", "" ] ]
Traditional learning-based coreference resolvers operate by training the mention-pair model for determining whether two mentions are coreferent or not. Though conceptually simple and easy to understand, the mention-pair model is linguistically rather unappealing and lags far behind the heuristic-based coreference models proposed in the pre-statistical NLP era in terms of sophistication. Two independent lines of recent research have attempted to improve the mention-pair model, one by acquiring the mention-ranking model to rank preceding mentions for a given anaphor, and the other by training the entity-mention model to determine whether a preceding cluster is coreferent with a given mention. We propose a cluster-ranking approach to coreference resolution, which combines the strengths of the mention-ranking model and the entity-mention model, and is therefore theoretically more appealing than both of these models. In addition, we seek to improve cluster rankers via two extensions: (1) lexicalization and (2) incorporating knowledge of anaphoricity by jointly modeling anaphoricity determination and coreference resolution. Experimental results on the ACE data sets demonstrate the superior performance of cluster rankers to competing approaches as well as the effectiveness of our two extensions.
1412.2716
Paul Ginsparg
Daniel T. Citron and Paul Ginsparg
Patterns of Text Reuse in a Scientific Corpus
6 pages, plus 10 pages of supplementary material. To appear in PNAS (online 8 Dec 2014)
null
10.1073/pnas.1415135111
null
cs.DL physics.soc-ph
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
We consider the incidence of text "reuse" by researchers, via a systematic pairwise comparison of the text content of all articles deposited to arXiv.org from 1991--2012. We measure the global frequencies of three classes of text reuse, and measure how chronic text reuse is distributed among authors in the dataset. We infer a baseline for accepted practice, perhaps surprisingly permissive compared with other societal contexts, and a clearly delineated set of aberrant authors. We find a negative correlation between the amount of reused text in an article and its influence, as measured by subsequent citations. Finally, we consider the distribution of countries of origin of articles containing large amounts of reused text.
[ { "created": "Mon, 8 Dec 2014 20:01:17 GMT", "version": "v1" } ]
2014-12-09
[ [ "Citron", "Daniel T.", "" ], [ "Ginsparg", "Paul", "" ] ]
We consider the incidence of text "reuse" by researchers, via a systematic pairwise comparison of the text content of all articles deposited to arXiv.org from 1991--2012. We measure the global frequencies of three classes of text reuse, and measure how chronic text reuse is distributed among authors in the dataset. We infer a baseline for accepted practice, perhaps surprisingly permissive compared with other societal contexts, and a clearly delineated set of aberrant authors. We find a negative correlation between the amount of reused text in an article and its influence, as measured by subsequent citations. Finally, we consider the distribution of countries of origin of articles containing large amounts of reused text.
2205.11107
Lara Scavuzzo
Lara Scavuzzo and Feng Yang Chen and Didier Ch\'etelat and Maxime Gasse and Andrea Lodi and Neil Yorke-Smith and Karen Aardal
Learning to branch with Tree MDPs
10 pages, 2 figures, plus supplementary material
null
null
null
cs.LG math.OC
http://creativecommons.org/licenses/by-nc-sa/4.0/
State-of-the-art Mixed Integer Linear Program (MILP) solvers combine systematic tree search with a plethora of hard-coded heuristics, such as the branching rule. The idea of learning branching rules from data has received increasing attention recently, and promising results have been obtained by learning fast approximations of the strong branching expert. In this work, we instead propose to learn branching rules from scratch via Reinforcement Learning (RL). We revisit the work of Etheve et al. (2020) and propose tree Markov Decision Processes, or tree MDPs, a generalization of temporal MDPs that provides a more suitable framework for learning to branch. We derive a tree policy gradient theorem, which exhibits a better credit assignment compared to its temporal counterpart. We demonstrate through computational experiments that tree MDPs improve the learning convergence, and offer a promising framework for tackling the learning-to-branch problem in MILPs.
[ { "created": "Mon, 23 May 2022 07:57:32 GMT", "version": "v1" }, { "created": "Tue, 31 May 2022 11:05:56 GMT", "version": "v2" }, { "created": "Thu, 13 Oct 2022 13:37:42 GMT", "version": "v3" } ]
2022-10-14
[ [ "Scavuzzo", "Lara", "" ], [ "Chen", "Feng Yang", "" ], [ "Chételat", "Didier", "" ], [ "Gasse", "Maxime", "" ], [ "Lodi", "Andrea", "" ], [ "Yorke-Smith", "Neil", "" ], [ "Aardal", "Karen", "" ] ]
State-of-the-art Mixed Integer Linear Program (MILP) solvers combine systematic tree search with a plethora of hard-coded heuristics, such as the branching rule. The idea of learning branching rules from data has received increasing attention recently, and promising results have been obtained by learning fast approximations of the strong branching expert. In this work, we instead propose to learn branching rules from scratch via Reinforcement Learning (RL). We revisit the work of Etheve et al. (2020) and propose tree Markov Decision Processes, or tree MDPs, a generalization of temporal MDPs that provides a more suitable framework for learning to branch. We derive a tree policy gradient theorem, which exhibits a better credit assignment compared to its temporal counterpart. We demonstrate through computational experiments that tree MDPs improve the learning convergence, and offer a promising framework for tackling the learning-to-branch problem in MILPs.
2209.12413
P B Sujit Dr
Kasi Vishwanath, P.B. Sujit and Srikanth Saripalli
CAMEL: Learning Cost-maps Made Easy for Off-road Driving
null
null
null
null
cs.RO cs.CV
http://creativecommons.org/licenses/by/4.0/
Cost-maps are used by robotic vehicles to plan collision-free paths. The cost associated with each cell in the map represents the sensed environment information which is often determined manually after several trial-and-error efforts. In off-road environments, due to the presence of several types of features, it is challenging to handcraft the cost values associated with each feature. Moreover, different handcrafted cost values can lead to different paths for the same environment which is not desirable. In this paper, we address the problem of learning the cost-map values from the sensed environment for robust vehicle path planning. We propose a novel framework called as CAMEL using deep learning approach that learns the parameters through demonstrations yielding an adaptive and robust cost-map for path planning. CAMEL has been trained on multi-modal datasets such as RELLIS-3D. The evaluation of CAMEL is carried out on an off-road scene simulator (MAVS) and on field data from IISER-B campus. We also perform realworld implementation of CAMEL on a ground rover. The results shows flexible and robust motion of the vehicle without collisions in unstructured terrains.
[ { "created": "Mon, 26 Sep 2022 04:37:03 GMT", "version": "v1" }, { "created": "Tue, 18 Oct 2022 08:07:59 GMT", "version": "v2" } ]
2022-10-19
[ [ "Vishwanath", "Kasi", "" ], [ "Sujit", "P. B.", "" ], [ "Saripalli", "Srikanth", "" ] ]
Cost-maps are used by robotic vehicles to plan collision-free paths. The cost associated with each cell in the map represents the sensed environment information which is often determined manually after several trial-and-error efforts. In off-road environments, due to the presence of several types of features, it is challenging to handcraft the cost values associated with each feature. Moreover, different handcrafted cost values can lead to different paths for the same environment which is not desirable. In this paper, we address the problem of learning the cost-map values from the sensed environment for robust vehicle path planning. We propose a novel framework called as CAMEL using deep learning approach that learns the parameters through demonstrations yielding an adaptive and robust cost-map for path planning. CAMEL has been trained on multi-modal datasets such as RELLIS-3D. The evaluation of CAMEL is carried out on an off-road scene simulator (MAVS) and on field data from IISER-B campus. We also perform realworld implementation of CAMEL on a ground rover. The results shows flexible and robust motion of the vehicle without collisions in unstructured terrains.
1303.2764
Lin Na
Na Lin, Hong-Dong Liu, Chang-Qing Gong
Research and Simulation on Drivers' Route Choice Behavior Cognition Model
6 pages,8 figures,a table
null
null
null
cs.NI
http://creativecommons.org/licenses/by-nc-sa/3.0/
This paper studied the behavior-cognitive model of drivers during their travel based on the current research on driver behavior. Firstly, a route choice behavior-cognitive model was proposed for describing the decision-making mechanism of drivers during his travel; then, simulation experiments were carried out on the cosimulation VBc-vissim platform. From the experimental results, dynamic behavior features of drivers during their travel can be properly explained by the behavior-cognitive model, thus optimal path can be obtained from this model.
[ { "created": "Tue, 12 Mar 2013 02:59:15 GMT", "version": "v1" } ]
2013-03-13
[ [ "Lin", "Na", "" ], [ "Liu", "Hong-Dong", "" ], [ "Gong", "Chang-Qing", "" ] ]
This paper studied the behavior-cognitive model of drivers during their travel based on the current research on driver behavior. Firstly, a route choice behavior-cognitive model was proposed for describing the decision-making mechanism of drivers during his travel; then, simulation experiments were carried out on the cosimulation VBc-vissim platform. From the experimental results, dynamic behavior features of drivers during their travel can be properly explained by the behavior-cognitive model, thus optimal path can be obtained from this model.
1609.00559
Ted Pedersen
Bridget T. McInnes and Ted Pedersen
Improving Correlation with Human Judgments by Integrating Semantic Similarity with Second--Order Vectors
10 pages, Appears in the Proceedings of the 16th Workshop on Biomedical Natural Language Processing (BioNLP-2017), August 2017, Vancouver, BC
null
null
null
cs.CL
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Vector space methods that measure semantic similarity and relatedness often rely on distributional information such as co--occurrence frequencies or statistical measures of association to weight the importance of particular co--occurrences. In this paper, we extend these methods by incorporating a measure of semantic similarity based on a human curated taxonomy into a second--order vector representation. This results in a measure of semantic relatedness that combines both the contextual information available in a corpus--based vector space representation with the semantic knowledge found in a biomedical ontology. Our results show that incorporating semantic similarity into a second order co--occurrence matrices improves correlation with human judgments for both similarity and relatedness, and that our method compares favorably to various different word embedding methods that have recently been evaluated on the same reference standards we have used.
[ { "created": "Fri, 2 Sep 2016 11:44:17 GMT", "version": "v1" }, { "created": "Sat, 27 May 2017 00:23:06 GMT", "version": "v2" } ]
2017-05-30
[ [ "McInnes", "Bridget T.", "" ], [ "Pedersen", "Ted", "" ] ]
Vector space methods that measure semantic similarity and relatedness often rely on distributional information such as co--occurrence frequencies or statistical measures of association to weight the importance of particular co--occurrences. In this paper, we extend these methods by incorporating a measure of semantic similarity based on a human curated taxonomy into a second--order vector representation. This results in a measure of semantic relatedness that combines both the contextual information available in a corpus--based vector space representation with the semantic knowledge found in a biomedical ontology. Our results show that incorporating semantic similarity into a second order co--occurrence matrices improves correlation with human judgments for both similarity and relatedness, and that our method compares favorably to various different word embedding methods that have recently been evaluated on the same reference standards we have used.
2301.11716
Phuong-Hang Le
Phuong-Hang Le, Hongyu Gong, Changhan Wang, Juan Pino, Benjamin Lecouteux, Didier Schwab
Pre-training for Speech Translation: CTC Meets Optimal Transport
ICML 2023 (oral presentation). This version fixed URLs, updated affiliations & acknowledgements, and improved formatting
null
null
null
cs.CL cs.LG cs.SD eess.AS
http://creativecommons.org/licenses/by/4.0/
The gap between speech and text modalities is a major challenge in speech-to-text translation (ST). Different methods have been proposed to reduce this gap, but most of them require architectural changes in ST training. In this work, we propose to mitigate this issue at the pre-training stage, requiring no change in the ST model. First, we show that the connectionist temporal classification (CTC) loss can reduce the modality gap by design. We provide a quantitative comparison with the more common cross-entropy loss, showing that pre-training with CTC consistently achieves better final ST accuracy. Nevertheless, CTC is only a partial solution and thus, in our second contribution, we propose a novel pre-training method combining CTC and optimal transport to further reduce this gap. Our method pre-trains a Siamese-like model composed of two encoders, one for acoustic inputs and the other for textual inputs, such that they produce representations that are close to each other in the Wasserstein space. Extensive experiments on the standard CoVoST-2 and MuST-C datasets show that our pre-training method applied to the vanilla encoder-decoder Transformer achieves state-of-the-art performance under the no-external-data setting, and performs on par with recent strong multi-task learning systems trained with external data. Finally, our method can also be applied on top of these multi-task systems, leading to further improvements for these models. Code and pre-trained models are available at https://github.com/formiel/fairseq.
[ { "created": "Fri, 27 Jan 2023 14:03:09 GMT", "version": "v1" }, { "created": "Tue, 30 May 2023 09:06:22 GMT", "version": "v2" }, { "created": "Mon, 5 Jun 2023 11:44:02 GMT", "version": "v3" } ]
2023-06-06
[ [ "Le", "Phuong-Hang", "" ], [ "Gong", "Hongyu", "" ], [ "Wang", "Changhan", "" ], [ "Pino", "Juan", "" ], [ "Lecouteux", "Benjamin", "" ], [ "Schwab", "Didier", "" ] ]
The gap between speech and text modalities is a major challenge in speech-to-text translation (ST). Different methods have been proposed to reduce this gap, but most of them require architectural changes in ST training. In this work, we propose to mitigate this issue at the pre-training stage, requiring no change in the ST model. First, we show that the connectionist temporal classification (CTC) loss can reduce the modality gap by design. We provide a quantitative comparison with the more common cross-entropy loss, showing that pre-training with CTC consistently achieves better final ST accuracy. Nevertheless, CTC is only a partial solution and thus, in our second contribution, we propose a novel pre-training method combining CTC and optimal transport to further reduce this gap. Our method pre-trains a Siamese-like model composed of two encoders, one for acoustic inputs and the other for textual inputs, such that they produce representations that are close to each other in the Wasserstein space. Extensive experiments on the standard CoVoST-2 and MuST-C datasets show that our pre-training method applied to the vanilla encoder-decoder Transformer achieves state-of-the-art performance under the no-external-data setting, and performs on par with recent strong multi-task learning systems trained with external data. Finally, our method can also be applied on top of these multi-task systems, leading to further improvements for these models. Code and pre-trained models are available at https://github.com/formiel/fairseq.
1804.07663
Andreas Steyven
Andreas Steyven, Emma Hart, Ben Paechter
An Investigation of Environmental Influence on the Benefits of Adaptation Mechanisms in Evolutionary Swarm Robotics
In GECCO 2017
null
10.1145/3071178.3071232
null
cs.NE
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
A robotic swarm that is required to operate for long periods in a potentially unknown environment can use both evolution and individual learning methods in order to adapt. However, the role played by the environment in influencing the effectiveness of each type of learning is not well understood. In this paper, we address this question by analysing the performance of a swarm in a range of simulated, dynamic environments where a distributed evolutionary algorithm for evolving a controller is augmented with a number of different individual learning mechanisms. The learning mechanisms themselves are defined by parameters which can be either fixed or inherited. We conduct experiments in a range of dynamic environments whose characteristics are varied so as to present different opportunities for learning. Results enable us to map environmental characteristics to the most effective learning algorithm.
[ { "created": "Fri, 20 Apr 2018 15:13:47 GMT", "version": "v1" } ]
2018-04-23
[ [ "Steyven", "Andreas", "" ], [ "Hart", "Emma", "" ], [ "Paechter", "Ben", "" ] ]
A robotic swarm that is required to operate for long periods in a potentially unknown environment can use both evolution and individual learning methods in order to adapt. However, the role played by the environment in influencing the effectiveness of each type of learning is not well understood. In this paper, we address this question by analysing the performance of a swarm in a range of simulated, dynamic environments where a distributed evolutionary algorithm for evolving a controller is augmented with a number of different individual learning mechanisms. The learning mechanisms themselves are defined by parameters which can be either fixed or inherited. We conduct experiments in a range of dynamic environments whose characteristics are varied so as to present different opportunities for learning. Results enable us to map environmental characteristics to the most effective learning algorithm.
1306.0322
Hector Zenil
Hector Zenil, Fernando Soler-Toscano, Kamaludin Dingle and Ard A. Louis
Correlation of Automorphism Group Size and Topological Properties with Program-size Complexity Evaluations of Graphs and Complex Networks
15 2-column pages, 20 figures. Forthcoming in Physica A: Statistical Mechanics and its Applications
null
10.1016/j.physa.2014.02.060
null
cs.IT cs.CC cs.CG math.IT q-bio.MN
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
We show that numerical approximations of Kolmogorov complexity (K) applied to graph adjacency matrices capture some group-theoretic and topological properties of graphs and empirical networks ranging from metabolic to social networks. That K and the size of the group of automorphisms of a graph are correlated opens up interesting connections to problems in computational geometry, and thus connects several measures and concepts from complexity science. We show that approximations of K characterise synthetic and natural networks by their generating mechanisms, assigning lower algorithmic randomness to complex network models (Watts-Strogatz and Barabasi-Albert networks) and high Kolmogorov complexity to (random) Erdos-Renyi graphs. We derive these results via two different Kolmogorov complexity approximation methods applied to the adjacency matrices of the graphs and networks. The methods used are the traditional lossless compression approach to Kolmogorov complexity, and a normalised version of a Block Decomposition Method (BDM) measure, based on algorithmic probability theory.
[ { "created": "Mon, 3 Jun 2013 08:36:11 GMT", "version": "v1" }, { "created": "Mon, 17 Jun 2013 11:32:00 GMT", "version": "v2" }, { "created": "Sun, 23 Feb 2014 01:42:27 GMT", "version": "v3" } ]
2015-06-16
[ [ "Zenil", "Hector", "" ], [ "Soler-Toscano", "Fernando", "" ], [ "Dingle", "Kamaludin", "" ], [ "Louis", "Ard A.", "" ] ]
We show that numerical approximations of Kolmogorov complexity (K) applied to graph adjacency matrices capture some group-theoretic and topological properties of graphs and empirical networks ranging from metabolic to social networks. That K and the size of the group of automorphisms of a graph are correlated opens up interesting connections to problems in computational geometry, and thus connects several measures and concepts from complexity science. We show that approximations of K characterise synthetic and natural networks by their generating mechanisms, assigning lower algorithmic randomness to complex network models (Watts-Strogatz and Barabasi-Albert networks) and high Kolmogorov complexity to (random) Erdos-Renyi graphs. We derive these results via two different Kolmogorov complexity approximation methods applied to the adjacency matrices of the graphs and networks. The methods used are the traditional lossless compression approach to Kolmogorov complexity, and a normalised version of a Block Decomposition Method (BDM) measure, based on algorithmic probability theory.
2307.01831
Shentong Mo
Shentong Mo, Enze Xie, Ruihang Chu, Lewei Yao, Lanqing Hong, Matthias Nie{\ss}ner, Zhenguo Li
DiT-3D: Exploring Plain Diffusion Transformers for 3D Shape Generation
Project Page: https://dit-3d.github.io/
null
null
null
cs.CV cs.AI cs.LG
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Recent Diffusion Transformers (e.g., DiT) have demonstrated their powerful effectiveness in generating high-quality 2D images. However, it is still being determined whether the Transformer architecture performs equally well in 3D shape generation, as previous 3D diffusion methods mostly adopted the U-Net architecture. To bridge this gap, we propose a novel Diffusion Transformer for 3D shape generation, namely DiT-3D, which can directly operate the denoising process on voxelized point clouds using plain Transformers. Compared to existing U-Net approaches, our DiT-3D is more scalable in model size and produces much higher quality generations. Specifically, the DiT-3D adopts the design philosophy of DiT but modifies it by incorporating 3D positional and patch embeddings to adaptively aggregate input from voxelized point clouds. To reduce the computational cost of self-attention in 3D shape generation, we incorporate 3D window attention into Transformer blocks, as the increased 3D token length resulting from the additional dimension of voxels can lead to high computation. Finally, linear and devoxelization layers are used to predict the denoised point clouds. In addition, our transformer architecture supports efficient fine-tuning from 2D to 3D, where the pre-trained DiT-2D checkpoint on ImageNet can significantly improve DiT-3D on ShapeNet. Experimental results on the ShapeNet dataset demonstrate that the proposed DiT-3D achieves state-of-the-art performance in high-fidelity and diverse 3D point cloud generation. In particular, our DiT-3D decreases the 1-Nearest Neighbor Accuracy of the state-of-the-art method by 4.59 and increases the Coverage metric by 3.51 when evaluated on Chamfer Distance.
[ { "created": "Tue, 4 Jul 2023 17:15:46 GMT", "version": "v1" } ]
2023-07-06
[ [ "Mo", "Shentong", "" ], [ "Xie", "Enze", "" ], [ "Chu", "Ruihang", "" ], [ "Yao", "Lewei", "" ], [ "Hong", "Lanqing", "" ], [ "Nießner", "Matthias", "" ], [ "Li", "Zhenguo", "" ] ]
Recent Diffusion Transformers (e.g., DiT) have demonstrated their powerful effectiveness in generating high-quality 2D images. However, it is still being determined whether the Transformer architecture performs equally well in 3D shape generation, as previous 3D diffusion methods mostly adopted the U-Net architecture. To bridge this gap, we propose a novel Diffusion Transformer for 3D shape generation, namely DiT-3D, which can directly operate the denoising process on voxelized point clouds using plain Transformers. Compared to existing U-Net approaches, our DiT-3D is more scalable in model size and produces much higher quality generations. Specifically, the DiT-3D adopts the design philosophy of DiT but modifies it by incorporating 3D positional and patch embeddings to adaptively aggregate input from voxelized point clouds. To reduce the computational cost of self-attention in 3D shape generation, we incorporate 3D window attention into Transformer blocks, as the increased 3D token length resulting from the additional dimension of voxels can lead to high computation. Finally, linear and devoxelization layers are used to predict the denoised point clouds. In addition, our transformer architecture supports efficient fine-tuning from 2D to 3D, where the pre-trained DiT-2D checkpoint on ImageNet can significantly improve DiT-3D on ShapeNet. Experimental results on the ShapeNet dataset demonstrate that the proposed DiT-3D achieves state-of-the-art performance in high-fidelity and diverse 3D point cloud generation. In particular, our DiT-3D decreases the 1-Nearest Neighbor Accuracy of the state-of-the-art method by 4.59 and increases the Coverage metric by 3.51 when evaluated on Chamfer Distance.
2206.00564
Laurie Burchell
Laurie Burchell, Alexandra Birch, Kenneth Heafield
Exploring Diversity in Back Translation for Low-Resource Machine Translation
null
null
10.18653/v1/2022.deeplo-1.8
null
cs.CL
http://creativecommons.org/licenses/by/4.0/
Back translation is one of the most widely used methods for improving the performance of neural machine translation systems. Recent research has sought to enhance the effectiveness of this method by increasing the 'diversity' of the generated translations. We argue that the definitions and metrics used to quantify 'diversity' in previous work have been insufficient. This work puts forward a more nuanced framework for understanding diversity in training data, splitting it into lexical diversity and syntactic diversity. We present novel metrics for measuring these different aspects of diversity and carry out empirical analysis into the effect of these types of diversity on final neural machine translation model performance for low-resource English$\leftrightarrow$Turkish and mid-resource English$\leftrightarrow$Icelandic. Our findings show that generating back translation using nucleus sampling results in higher final model performance, and that this method of generation has high levels of both lexical and syntactic diversity. We also find evidence that lexical diversity is more important than syntactic for back translation performance.
[ { "created": "Wed, 1 Jun 2022 15:21:16 GMT", "version": "v1" } ]
2023-09-01
[ [ "Burchell", "Laurie", "" ], [ "Birch", "Alexandra", "" ], [ "Heafield", "Kenneth", "" ] ]
Back translation is one of the most widely used methods for improving the performance of neural machine translation systems. Recent research has sought to enhance the effectiveness of this method by increasing the 'diversity' of the generated translations. We argue that the definitions and metrics used to quantify 'diversity' in previous work have been insufficient. This work puts forward a more nuanced framework for understanding diversity in training data, splitting it into lexical diversity and syntactic diversity. We present novel metrics for measuring these different aspects of diversity and carry out empirical analysis into the effect of these types of diversity on final neural machine translation model performance for low-resource English$\leftrightarrow$Turkish and mid-resource English$\leftrightarrow$Icelandic. Our findings show that generating back translation using nucleus sampling results in higher final model performance, and that this method of generation has high levels of both lexical and syntactic diversity. We also find evidence that lexical diversity is more important than syntactic for back translation performance.
2207.01171
Alessandra Carneiro
Alessandra Carneiro and Lorena Nascimento and Mauricio Noernberg and Carmem Hara and Aurora Pozo
Portuguese Man-of-War Image Classification with Convolutional Neural Networks
null
null
null
null
cs.CV cs.AI cs.LG
http://creativecommons.org/licenses/by/4.0/
Portuguese man-of-war (PMW) is a gelatinous organism with long tentacles capable of causing severe burns, thus leading to negative impacts on human activities, such as tourism and fishing. There is a lack of information about the spatio-temporal dynamics of this species. Therefore, the use of alternative methods for collecting data can contribute to their monitoring. Given the widespread use of social networks and the eye-catching look of PMW, Instagram posts can be a promising data source for monitoring. The first task to follow this approach is to identify posts that refer to PMW. This paper reports on the use of convolutional neural networks for PMW images classification, in order to automate the recognition of Instagram posts. We created a suitable dataset, and trained three different neural networks: VGG-16, ResNet50, and InceptionV3, with and without a pre-trained step with the ImageNet dataset. We analyzed their results using accuracy, precision, recall, and F1 score metrics. The pre-trained ResNet50 network presented the best results, obtaining 94% of accuracy and 95% of precision, recall, and F1 score. These results show that convolutional neural networks can be very effective for recognizing PMW images from the Instagram social media.
[ { "created": "Mon, 4 Jul 2022 03:06:45 GMT", "version": "v1" } ]
2022-07-05
[ [ "Carneiro", "Alessandra", "" ], [ "Nascimento", "Lorena", "" ], [ "Noernberg", "Mauricio", "" ], [ "Hara", "Carmem", "" ], [ "Pozo", "Aurora", "" ] ]
Portuguese man-of-war (PMW) is a gelatinous organism with long tentacles capable of causing severe burns, thus leading to negative impacts on human activities, such as tourism and fishing. There is a lack of information about the spatio-temporal dynamics of this species. Therefore, the use of alternative methods for collecting data can contribute to their monitoring. Given the widespread use of social networks and the eye-catching look of PMW, Instagram posts can be a promising data source for monitoring. The first task to follow this approach is to identify posts that refer to PMW. This paper reports on the use of convolutional neural networks for PMW images classification, in order to automate the recognition of Instagram posts. We created a suitable dataset, and trained three different neural networks: VGG-16, ResNet50, and InceptionV3, with and without a pre-trained step with the ImageNet dataset. We analyzed their results using accuracy, precision, recall, and F1 score metrics. The pre-trained ResNet50 network presented the best results, obtaining 94% of accuracy and 95% of precision, recall, and F1 score. These results show that convolutional neural networks can be very effective for recognizing PMW images from the Instagram social media.
2107.13165
Kushal Chawla
Kushal Chawla, Rene Clever, Jaysa Ramirez, Gale Lucas, Jonathan Gratch
Towards Emotion-Aware Agents For Negotiation Dialogues
Accepted at 9th International Conference on Affective Computing & Intelligent Interaction (ACII 2021)
null
null
null
cs.HC cs.AI cs.CL
http://creativecommons.org/licenses/by/4.0/
Negotiation is a complex social interaction that encapsulates emotional encounters in human decision-making. Virtual agents that can negotiate with humans are useful in pedagogy and conversational AI. To advance the development of such agents, we explore the prediction of two important subjective goals in a negotiation - outcome satisfaction and partner perception. Specifically, we analyze the extent to which emotion attributes extracted from the negotiation help in the prediction, above and beyond the individual difference variables. We focus on a recent dataset in chat-based negotiations, grounded in a realistic camping scenario. We study three degrees of emotion dimensions - emoticons, lexical, and contextual by leveraging affective lexicons and a state-of-the-art deep learning architecture. Our insights will be helpful in designing adaptive negotiation agents that interact through realistic communication interfaces.
[ { "created": "Wed, 28 Jul 2021 04:42:36 GMT", "version": "v1" } ]
2021-07-29
[ [ "Chawla", "Kushal", "" ], [ "Clever", "Rene", "" ], [ "Ramirez", "Jaysa", "" ], [ "Lucas", "Gale", "" ], [ "Gratch", "Jonathan", "" ] ]
Negotiation is a complex social interaction that encapsulates emotional encounters in human decision-making. Virtual agents that can negotiate with humans are useful in pedagogy and conversational AI. To advance the development of such agents, we explore the prediction of two important subjective goals in a negotiation - outcome satisfaction and partner perception. Specifically, we analyze the extent to which emotion attributes extracted from the negotiation help in the prediction, above and beyond the individual difference variables. We focus on a recent dataset in chat-based negotiations, grounded in a realistic camping scenario. We study three degrees of emotion dimensions - emoticons, lexical, and contextual by leveraging affective lexicons and a state-of-the-art deep learning architecture. Our insights will be helpful in designing adaptive negotiation agents that interact through realistic communication interfaces.
1410.1237
Mahantesh Halappanavar
Hao Lu and Mahantesh Halappanavar and Ananth Kalyanaraman
Parallel Heuristics for Scalable Community Detection
Submitted to a journal
null
null
null
cs.SI cs.DC physics.soc-ph
http://creativecommons.org/licenses/publicdomain/
Community detection has become a fundamental operation in numerous graph-theoretic applications. It is used to reveal natural divisions that exist within real world networks without imposing prior size or cardinality constraints on the set of communities. Despite its potential for application, there is only limited support for community detection on large-scale parallel computers, largely owing to the irregular and inherently sequential nature of the underlying heuristics. In this paper, we present parallelization heuristics for fast community detection using the Louvain method as the serial template. The Louvain method is an iterative heuristic for modularity optimization. Originally developed by Blondel et al. in 2008, the method has become increasingly popular owing to its ability to detect high modularity community partitions in a fast and memory-efficient manner. However, the method is also inherently sequential, thereby limiting its scalability. Here, we observe certain key properties of this method that present challenges for its parallelization, and consequently propose heuristics that are designed to break the sequential barrier. For evaluation purposes, we implemented our heuristics using OpenMP multithreading, and tested them over real world graphs derived from multiple application domains (e.g., internet, citation, biological). Compared to the serial Louvain implementation, our parallel implementation is able to produce community outputs with a higher modularity for most of the inputs tested, in comparable number or fewer iterations, while providing absolute speedups of up to 16x using 32 threads.
[ { "created": "Mon, 6 Oct 2014 01:54:15 GMT", "version": "v1" }, { "created": "Tue, 7 Oct 2014 01:01:12 GMT", "version": "v2" } ]
2014-10-08
[ [ "Lu", "Hao", "" ], [ "Halappanavar", "Mahantesh", "" ], [ "Kalyanaraman", "Ananth", "" ] ]
Community detection has become a fundamental operation in numerous graph-theoretic applications. It is used to reveal natural divisions that exist within real world networks without imposing prior size or cardinality constraints on the set of communities. Despite its potential for application, there is only limited support for community detection on large-scale parallel computers, largely owing to the irregular and inherently sequential nature of the underlying heuristics. In this paper, we present parallelization heuristics for fast community detection using the Louvain method as the serial template. The Louvain method is an iterative heuristic for modularity optimization. Originally developed by Blondel et al. in 2008, the method has become increasingly popular owing to its ability to detect high modularity community partitions in a fast and memory-efficient manner. However, the method is also inherently sequential, thereby limiting its scalability. Here, we observe certain key properties of this method that present challenges for its parallelization, and consequently propose heuristics that are designed to break the sequential barrier. For evaluation purposes, we implemented our heuristics using OpenMP multithreading, and tested them over real world graphs derived from multiple application domains (e.g., internet, citation, biological). Compared to the serial Louvain implementation, our parallel implementation is able to produce community outputs with a higher modularity for most of the inputs tested, in comparable number or fewer iterations, while providing absolute speedups of up to 16x using 32 threads.
1907.05016
Jing Li
Jing Li and Dongning Guo
On Analysis of the Bitcoin and Prism Backbone Protocols
null
null
null
null
cs.CR cs.DC cs.IT math.IT
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Bitcoin is a peer-to-peer payment system proposed by Nakamoto in 2008. Properties of the bitcoin backbone protocol have been investigated in some depth: the blockchain growth property quantifies the number of blocks added to the blockchain during any time intervals; the blockchain quality property ensures the honest miners always contribute at least a certain fraction of the blockchain; the common prefix property ensures if a block is deep enough, it will eventually be adopted by all honest miners with high probability. Following the spirit of decoupling various functionalities of the blockchain, the Prism protocol is proposed to dramatically improve the throughput while maintaining the same level of security. Prior analyses of the bitcoin and Prism backbone protocols assume the lifespan of blockchain is finite. This paper presents a streamlined and strengthened analysis without the finite horizon assumption. Specifically, the results include a blockchain growth property, a blockchain quality property, and a common prefix property of the bitcoin backbone protocol, as well as the liveness and persistence of the Prism backbone protocol regardless of whether the blockchains have a infinite lifespan. We also express the properties of bitcoin and Prism backbone protocols in explicit expressions rather than order optimal results, which lead to tighter bounds and practical references for public transaction ledger protocol design.
[ { "created": "Thu, 11 Jul 2019 06:35:05 GMT", "version": "v1" }, { "created": "Sun, 20 Oct 2019 01:10:42 GMT", "version": "v2" } ]
2019-10-22
[ [ "Li", "Jing", "" ], [ "Guo", "Dongning", "" ] ]
Bitcoin is a peer-to-peer payment system proposed by Nakamoto in 2008. Properties of the bitcoin backbone protocol have been investigated in some depth: the blockchain growth property quantifies the number of blocks added to the blockchain during any time intervals; the blockchain quality property ensures the honest miners always contribute at least a certain fraction of the blockchain; the common prefix property ensures if a block is deep enough, it will eventually be adopted by all honest miners with high probability. Following the spirit of decoupling various functionalities of the blockchain, the Prism protocol is proposed to dramatically improve the throughput while maintaining the same level of security. Prior analyses of the bitcoin and Prism backbone protocols assume the lifespan of blockchain is finite. This paper presents a streamlined and strengthened analysis without the finite horizon assumption. Specifically, the results include a blockchain growth property, a blockchain quality property, and a common prefix property of the bitcoin backbone protocol, as well as the liveness and persistence of the Prism backbone protocol regardless of whether the blockchains have a infinite lifespan. We also express the properties of bitcoin and Prism backbone protocols in explicit expressions rather than order optimal results, which lead to tighter bounds and practical references for public transaction ledger protocol design.
1512.01872
Pranav Rajpurkar
Pranav Rajpurkar, Toki Migimatsu, Jeff Kiske, Royce Cheng-Yue, Sameep Tandon, Tao Wang, Andrew Ng
Driverseat: Crowdstrapping Learning Tasks for Autonomous Driving
null
null
null
null
cs.HC cs.AI cs.RO
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
While emerging deep-learning systems have outclassed knowledge-based approaches in many tasks, their application to detection tasks for autonomous technologies remains an open field for scientific exploration. Broadly, there are two major developmental bottlenecks: the unavailability of comprehensively labeled datasets and of expressive evaluation strategies. Approaches for labeling datasets have relied on intensive hand-engineering, and strategies for evaluating learning systems have been unable to identify failure-case scenarios. Human intelligence offers an untapped approach for breaking through these bottlenecks. This paper introduces Driverseat, a technology for embedding crowds around learning systems for autonomous driving. Driverseat utilizes crowd contributions for (a) collecting complex 3D labels and (b) tagging diverse scenarios for ready evaluation of learning systems. We demonstrate how Driverseat can crowdstrap a convolutional neural network on the lane-detection task. More generally, crowdstrapping introduces a valuable paradigm for any technology that can benefit from leveraging the powerful combination of human and computer intelligence.
[ { "created": "Mon, 7 Dec 2015 01:34:23 GMT", "version": "v1" } ]
2015-12-08
[ [ "Rajpurkar", "Pranav", "" ], [ "Migimatsu", "Toki", "" ], [ "Kiske", "Jeff", "" ], [ "Cheng-Yue", "Royce", "" ], [ "Tandon", "Sameep", "" ], [ "Wang", "Tao", "" ], [ "Ng", "Andrew", "" ] ]
While emerging deep-learning systems have outclassed knowledge-based approaches in many tasks, their application to detection tasks for autonomous technologies remains an open field for scientific exploration. Broadly, there are two major developmental bottlenecks: the unavailability of comprehensively labeled datasets and of expressive evaluation strategies. Approaches for labeling datasets have relied on intensive hand-engineering, and strategies for evaluating learning systems have been unable to identify failure-case scenarios. Human intelligence offers an untapped approach for breaking through these bottlenecks. This paper introduces Driverseat, a technology for embedding crowds around learning systems for autonomous driving. Driverseat utilizes crowd contributions for (a) collecting complex 3D labels and (b) tagging diverse scenarios for ready evaluation of learning systems. We demonstrate how Driverseat can crowdstrap a convolutional neural network on the lane-detection task. More generally, crowdstrapping introduces a valuable paradigm for any technology that can benefit from leveraging the powerful combination of human and computer intelligence.
1803.09413
Sumita Mishra
Sachin Kumar, Sumita Mishra, Pooja Khanna, Pragya
Precision Sugarcane Monitoring Using SVM Classifier
This is a pre-print of an article published in [Procedia Computer Science 2017]
Procedia Computer Science,2017,vol.122,pp. 881-887
10.1016/j.procs.2017.11.450
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
India is agriculture based economy and sugarcane is one of the major crops produced in northern India. Productivity of sugarcane decreases due to inappropriate soil conditions and infections caused by various types of diseases , timely and accurate disease diagnosis, plays an important role towards optimizing crop yield. This paper presents a system model for monitoring of sugarcane crop, the proposed model continuously monitor parameters (temperature, humidity and moisture) responsible for healthy growth of the crop in addition KNN clustering along with SVM classifier is utilized for infection identification if any through images obtained at regular intervals. The data has been transmitted wirelessly from the site to the control unit. Model achieves an accuracy of 96% on a sample of 200 images, the model was tested at Lolai, near Malhaur, Gomti Nagar Extension.
[ { "created": "Mon, 26 Mar 2018 05:05:25 GMT", "version": "v1" } ]
2018-03-28
[ [ "Kumar", "Sachin", "" ], [ "Mishra", "Sumita", "" ], [ "Khanna", "Pooja", "" ], [ "Pragya", "", "" ] ]
India is agriculture based economy and sugarcane is one of the major crops produced in northern India. Productivity of sugarcane decreases due to inappropriate soil conditions and infections caused by various types of diseases , timely and accurate disease diagnosis, plays an important role towards optimizing crop yield. This paper presents a system model for monitoring of sugarcane crop, the proposed model continuously monitor parameters (temperature, humidity and moisture) responsible for healthy growth of the crop in addition KNN clustering along with SVM classifier is utilized for infection identification if any through images obtained at regular intervals. The data has been transmitted wirelessly from the site to the control unit. Model achieves an accuracy of 96% on a sample of 200 images, the model was tested at Lolai, near Malhaur, Gomti Nagar Extension.
2402.14162
Minh-Hao Van
Minh-Hao Van, Prateek Verma, Xintao Wu
On Large Visual Language Models for Medical Imaging Analysis: An Empirical Study
null
null
null
null
cs.CV cs.AI
http://creativecommons.org/licenses/by/4.0/
Recently, large language models (LLMs) have taken the spotlight in natural language processing. Further, integrating LLMs with vision enables the users to explore emergent abilities with multimodal data. Visual language models (VLMs), such as LLaVA, Flamingo, or CLIP, have demonstrated impressive performance on various visio-linguistic tasks. Consequently, there are enormous applications of large models that could be potentially used in the biomedical imaging field. Along that direction, there is a lack of related work to show the ability of large models to diagnose the diseases. In this work, we study the zero-shot and few-shot robustness of VLMs on the medical imaging analysis tasks. Our comprehensive experiments demonstrate the effectiveness of VLMs in analyzing biomedical images such as brain MRIs, microscopic images of blood cells, and chest X-rays.
[ { "created": "Wed, 21 Feb 2024 23:01:38 GMT", "version": "v1" } ]
2024-02-23
[ [ "Van", "Minh-Hao", "" ], [ "Verma", "Prateek", "" ], [ "Wu", "Xintao", "" ] ]
Recently, large language models (LLMs) have taken the spotlight in natural language processing. Further, integrating LLMs with vision enables the users to explore emergent abilities with multimodal data. Visual language models (VLMs), such as LLaVA, Flamingo, or CLIP, have demonstrated impressive performance on various visio-linguistic tasks. Consequently, there are enormous applications of large models that could be potentially used in the biomedical imaging field. Along that direction, there is a lack of related work to show the ability of large models to diagnose the diseases. In this work, we study the zero-shot and few-shot robustness of VLMs on the medical imaging analysis tasks. Our comprehensive experiments demonstrate the effectiveness of VLMs in analyzing biomedical images such as brain MRIs, microscopic images of blood cells, and chest X-rays.
2312.07381
Hallee Wong
Hallee E. Wong, Marianne Rakic, John Guttag, Adrian V. Dalca
ScribblePrompt: Fast and Flexible Interactive Segmentation for Any Biomedical Image
Accepted by ECCV 2024. Project Website: https://scribbleprompt.csail.mit.edu Keywords: Interactive Segmentation, Medical Imaging, Segment Anything Model, SAM, Scribble Annotations, Prompt
null
null
null
cs.CV eess.IV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Biomedical image segmentation is a crucial part of both scientific research and clinical care. With enough labelled data, deep learning models can be trained to accurately automate specific biomedical image segmentation tasks. However, manually segmenting images to create training data is highly labor intensive and requires domain expertise. We present \emph{ScribblePrompt}, a flexible neural network based interactive segmentation tool for biomedical imaging that enables human annotators to segment previously unseen structures using scribbles, clicks, and bounding boxes. Through rigorous quantitative experiments, we demonstrate that given comparable amounts of interaction, ScribblePrompt produces more accurate segmentations than previous methods on datasets unseen during training. In a user study with domain experts, ScribblePrompt reduced annotation time by 28% while improving Dice by 15% compared to the next best method. ScribblePrompt's success rests on a set of careful design decisions. These include a training strategy that incorporates both a highly diverse set of images and tasks, novel algorithms for simulated user interactions and labels, and a network that enables fast inference. We showcase ScribblePrompt in an interactive demo, provide code, and release a dataset of scribble annotations at https://scribbleprompt.csail.mit.edu
[ { "created": "Tue, 12 Dec 2023 15:57:03 GMT", "version": "v1" }, { "created": "Fri, 12 Apr 2024 20:41:14 GMT", "version": "v2" }, { "created": "Tue, 16 Jul 2024 21:21:42 GMT", "version": "v3" } ]
2024-07-18
[ [ "Wong", "Hallee E.", "" ], [ "Rakic", "Marianne", "" ], [ "Guttag", "John", "" ], [ "Dalca", "Adrian V.", "" ] ]
Biomedical image segmentation is a crucial part of both scientific research and clinical care. With enough labelled data, deep learning models can be trained to accurately automate specific biomedical image segmentation tasks. However, manually segmenting images to create training data is highly labor intensive and requires domain expertise. We present \emph{ScribblePrompt}, a flexible neural network based interactive segmentation tool for biomedical imaging that enables human annotators to segment previously unseen structures using scribbles, clicks, and bounding boxes. Through rigorous quantitative experiments, we demonstrate that given comparable amounts of interaction, ScribblePrompt produces more accurate segmentations than previous methods on datasets unseen during training. In a user study with domain experts, ScribblePrompt reduced annotation time by 28% while improving Dice by 15% compared to the next best method. ScribblePrompt's success rests on a set of careful design decisions. These include a training strategy that incorporates both a highly diverse set of images and tasks, novel algorithms for simulated user interactions and labels, and a network that enables fast inference. We showcase ScribblePrompt in an interactive demo, provide code, and release a dataset of scribble annotations at https://scribbleprompt.csail.mit.edu
1712.06897
Jie Lyu
Jie Lyu, Zejian Yuan, Dapeng Chen
Learning Fixation Point Strategy for Object Detection and Classification
null
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
We propose a novel recurrent attentional structure to localize and recognize objects jointly. The network can learn to extract a sequence of local observations with detailed appearance and rough context, instead of sliding windows or convolutions on the entire image. Meanwhile, those observations are fused to complete detection and classification tasks. On training, we present a hybrid loss function to learn the parameters of the multi-task network end-to-end. Particularly, the combination of stochastic and object-awareness strategy, named SA, can select more abundant context and ensure the last fixation close to the object. In addition, we build a real-world dataset to verify the capacity of our method in detecting the object of interest including those small ones. Our method can predict a precise bounding box on an image, and achieve high speed on large images without pooling operations. Experimental results indicate that the proposed method can mine effective context by several local observations. Moreover, the precision and speed are easily improved by changing the number of recurrent steps. Finally, we will open the source code of our proposed approach.
[ { "created": "Tue, 19 Dec 2017 12:28:01 GMT", "version": "v1" } ]
2017-12-20
[ [ "Lyu", "Jie", "" ], [ "Yuan", "Zejian", "" ], [ "Chen", "Dapeng", "" ] ]
We propose a novel recurrent attentional structure to localize and recognize objects jointly. The network can learn to extract a sequence of local observations with detailed appearance and rough context, instead of sliding windows or convolutions on the entire image. Meanwhile, those observations are fused to complete detection and classification tasks. On training, we present a hybrid loss function to learn the parameters of the multi-task network end-to-end. Particularly, the combination of stochastic and object-awareness strategy, named SA, can select more abundant context and ensure the last fixation close to the object. In addition, we build a real-world dataset to verify the capacity of our method in detecting the object of interest including those small ones. Our method can predict a precise bounding box on an image, and achieve high speed on large images without pooling operations. Experimental results indicate that the proposed method can mine effective context by several local observations. Moreover, the precision and speed are easily improved by changing the number of recurrent steps. Finally, we will open the source code of our proposed approach.
2310.14162
Rohan Gupta
Rohan Gupta
Augmenting End-to-End Steering Angle Prediction with CAN Bus Data
5 pages
null
null
null
cs.CV cs.AI
http://creativecommons.org/publicdomain/zero/1.0/
In recent years, end to end steering prediction for autonomous vehicles has become a major area of research. The primary method for achieving end to end steering was to use computer vision models on a live feed of video data. However, to further increase accuracy, many companies have added data from light detection and ranging (LiDAR) and or radar sensors through sensor fusion. However, the addition of lasers and sensors comes at a high financial cost. In this paper, I address both of these issues by increasing the accuracy of the computer vision models without the increased cost of using LiDAR and or sensors. I achieved this by improving the accuracy of computer vision models by sensor fusing CAN bus data, a vehicle protocol, with video data. CAN bus data is a rich source of information about the vehicle's state, including its speed, steering angle, and acceleration. By fusing this data with video data, the accuracy of the computer vision model's predictions can be improved. When I trained the model without CAN bus data, I obtained an RMSE of 0.02492, while the model trained with the CAN bus data achieved an RMSE of 0.01970. This finding indicates that fusing CAN Bus data with video data can reduce the computer vision model's prediction error by 20% with some models decreasing the error by 80%.
[ { "created": "Sun, 22 Oct 2023 03:24:53 GMT", "version": "v1" } ]
2023-10-24
[ [ "Gupta", "Rohan", "" ] ]
In recent years, end to end steering prediction for autonomous vehicles has become a major area of research. The primary method for achieving end to end steering was to use computer vision models on a live feed of video data. However, to further increase accuracy, many companies have added data from light detection and ranging (LiDAR) and or radar sensors through sensor fusion. However, the addition of lasers and sensors comes at a high financial cost. In this paper, I address both of these issues by increasing the accuracy of the computer vision models without the increased cost of using LiDAR and or sensors. I achieved this by improving the accuracy of computer vision models by sensor fusing CAN bus data, a vehicle protocol, with video data. CAN bus data is a rich source of information about the vehicle's state, including its speed, steering angle, and acceleration. By fusing this data with video data, the accuracy of the computer vision model's predictions can be improved. When I trained the model without CAN bus data, I obtained an RMSE of 0.02492, while the model trained with the CAN bus data achieved an RMSE of 0.01970. This finding indicates that fusing CAN Bus data with video data can reduce the computer vision model's prediction error by 20% with some models decreasing the error by 80%.
1309.6455
Gwen Spencer PhD
Gwen Spencer and Richard Howarth
Maximizing the Spread of Stable Influence: Leveraging Norm-driven Moral-Motivation for Green Behavior Change in Networks
null
null
null
null
cs.SI physics.soc-ph
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
In an effort to understand why individuals choose to participate in personally-expensive pro-environmental behaviors, environmental and behavioral economists have examined a moral-motivation model in which the decision to adopt a pro-environmental behavior depends on the society-wide market share of that behavior. An increasing body of practical research on adoption of pro-environmental behavior emphasizes the importance of encouragement from local social contacts and messaging about locally-embraced norms: we respond by extending the moral-motivation model to a social networks setting. We obtain a new decision rule: an individual adopts a pro-environmental behavior if he or she observes a certain threshold of adoption within their local social neighborhood. This gives rise to a concurrent update process which describes adoption of a pro-environmental behavior spreading through a network. The original moral-motivation model corresponds to the special case of our network version in a complete graph. By improving convergence results, we formulate modest-size Integer Programs that accurately (but not efficiently) find minimum-size sets of nodes that convert the entire network, or alternately that maximize long-term adoption in the network given a limited number of nodes which may be temporarily converted. Issues of stability in determining long-term adoption are key. We give hardness of approximation results for these optimization problems. We demonstrate that there exist classes of networks which qualitatively have severely different behavior than the non-networked version, and provide preliminary computational results in in modestly-sized highly-clustered small-world networks related to the famous small-world networks of Watts and Strogatz.
[ { "created": "Wed, 25 Sep 2013 10:30:57 GMT", "version": "v1" } ]
2013-09-26
[ [ "Spencer", "Gwen", "" ], [ "Howarth", "Richard", "" ] ]
In an effort to understand why individuals choose to participate in personally-expensive pro-environmental behaviors, environmental and behavioral economists have examined a moral-motivation model in which the decision to adopt a pro-environmental behavior depends on the society-wide market share of that behavior. An increasing body of practical research on adoption of pro-environmental behavior emphasizes the importance of encouragement from local social contacts and messaging about locally-embraced norms: we respond by extending the moral-motivation model to a social networks setting. We obtain a new decision rule: an individual adopts a pro-environmental behavior if he or she observes a certain threshold of adoption within their local social neighborhood. This gives rise to a concurrent update process which describes adoption of a pro-environmental behavior spreading through a network. The original moral-motivation model corresponds to the special case of our network version in a complete graph. By improving convergence results, we formulate modest-size Integer Programs that accurately (but not efficiently) find minimum-size sets of nodes that convert the entire network, or alternately that maximize long-term adoption in the network given a limited number of nodes which may be temporarily converted. Issues of stability in determining long-term adoption are key. We give hardness of approximation results for these optimization problems. We demonstrate that there exist classes of networks which qualitatively have severely different behavior than the non-networked version, and provide preliminary computational results in in modestly-sized highly-clustered small-world networks related to the famous small-world networks of Watts and Strogatz.
2305.02309
Erik Nijkamp Dr.
Erik Nijkamp, Hiroaki Hayashi, Caiming Xiong, Silvio Savarese, Yingbo Zhou
CodeGen2: Lessons for Training LLMs on Programming and Natural Languages
null
null
null
null
cs.LG
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Large language models (LLMs) have demonstrated remarkable abilities in representation learning for program synthesis and understanding tasks. The quality of the learned representations appears to be dictated by the neural scaling laws as a function of the number of model parameters and observations, while imposing upper bounds on the model performance by the amount of available data and compute, which is costly. In this study, we attempt to render the training of LLMs for program synthesis more efficient by unifying four key components: (1) model architectures, (2) learning methods, (3) infill sampling, and, (4) data distributions. Specifically, for the model architecture, we attempt to unify encoder and decoder-based models into a single prefix-LM. For learning methods, (i) causal language modeling, (ii) span corruption, (iii) infilling are unified into a simple learning algorithm. For infill sampling, we explore the claim of a "free lunch" hypothesis. For data distributions, the effect of a mixture distribution and multi-epoch training of programming and natural languages on model performance is explored. We conduct a comprehensive series of empirical experiments on 1B LLMs, for which failures and successes of this exploration are distilled into five lessons. We will provide a final recipe for training and release CodeGen2 models in size 1B, 3.7B, 7B, and, 16B parameters, along with the training framework as open-source: https://github.com/salesforce/CodeGen.
[ { "created": "Wed, 3 May 2023 17:55:25 GMT", "version": "v1" }, { "created": "Tue, 11 Jul 2023 21:11:23 GMT", "version": "v2" } ]
2023-07-13
[ [ "Nijkamp", "Erik", "" ], [ "Hayashi", "Hiroaki", "" ], [ "Xiong", "Caiming", "" ], [ "Savarese", "Silvio", "" ], [ "Zhou", "Yingbo", "" ] ]
Large language models (LLMs) have demonstrated remarkable abilities in representation learning for program synthesis and understanding tasks. The quality of the learned representations appears to be dictated by the neural scaling laws as a function of the number of model parameters and observations, while imposing upper bounds on the model performance by the amount of available data and compute, which is costly. In this study, we attempt to render the training of LLMs for program synthesis more efficient by unifying four key components: (1) model architectures, (2) learning methods, (3) infill sampling, and, (4) data distributions. Specifically, for the model architecture, we attempt to unify encoder and decoder-based models into a single prefix-LM. For learning methods, (i) causal language modeling, (ii) span corruption, (iii) infilling are unified into a simple learning algorithm. For infill sampling, we explore the claim of a "free lunch" hypothesis. For data distributions, the effect of a mixture distribution and multi-epoch training of programming and natural languages on model performance is explored. We conduct a comprehensive series of empirical experiments on 1B LLMs, for which failures and successes of this exploration are distilled into five lessons. We will provide a final recipe for training and release CodeGen2 models in size 1B, 3.7B, 7B, and, 16B parameters, along with the training framework as open-source: https://github.com/salesforce/CodeGen.
2007.03338
Mehdi Ghatee Dr.
Marzieh Heidari, Mehdi Ghatee, Ahmad Nickabadi, Arash Pourhasan Nezhad
Diverse and Styled Image Captioning Using SVD-Based Mixture of Recurrent Experts
13 pages, 4 figures and 5 tables, extracted from an MSc thesis in the Amirkabir University of Technology, Tehran, Iran
Concurrency and Computation: Practice and Experience, 2022
10.1002/cpe.6866
null
cs.CV cs.CL cs.LG
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
With great advances in vision and natural language processing, the generation of image captions becomes a need. In a recent paper, Mathews, Xie and He [1], extended a new model to generate styled captions by separating semantics and style. In continuation of this work, here a new captioning model is developed including an image encoder to extract the features, a mixture of recurrent networks to embed the set of extracted features to a set of words, and a sentence generator that combines the obtained words as a stylized sentence. The resulted system that entitled as Mixture of Recurrent Experts (MoRE), uses a new training algorithm that derives singular value decomposition (SVD) from weighting matrices of Recurrent Neural Networks (RNNs) to increase the diversity of captions. Each decomposition step depends on a distinctive factor based on the number of RNNs in MoRE. Since the used sentence generator gives a stylized language corpus without paired images, our captioning model can do the same. Besides, the styled and diverse captions are extracted without training on a densely labeled or styled dataset. To validate this captioning model, we use Microsoft COCO which is a standard factual image caption dataset. We show that the proposed captioning model can generate a diverse and stylized image captions without the necessity of extra-labeling. The results also show better descriptions in terms of content accuracy.
[ { "created": "Tue, 7 Jul 2020 11:00:27 GMT", "version": "v1" } ]
2022-02-03
[ [ "Heidari", "Marzieh", "" ], [ "Ghatee", "Mehdi", "" ], [ "Nickabadi", "Ahmad", "" ], [ "Nezhad", "Arash Pourhasan", "" ] ]
With great advances in vision and natural language processing, the generation of image captions becomes a need. In a recent paper, Mathews, Xie and He [1], extended a new model to generate styled captions by separating semantics and style. In continuation of this work, here a new captioning model is developed including an image encoder to extract the features, a mixture of recurrent networks to embed the set of extracted features to a set of words, and a sentence generator that combines the obtained words as a stylized sentence. The resulted system that entitled as Mixture of Recurrent Experts (MoRE), uses a new training algorithm that derives singular value decomposition (SVD) from weighting matrices of Recurrent Neural Networks (RNNs) to increase the diversity of captions. Each decomposition step depends on a distinctive factor based on the number of RNNs in MoRE. Since the used sentence generator gives a stylized language corpus without paired images, our captioning model can do the same. Besides, the styled and diverse captions are extracted without training on a densely labeled or styled dataset. To validate this captioning model, we use Microsoft COCO which is a standard factual image caption dataset. We show that the proposed captioning model can generate a diverse and stylized image captions without the necessity of extra-labeling. The results also show better descriptions in terms of content accuracy.
1905.03919
Filippo Menczer
Kazutoshi Sasahara, Wen Chen, Hao Peng, Giovanni Luca Ciampaglia, Alessandro Flammini, Filippo Menczer
Social Influence and Unfollowing Accelerate the Emergence of Echo Chambers
28 pages, 11 figures. Forthcoming in Journal of Computational Social Science
J Comput Soc Sc (2020)
10.1007/s42001-020-00084-7
null
cs.CY cs.SI physics.soc-ph
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
While social media make it easy to connect with and access information from anyone, they also facilitate basic influence and unfriending mechanisms that may lead to segregated and polarized clusters known as "echo chambers." Here we study the conditions in which such echo chambers emerge by introducing a simple model of information sharing in online social networks with the two ingredients of influence and unfriending. Users can change both their opinions and social connections based on the information to which they are exposed through sharing. The model dynamics show that even with minimal amounts of influence and unfriending, the social network rapidly devolves into segregated, homogeneous communities. These predictions are consistent with empirical data from Twitter. Although our findings suggest that echo chambers are somewhat inevitable given the mechanisms at play in online social media, they also provide insights into possible mitigation strategies.
[ { "created": "Fri, 10 May 2019 03:08:23 GMT", "version": "v1" }, { "created": "Mon, 20 May 2019 16:40:50 GMT", "version": "v2" }, { "created": "Tue, 25 Aug 2020 02:27:40 GMT", "version": "v3" } ]
2020-09-15
[ [ "Sasahara", "Kazutoshi", "" ], [ "Chen", "Wen", "" ], [ "Peng", "Hao", "" ], [ "Ciampaglia", "Giovanni Luca", "" ], [ "Flammini", "Alessandro", "" ], [ "Menczer", "Filippo", "" ] ]
While social media make it easy to connect with and access information from anyone, they also facilitate basic influence and unfriending mechanisms that may lead to segregated and polarized clusters known as "echo chambers." Here we study the conditions in which such echo chambers emerge by introducing a simple model of information sharing in online social networks with the two ingredients of influence and unfriending. Users can change both their opinions and social connections based on the information to which they are exposed through sharing. The model dynamics show that even with minimal amounts of influence and unfriending, the social network rapidly devolves into segregated, homogeneous communities. These predictions are consistent with empirical data from Twitter. Although our findings suggest that echo chambers are somewhat inevitable given the mechanisms at play in online social media, they also provide insights into possible mitigation strategies.
1702.07478
Igor Tarasyuk
Igor V. Tarasyuk, Hermenegilda Maci\`a, Valent\'in Valero
Stochastic equivalence for performance analysis of concurrent systems in dtsiPBC
Prepared for submission to Discrete Mathematics and Theoretical Computer Science
null
null
null
cs.LO
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
We propose an extension with immediate multiactions of discrete time stochastic Petri Box Calculus (dtsPBC), presented by I.V. Tarasyuk. The resulting algebra dtsiPBC is a discrete time analogue of stochastic Petri Box Calculus (sPBC) with immediate multiactions, designed by H. Maci\`a, V. Valero et al. within a continuous time domain. The step operational semantics is constructed via labeled probabilistic transition systems. The denotational semantics is based on labeled discrete time stochastic Petri nets with immediate transitions. To evaluate performance, the corresponding semi-Markov chains are analyzed. We define step stochastic bisimulation equivalence of expressions that is applied to reduce their transition systems and underlying semi-Markov chains while preserving the functionality and performance characteristics. We explain how this equivalence can be used to simplify performance analysis of the algebraic processes. In a case study, a method of modeling, performance evaluation and behaviour reduction for concurrent systems is outlined and applied to the shared memory system.
[ { "created": "Fri, 24 Feb 2017 07:07:24 GMT", "version": "v1" } ]
2017-02-27
[ [ "Tarasyuk", "Igor V.", "" ], [ "Macià", "Hermenegilda", "" ], [ "Valero", "Valentín", "" ] ]
We propose an extension with immediate multiactions of discrete time stochastic Petri Box Calculus (dtsPBC), presented by I.V. Tarasyuk. The resulting algebra dtsiPBC is a discrete time analogue of stochastic Petri Box Calculus (sPBC) with immediate multiactions, designed by H. Maci\`a, V. Valero et al. within a continuous time domain. The step operational semantics is constructed via labeled probabilistic transition systems. The denotational semantics is based on labeled discrete time stochastic Petri nets with immediate transitions. To evaluate performance, the corresponding semi-Markov chains are analyzed. We define step stochastic bisimulation equivalence of expressions that is applied to reduce their transition systems and underlying semi-Markov chains while preserving the functionality and performance characteristics. We explain how this equivalence can be used to simplify performance analysis of the algebraic processes. In a case study, a method of modeling, performance evaluation and behaviour reduction for concurrent systems is outlined and applied to the shared memory system.
1810.01966
Ekram Hossain
Mohammad Salehi, Hina Tabassum, and Ekram Hossain
Accuracy of Distance-Based Ranking of Users in the Analysis of NOMA Systems
null
null
null
null
cs.IT math.IT
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
We characterize the accuracy of analyzing the performance of a NOMA system where users are ranked according to their distances instead of instantaneous channel gains, i.e., product of distance-based path-loss and fading channel gains. Distance-based ranking is analytically tractable and can lead to important insights. However, it may not be appropriate in a multipath fading environment where a near user suffers from severe fading while a far user experiences weak fading. Since the ranking of users in a NOMA system has a direct impact on coverage probability analysis, impact of the traditional distance-based ranking, as opposed to instantaneous signal power-based ranking, needs to be understood. This will enable us to identify scenarios where distance-based ranking, which is easier to implement compared to instantaneous signal power-based ranking, is acceptable for system performance analysis. To this end, in this paper, we derive the probability of the event when distance-based ranking yields the same results as instantaneous signal power-based ranking, which is referred to as the accuracy probability. We characterize the probability of accuracy considering Nakagami-m fading channels and three different spatial distribution models of user locations in NOMA. We illustrate the impact of accuracy probability on uplink and downlink coverage probability.
[ { "created": "Wed, 3 Oct 2018 20:50:32 GMT", "version": "v1" } ]
2018-10-05
[ [ "Salehi", "Mohammad", "" ], [ "Tabassum", "Hina", "" ], [ "Hossain", "Ekram", "" ] ]
We characterize the accuracy of analyzing the performance of a NOMA system where users are ranked according to their distances instead of instantaneous channel gains, i.e., product of distance-based path-loss and fading channel gains. Distance-based ranking is analytically tractable and can lead to important insights. However, it may not be appropriate in a multipath fading environment where a near user suffers from severe fading while a far user experiences weak fading. Since the ranking of users in a NOMA system has a direct impact on coverage probability analysis, impact of the traditional distance-based ranking, as opposed to instantaneous signal power-based ranking, needs to be understood. This will enable us to identify scenarios where distance-based ranking, which is easier to implement compared to instantaneous signal power-based ranking, is acceptable for system performance analysis. To this end, in this paper, we derive the probability of the event when distance-based ranking yields the same results as instantaneous signal power-based ranking, which is referred to as the accuracy probability. We characterize the probability of accuracy considering Nakagami-m fading channels and three different spatial distribution models of user locations in NOMA. We illustrate the impact of accuracy probability on uplink and downlink coverage probability.
1203.0587
Alex Brik
Alex Brik, Jeffrey B. Remmel
Expressing Preferences using Preference Set Constraint Atoms
9 pages
null
null
null
cs.LO
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
This paper introduces an extension of Answer Set Programming called Preference Set Constraint Programming which is a convenient and general formalism to reason with preferences. PSC programming extends Set Constraint Programming introduced by Marek and Remmel (Marek and Remmel 2004) by introducing two types of preference set constraint atoms, measure preference set constraint atoms and pre-ordered preference set constraint atoms, which are extensions of set constraint atoms. We show that the question of whether a PSC program has a preferred stable model is CoNP-complete. We give examples of the uses of the preference set constraint atoms and show that Answer Set Optimization (Brewka, Niemel\"a, and Truszczynski 2003) and General Preference (Son and Pontelli 2006) can be expressed using preference set constraint atoms.
[ { "created": "Fri, 2 Mar 2012 23:25:07 GMT", "version": "v1" } ]
2012-03-06
[ [ "Brik", "Alex", "" ], [ "Remmel", "Jeffrey B.", "" ] ]
This paper introduces an extension of Answer Set Programming called Preference Set Constraint Programming which is a convenient and general formalism to reason with preferences. PSC programming extends Set Constraint Programming introduced by Marek and Remmel (Marek and Remmel 2004) by introducing two types of preference set constraint atoms, measure preference set constraint atoms and pre-ordered preference set constraint atoms, which are extensions of set constraint atoms. We show that the question of whether a PSC program has a preferred stable model is CoNP-complete. We give examples of the uses of the preference set constraint atoms and show that Answer Set Optimization (Brewka, Niemel\"a, and Truszczynski 2003) and General Preference (Son and Pontelli 2006) can be expressed using preference set constraint atoms.
1711.09008
Yuming Jiang
Atef Abdelkefi and Yuming Jiang and Sachin Sharma
SENATUS: An Approach to Joint Traffic Anomaly Detection and Root Cause Analysis
null
null
null
null
cs.NI
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
In this paper, we propose a novel approach, called SENATUS, for joint traffic anomaly detection and root-cause analysis. Inspired from the concept of a senate, the key idea of the proposed approach is divided into three stages: election, voting and decision. At the election stage, a small number of \nop{traffic flow sets (termed as senator flows)}senator flows are chosen\nop{, which are used} to represent approximately the total (usually huge) set of traffic flows. In the voting stage, anomaly detection is applied on the senator flows and the detected anomalies are correlated to identify the most possible anomalous time bins. Finally in the decision stage, a machine learning technique is applied to the senator flows of each anomalous time bin to find the root cause of the anomalies. We evaluate SENATUS using traffic traces collected from the Pan European network, GEANT, and compare against another approach which detects anomalies using lossless compression of traffic histograms. We show the effectiveness of SENATUS in diagnosing anomaly types: network scans and DoS/DDoS attacks.
[ { "created": "Fri, 24 Nov 2017 15:14:50 GMT", "version": "v1" } ]
2017-11-27
[ [ "Abdelkefi", "Atef", "" ], [ "Jiang", "Yuming", "" ], [ "Sharma", "Sachin", "" ] ]
In this paper, we propose a novel approach, called SENATUS, for joint traffic anomaly detection and root-cause analysis. Inspired from the concept of a senate, the key idea of the proposed approach is divided into three stages: election, voting and decision. At the election stage, a small number of \nop{traffic flow sets (termed as senator flows)}senator flows are chosen\nop{, which are used} to represent approximately the total (usually huge) set of traffic flows. In the voting stage, anomaly detection is applied on the senator flows and the detected anomalies are correlated to identify the most possible anomalous time bins. Finally in the decision stage, a machine learning technique is applied to the senator flows of each anomalous time bin to find the root cause of the anomalies. We evaluate SENATUS using traffic traces collected from the Pan European network, GEANT, and compare against another approach which detects anomalies using lossless compression of traffic histograms. We show the effectiveness of SENATUS in diagnosing anomaly types: network scans and DoS/DDoS attacks.
1602.03599
EPTCS
Juliana Franco (Imperial College London), Sophia Drossopoulou (Imperial College London)
Behavioural types for non-uniform memory accesses
In Proceedings PLACES 2015, arXiv:1602.03254
EPTCS 203, 2016, pp. 109-120
10.4204/EPTCS.203.9
null
cs.PL
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Concurrent programs executing on NUMA architectures consist of concurrent entities (e.g. threads, actors) and data placed on different nodes. Execution of these concurrent entities often reads or updates states from remote nodes. The performance of such systems depends on the extent to which the concurrent entities can be executing in parallel, and on the amount of the remote reads and writes. We consider an actor-based object oriented language, and propose a type system which expresses the topology of the program (the placement of the actors and data on the nodes), and an effect system which characterises remote reads and writes (in terms of which node reads/writes from which other nodes). We use a variant of ownership types for the topology, and a combination of behavioural and ownership types for the effect system.
[ { "created": "Thu, 11 Feb 2016 01:21:09 GMT", "version": "v1" } ]
2016-02-12
[ [ "Franco", "Juliana", "", "Imperial College London" ], [ "Drossopoulou", "Sophia", "", "Imperial College London" ] ]
Concurrent programs executing on NUMA architectures consist of concurrent entities (e.g. threads, actors) and data placed on different nodes. Execution of these concurrent entities often reads or updates states from remote nodes. The performance of such systems depends on the extent to which the concurrent entities can be executing in parallel, and on the amount of the remote reads and writes. We consider an actor-based object oriented language, and propose a type system which expresses the topology of the program (the placement of the actors and data on the nodes), and an effect system which characterises remote reads and writes (in terms of which node reads/writes from which other nodes). We use a variant of ownership types for the topology, and a combination of behavioural and ownership types for the effect system.
2105.03655
Huy Ha
Huy Ha, Shuran Song
FlingBot: The Unreasonable Effectiveness of Dynamic Manipulation for Cloth Unfolding
11 pages, 6 figures. Code, data, and simulation environment publicly available at https://flingbot.cs.columbia.edu
Conference on Robot Learning (CoRL 2021)
null
null
cs.RO
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
High-velocity dynamic actions (e.g., fling or throw) play a crucial role in our everyday interaction with deformable objects by improving our efficiency and effectively expanding our physical reach range. Yet, most prior works have tackled cloth manipulation using exclusively single-arm quasi-static actions, which requires a large number of interactions for challenging initial cloth configurations and strictly limits the maximum cloth size by the robot's reach range. In this work, we demonstrate the effectiveness of dynamic flinging actions for cloth unfolding with our proposed self-supervised learning framework, FlingBot. Our approach learns how to unfold a piece of fabric from arbitrary initial configurations using a pick, stretch, and fling primitive for a dual-arm setup from visual observations. The final system achieves over 80% coverage within 3 actions on novel cloths, can unfold cloths larger than the system's reach range, and generalizes to T-shirts despite being trained on only rectangular cloths. We also finetuned FlingBot on a real-world dual-arm robot platform, where it increased the cloth coverage over 4 times more than the quasi-static baseline did. The simplicity of FlingBot combined with its superior performance over quasi-static baselines demonstrates the effectiveness of dynamic actions for deformable object manipulation.
[ { "created": "Sat, 8 May 2021 09:48:15 GMT", "version": "v1" }, { "created": "Tue, 29 Jun 2021 05:47:27 GMT", "version": "v2" }, { "created": "Mon, 18 Oct 2021 19:03:27 GMT", "version": "v3" } ]
2021-10-20
[ [ "Ha", "Huy", "" ], [ "Song", "Shuran", "" ] ]
High-velocity dynamic actions (e.g., fling or throw) play a crucial role in our everyday interaction with deformable objects by improving our efficiency and effectively expanding our physical reach range. Yet, most prior works have tackled cloth manipulation using exclusively single-arm quasi-static actions, which requires a large number of interactions for challenging initial cloth configurations and strictly limits the maximum cloth size by the robot's reach range. In this work, we demonstrate the effectiveness of dynamic flinging actions for cloth unfolding with our proposed self-supervised learning framework, FlingBot. Our approach learns how to unfold a piece of fabric from arbitrary initial configurations using a pick, stretch, and fling primitive for a dual-arm setup from visual observations. The final system achieves over 80% coverage within 3 actions on novel cloths, can unfold cloths larger than the system's reach range, and generalizes to T-shirts despite being trained on only rectangular cloths. We also finetuned FlingBot on a real-world dual-arm robot platform, where it increased the cloth coverage over 4 times more than the quasi-static baseline did. The simplicity of FlingBot combined with its superior performance over quasi-static baselines demonstrates the effectiveness of dynamic actions for deformable object manipulation.
2405.16091
Myong Chol Jung
Myong Chol Jung, He Zhao, Joanna Dipnall, Belinda Gabbe, Lan Du
Enhancing Near OOD Detection in Prompt Learning: Maximum Gains, Minimal Costs
null
null
null
null
cs.CV
http://creativecommons.org/licenses/by/4.0/
Prompt learning has shown to be an efficient and effective fine-tuning method for vision-language models like CLIP. While numerous studies have focused on the generalisation of these models in few-shot classification, their capability in near out-of-distribution (OOD) detection has been overlooked. A few recent works have highlighted the promising performance of prompt learning in far OOD detection. However, the more challenging task of few-shot near OOD detection has not yet been addressed. In this study, we investigate the near OOD detection capabilities of prompt learning models and observe that commonly used OOD scores have limited performance in near OOD detection. To enhance the performance, we propose a fast and simple post-hoc method that complements existing logit-based scores, improving near OOD detection AUROC by up to 11.67% with minimal computational cost. Our method can be easily applied to any prompt learning model without change in architecture or re-training the models. Comprehensive empirical evaluations across 13 datasets and 8 models demonstrate the effectiveness and adaptability of our method.
[ { "created": "Sat, 25 May 2024 06:46:16 GMT", "version": "v1" } ]
2024-05-28
[ [ "Jung", "Myong Chol", "" ], [ "Zhao", "He", "" ], [ "Dipnall", "Joanna", "" ], [ "Gabbe", "Belinda", "" ], [ "Du", "Lan", "" ] ]
Prompt learning has shown to be an efficient and effective fine-tuning method for vision-language models like CLIP. While numerous studies have focused on the generalisation of these models in few-shot classification, their capability in near out-of-distribution (OOD) detection has been overlooked. A few recent works have highlighted the promising performance of prompt learning in far OOD detection. However, the more challenging task of few-shot near OOD detection has not yet been addressed. In this study, we investigate the near OOD detection capabilities of prompt learning models and observe that commonly used OOD scores have limited performance in near OOD detection. To enhance the performance, we propose a fast and simple post-hoc method that complements existing logit-based scores, improving near OOD detection AUROC by up to 11.67% with minimal computational cost. Our method can be easily applied to any prompt learning model without change in architecture or re-training the models. Comprehensive empirical evaluations across 13 datasets and 8 models demonstrate the effectiveness and adaptability of our method.
1401.3682
Minglai Cai
Holger Boche, Minglai Cai, and Christian Deppe
Broadcast Classical-Quantum Capacity Region of Two-Phase Bidirectional Relaying Channel
null
Quantum Information Processing: Volume 14, Issue 10 (2015), Page 3879-3897
10.1007/s11128-015-1065-2
null
cs.IT math.IT math.QA quant-ph
http://creativecommons.org/licenses/by/4.0/
We study a three-node quantum network which enables bidirectional communication between two nodes with a half-duplex relay node. A decode-and-forward protocol is used to perform the communication in two phases. In the first phase, the messages of two nodes are transmitted to the relay node. In the second phase, the relay node broadcasts a re-encoded composition to the two nodes. We determine the capacity region of the broadcast phase.
[ { "created": "Wed, 15 Jan 2014 17:29:23 GMT", "version": "v1" }, { "created": "Tue, 21 Jan 2014 10:27:07 GMT", "version": "v2" }, { "created": "Mon, 27 Jul 2015 13:02:18 GMT", "version": "v3" }, { "created": "Mon, 28 Sep 2015 15:23:00 GMT", "version": "v4" } ]
2015-09-29
[ [ "Boche", "Holger", "" ], [ "Cai", "Minglai", "" ], [ "Deppe", "Christian", "" ] ]
We study a three-node quantum network which enables bidirectional communication between two nodes with a half-duplex relay node. A decode-and-forward protocol is used to perform the communication in two phases. In the first phase, the messages of two nodes are transmitted to the relay node. In the second phase, the relay node broadcasts a re-encoded composition to the two nodes. We determine the capacity region of the broadcast phase.
2306.01996
Han Wang
Han Wang, Ming Tang, Ke Xu, Quancheng Wang
BandwidthBreach: Unleashing Covert and Side Channels through Cache Bandwidth Exploitation
null
null
null
null
cs.CR
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
In the modern CPU architecture, enhancements such as the Line Fill Buffer (LFB) and Super Queue (SQ), which are designed to track pending cache requests, have significantly boosted performance. To exploit this structures, we deliberately engineered blockages in the L2 to L1d route by controlling LFB conflict and triggering prefetch prediction failures, while consciously dismissing other plausible influencing factors. This approach was subsequently extended to the L3 to L2 and L2 to L1i pathways, resulting in three potent covert channels, termed L2CC, L3CC, and LiCC, with capacities of 10.02 Mbps, 10.37 Mbps, and 1.83 Mbps, respectively. Strikingly, the capacities of L2CC and L3CC surpass those of earlier non-shared-memory-based covert channels, reaching a level comparable to their shared memory-dependent equivalents. Leveraging this congestion further facilitated the extraction of key bits from RSA and EdDSA implementations. Coupled with SpectreV1 and V2, our covert channels effectively evade the majority of traditional Spectre defenses. Their confluence with Branch Prediction (BP) Timing assaults additionally undercuts balanced branch protections, hence broadening their capability to infiltrate a wide range of cryptography libraries.
[ { "created": "Sat, 3 Jun 2023 04:09:07 GMT", "version": "v1" } ]
2023-06-06
[ [ "Wang", "Han", "" ], [ "Tang", "Ming", "" ], [ "Xu", "Ke", "" ], [ "Wang", "Quancheng", "" ] ]
In the modern CPU architecture, enhancements such as the Line Fill Buffer (LFB) and Super Queue (SQ), which are designed to track pending cache requests, have significantly boosted performance. To exploit this structures, we deliberately engineered blockages in the L2 to L1d route by controlling LFB conflict and triggering prefetch prediction failures, while consciously dismissing other plausible influencing factors. This approach was subsequently extended to the L3 to L2 and L2 to L1i pathways, resulting in three potent covert channels, termed L2CC, L3CC, and LiCC, with capacities of 10.02 Mbps, 10.37 Mbps, and 1.83 Mbps, respectively. Strikingly, the capacities of L2CC and L3CC surpass those of earlier non-shared-memory-based covert channels, reaching a level comparable to their shared memory-dependent equivalents. Leveraging this congestion further facilitated the extraction of key bits from RSA and EdDSA implementations. Coupled with SpectreV1 and V2, our covert channels effectively evade the majority of traditional Spectre defenses. Their confluence with Branch Prediction (BP) Timing assaults additionally undercuts balanced branch protections, hence broadening their capability to infiltrate a wide range of cryptography libraries.
2108.06832
Sanjiang Li
Xueqing Yan, Yongming Li, Sanjiang Li
A Fast Algorithm for Computing the Deficiency Number of a Mahjong Hand
32 pages, 3 figures
null
null
null
cs.AI cs.DM cs.MA
http://creativecommons.org/licenses/by/4.0/
The tile-based multiplayer game Mahjong is widely played in Asia and has also become increasingly popular worldwide. Face-to-face or online, each player begins with a hand of 13 tiles and players draw and discard tiles in turn until they complete a winning hand. An important notion in Mahjong is the deficiency number (a.k.a. shanten number in Japanese Mahjong) of a hand, which estimates how many tile changes are necessary to complete the hand into a winning hand. The deficiency number plays an essential role in major decision-making tasks such as selecting a tile to discard. This paper proposes a fast algorithm for computing the deficiency number of a Mahjong hand. Compared with the baseline algorithm, the new algorithm is usually 100 times faster and, more importantly, respects the agent's knowledge about available tiles. The algorithm can be used as a basic procedure in all Mahjong variants by both rule-based and machine learning-based Mahjong AI.
[ { "created": "Sun, 15 Aug 2021 22:44:14 GMT", "version": "v1" } ]
2021-08-17
[ [ "Yan", "Xueqing", "" ], [ "Li", "Yongming", "" ], [ "Li", "Sanjiang", "" ] ]
The tile-based multiplayer game Mahjong is widely played in Asia and has also become increasingly popular worldwide. Face-to-face or online, each player begins with a hand of 13 tiles and players draw and discard tiles in turn until they complete a winning hand. An important notion in Mahjong is the deficiency number (a.k.a. shanten number in Japanese Mahjong) of a hand, which estimates how many tile changes are necessary to complete the hand into a winning hand. The deficiency number plays an essential role in major decision-making tasks such as selecting a tile to discard. This paper proposes a fast algorithm for computing the deficiency number of a Mahjong hand. Compared with the baseline algorithm, the new algorithm is usually 100 times faster and, more importantly, respects the agent's knowledge about available tiles. The algorithm can be used as a basic procedure in all Mahjong variants by both rule-based and machine learning-based Mahjong AI.
1601.04952
Andrea Baronchelli
Vito Trianni, Daniele De Simone, Andreagiovanni Reina, Andrea Baronchelli
Emergence of Consensus in a Multi-Robot Network: from Abstract Models to Empirical Validation
A supporting video is available here: https://mail.google.com/mail/u/0/#search/vito.trianni%40istc.cnr.it/15244cd6f27f0e99?projector=1
Robotics and Automation Letters, IEEE , vol.PP, no.99, pp.1 (2016)
10.1109/LRA.2016.2519537
null
cs.MA cs.RO cs.SI physics.soc-ph
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Consensus dynamics in decentralised multiagent systems are subject to intense studies, and several different models have been proposed and analysed. Among these, the naming game stands out for its simplicity and applicability to a wide range of phenomena and applications, from semiotics to engineering. Despite the wide range of studies available, the implementation of theoretical models in real distributed systems is not always straightforward, as the physical platform imposes several constraints that may have a bearing on the consensus dynamics. In this paper, we investigate the effects of an implementation of the naming game for the kilobot robotic platform, in which we consider concurrent execution of games and physical interferences. Consensus dynamics are analysed in the light of the continuously evolving communication network created by the robots, highlighting how the different regimes crucially depend on the robot density and on their ability to spread widely in the experimental arena. We find that physical interferences reduce the benefits resulting from robot mobility in terms of consensus time, but also result in lower cognitive load for individual agents.
[ { "created": "Tue, 19 Jan 2016 15:29:52 GMT", "version": "v1" } ]
2016-01-21
[ [ "Trianni", "Vito", "" ], [ "De Simone", "Daniele", "" ], [ "Reina", "Andreagiovanni", "" ], [ "Baronchelli", "Andrea", "" ] ]
Consensus dynamics in decentralised multiagent systems are subject to intense studies, and several different models have been proposed and analysed. Among these, the naming game stands out for its simplicity and applicability to a wide range of phenomena and applications, from semiotics to engineering. Despite the wide range of studies available, the implementation of theoretical models in real distributed systems is not always straightforward, as the physical platform imposes several constraints that may have a bearing on the consensus dynamics. In this paper, we investigate the effects of an implementation of the naming game for the kilobot robotic platform, in which we consider concurrent execution of games and physical interferences. Consensus dynamics are analysed in the light of the continuously evolving communication network created by the robots, highlighting how the different regimes crucially depend on the robot density and on their ability to spread widely in the experimental arena. We find that physical interferences reduce the benefits resulting from robot mobility in terms of consensus time, but also result in lower cognitive load for individual agents.
1804.07419
Rafael Menelau Oliveira E Cruz
Felipe N. Walmsley, George D. C. Cavalcanti, Dayvid V. R. Oliveira, Rafael M. O. Cruz and Robert Sabourin
An Ensemble Generation Method Based on Instance Hardness
Paper accepted for publication on IJCNN 2018
null
10.1109/IJCNN.2018.8489269
null
cs.LG cs.AI stat.ML
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
In Machine Learning, ensemble methods have been receiving a great deal of attention. Techniques such as Bagging and Boosting have been successfully applied to a variety of problems. Nevertheless, such techniques are still susceptible to the effects of noise and outliers in the training data. We propose a new method for the generation of pools of classifiers based on Bagging, in which the probability of an instance being selected during the resampling process is inversely proportional to its instance hardness, which can be understood as the likelihood of an instance being misclassified, regardless of the choice of classifier. The goal of the proposed method is to remove noisy data without sacrificing the hard instances which are likely to be found on class boundaries. We evaluate the performance of the method in nineteen public data sets, and compare it to the performance of the Bagging and Random Subspace algorithms. Our experiments show that in high noise scenarios the accuracy of our method is significantly better than that of Bagging.
[ { "created": "Fri, 20 Apr 2018 01:29:47 GMT", "version": "v1" }, { "created": "Mon, 30 Apr 2018 07:18:12 GMT", "version": "v2" } ]
2018-11-01
[ [ "Walmsley", "Felipe N.", "" ], [ "Cavalcanti", "George D. C.", "" ], [ "Oliveira", "Dayvid V. R.", "" ], [ "Cruz", "Rafael M. O.", "" ], [ "Sabourin", "Robert", "" ] ]
In Machine Learning, ensemble methods have been receiving a great deal of attention. Techniques such as Bagging and Boosting have been successfully applied to a variety of problems. Nevertheless, such techniques are still susceptible to the effects of noise and outliers in the training data. We propose a new method for the generation of pools of classifiers based on Bagging, in which the probability of an instance being selected during the resampling process is inversely proportional to its instance hardness, which can be understood as the likelihood of an instance being misclassified, regardless of the choice of classifier. The goal of the proposed method is to remove noisy data without sacrificing the hard instances which are likely to be found on class boundaries. We evaluate the performance of the method in nineteen public data sets, and compare it to the performance of the Bagging and Random Subspace algorithms. Our experiments show that in high noise scenarios the accuracy of our method is significantly better than that of Bagging.
2202.11423
Kailun Yang
Kunyu Peng, Alina Roitberg, Kailun Yang, Jiaming Zhang, Rainer Stiefelhagen
Delving Deep into One-Shot Skeleton-based Action Recognition with Diverse Occlusions
Accepted to IEEE Transactions on Multimedia (TMM). Code is publicly available at https://github.com/KPeng9510/Trans4SOAR
null
null
null
cs.CV cs.RO
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Occlusions are universal disruptions constantly present in the real world. Especially for sparse representations, such as human skeletons, a few occluded points might destroy the geometrical and temporal continuity critically affecting the results. Yet, the research of data-scarce recognition from skeleton sequences, such as one-shot action recognition, does not explicitly consider occlusions despite their everyday pervasiveness. In this work, we explicitly tackle body occlusions for Skeleton-based One-shot Action Recognition (SOAR). We mainly consider two occlusion variants: 1) random occlusions and 2) more realistic occlusions caused by diverse everyday objects, which we generate by projecting the existing IKEA 3D furniture models into the camera coordinate system of the 3D skeletons with different geometric parameters. We leverage the proposed pipeline to blend out portions of skeleton sequences of the three popular action recognition datasets and formalize the first benchmark for SOAR from partially occluded body poses. Another key property of our benchmark are the more realistic occlusions generated by everyday objects, as even in standard recognition from 3D skeletons, only randomly missing joints were considered. We re-evaluate existing state-of-the-art frameworks for SOAR in the light of this new task and further introduce Trans4SOAR - a new transformer-based model which leverages three data streams and mixed attention fusion mechanism to alleviate the adverse effects caused by occlusions. While our experiments demonstrate a clear decline in accuracy with missing skeleton portions, this effect is smaller with Trans4SOAR, which outperforms other architectures on all datasets. Although we specifically focus on occlusions, Trans4SOAR additionally yields state-of-the-art in the standard SOAR without occlusion, surpassing the best published approach by 2.85% on NTU-120.
[ { "created": "Wed, 23 Feb 2022 11:11:54 GMT", "version": "v1" }, { "created": "Wed, 13 Jul 2022 17:34:01 GMT", "version": "v2" }, { "created": "Mon, 9 Jan 2023 20:55:30 GMT", "version": "v3" } ]
2023-01-11
[ [ "Peng", "Kunyu", "" ], [ "Roitberg", "Alina", "" ], [ "Yang", "Kailun", "" ], [ "Zhang", "Jiaming", "" ], [ "Stiefelhagen", "Rainer", "" ] ]
Occlusions are universal disruptions constantly present in the real world. Especially for sparse representations, such as human skeletons, a few occluded points might destroy the geometrical and temporal continuity critically affecting the results. Yet, the research of data-scarce recognition from skeleton sequences, such as one-shot action recognition, does not explicitly consider occlusions despite their everyday pervasiveness. In this work, we explicitly tackle body occlusions for Skeleton-based One-shot Action Recognition (SOAR). We mainly consider two occlusion variants: 1) random occlusions and 2) more realistic occlusions caused by diverse everyday objects, which we generate by projecting the existing IKEA 3D furniture models into the camera coordinate system of the 3D skeletons with different geometric parameters. We leverage the proposed pipeline to blend out portions of skeleton sequences of the three popular action recognition datasets and formalize the first benchmark for SOAR from partially occluded body poses. Another key property of our benchmark are the more realistic occlusions generated by everyday objects, as even in standard recognition from 3D skeletons, only randomly missing joints were considered. We re-evaluate existing state-of-the-art frameworks for SOAR in the light of this new task and further introduce Trans4SOAR - a new transformer-based model which leverages three data streams and mixed attention fusion mechanism to alleviate the adverse effects caused by occlusions. While our experiments demonstrate a clear decline in accuracy with missing skeleton portions, this effect is smaller with Trans4SOAR, which outperforms other architectures on all datasets. Although we specifically focus on occlusions, Trans4SOAR additionally yields state-of-the-art in the standard SOAR without occlusion, surpassing the best published approach by 2.85% on NTU-120.
1210.6192
Kasturika B Ray
Rachita Misra, Kasturika B ray
Textural Approach to Palmprint Identification
9 pages
http://www.ijascse.in/publications-2012--2
null
null
cs.CV cs.CR cs.GR
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Biometrics which use of human physiological characteristics for identifying an individual is now a widespread method of identification and authentication. Biometric identification is a technology which uses several image processing techniques and describes the general procedure for identification and verification using feature extraction, storage and matching from the digitized image of biometric characters such as Finger Print, Face, Iris or Palm Print. The current paper uses palm print biometrics. Here we have presented an identification approach using textural properties of palm print images. The elegance of the method is that the conventional edge detection technique is extended to suitably describe the texture features. In this technique all the characteristics of the palm such as principal lines, edges and wrinkles are considered with equal importance.
[ { "created": "Tue, 23 Oct 2012 10:52:31 GMT", "version": "v1" } ]
2012-10-24
[ [ "Misra", "Rachita", "" ], [ "ray", "Kasturika B", "" ] ]
Biometrics which use of human physiological characteristics for identifying an individual is now a widespread method of identification and authentication. Biometric identification is a technology which uses several image processing techniques and describes the general procedure for identification and verification using feature extraction, storage and matching from the digitized image of biometric characters such as Finger Print, Face, Iris or Palm Print. The current paper uses palm print biometrics. Here we have presented an identification approach using textural properties of palm print images. The elegance of the method is that the conventional edge detection technique is extended to suitably describe the texture features. In this technique all the characteristics of the palm such as principal lines, edges and wrinkles are considered with equal importance.
2102.07312
Shoya Ishimaru
Shoya Ishimaru, Takanori Maruichi, Andreas Dengel and Koichi Kise
Confidence-Aware Learning Assistant
9 pages, 11 figures
null
null
null
cs.HC cs.LG
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Not only correctness but also self-confidence play an important role in improving the quality of knowledge. Undesirable situations such as confident incorrect and unconfident correct knowledge prevent learners from revising their knowledge because it is not always easy for them to perceive the situations. To solve this problem, we propose a system that estimates self-confidence while solving multiple-choice questions by eye tracking and gives feedback about which question should be reviewed carefully. We report the results of three studies measuring its effectiveness. (1) On a well-controlled dataset with 10 participants, our approach detected confidence and unconfidence with 81% and 79% average precision. (2) With the help of 20 participants, we observed that correct answer rates of questions were increased by 14% and 17% by giving feedback about correct answers without confidence and incorrect answers with confidence, respectively. (3) We conducted a large-scale data recording in a private school (72 high school students solved 14,302 questions) to investigate effective features and the number of required training samples.
[ { "created": "Mon, 15 Feb 2021 02:47:11 GMT", "version": "v1" } ]
2021-02-16
[ [ "Ishimaru", "Shoya", "" ], [ "Maruichi", "Takanori", "" ], [ "Dengel", "Andreas", "" ], [ "Kise", "Koichi", "" ] ]
Not only correctness but also self-confidence play an important role in improving the quality of knowledge. Undesirable situations such as confident incorrect and unconfident correct knowledge prevent learners from revising their knowledge because it is not always easy for them to perceive the situations. To solve this problem, we propose a system that estimates self-confidence while solving multiple-choice questions by eye tracking and gives feedback about which question should be reviewed carefully. We report the results of three studies measuring its effectiveness. (1) On a well-controlled dataset with 10 participants, our approach detected confidence and unconfidence with 81% and 79% average precision. (2) With the help of 20 participants, we observed that correct answer rates of questions were increased by 14% and 17% by giving feedback about correct answers without confidence and incorrect answers with confidence, respectively. (3) We conducted a large-scale data recording in a private school (72 high school students solved 14,302 questions) to investigate effective features and the number of required training samples.
1808.06075
Gehui Shen
Gehui Shen, Zhi-Hong Deng, Ting Huang and Xi Chen
Learning to Compose over Tree Structures via POS Tags
null
null
null
null
cs.CL
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Recursive Neural Network (RecNN), a type of models which compose words or phrases recursively over syntactic tree structures, has been proven to have superior ability to obtain sentence representation for a variety of NLP tasks. However, RecNN is born with a thorny problem that a shared compositional function for each node of trees can't capture the complex semantic compositionality so that the expressive power of model is limited. In this paper, in order to address this problem, we propose Tag-Guided HyperRecNN/TreeLSTM (TG-HRecNN/TreeLSTM), which introduces hypernetwork into RecNNs to take as inputs Part-of-Speech (POS) tags of word/phrase and generate the semantic composition parameters dynamically. Experimental results on five datasets for two typical NLP tasks show proposed models both obtain significant improvement compared with RecNN and TreeLSTM consistently. Our TG-HTreeLSTM outperforms all existing RecNN-based models and achieves or is competitive with state-of-the-art on four sentence classification benchmarks. The effectiveness of our models is also demonstrated by qualitative analysis.
[ { "created": "Sat, 18 Aug 2018 11:53:24 GMT", "version": "v1" }, { "created": "Tue, 21 Aug 2018 01:57:49 GMT", "version": "v2" } ]
2018-08-22
[ [ "Shen", "Gehui", "" ], [ "Deng", "Zhi-Hong", "" ], [ "Huang", "Ting", "" ], [ "Chen", "Xi", "" ] ]
Recursive Neural Network (RecNN), a type of models which compose words or phrases recursively over syntactic tree structures, has been proven to have superior ability to obtain sentence representation for a variety of NLP tasks. However, RecNN is born with a thorny problem that a shared compositional function for each node of trees can't capture the complex semantic compositionality so that the expressive power of model is limited. In this paper, in order to address this problem, we propose Tag-Guided HyperRecNN/TreeLSTM (TG-HRecNN/TreeLSTM), which introduces hypernetwork into RecNNs to take as inputs Part-of-Speech (POS) tags of word/phrase and generate the semantic composition parameters dynamically. Experimental results on five datasets for two typical NLP tasks show proposed models both obtain significant improvement compared with RecNN and TreeLSTM consistently. Our TG-HTreeLSTM outperforms all existing RecNN-based models and achieves or is competitive with state-of-the-art on four sentence classification benchmarks. The effectiveness of our models is also demonstrated by qualitative analysis.
2402.17766
Runpei Dong
Zekun Qi, Runpei Dong, Shaochen Zhang, Haoran Geng, Chunrui Han, Zheng Ge, Li Yi, Kaisheng Ma
ShapeLLM: Universal 3D Object Understanding for Embodied Interaction
Accepted at ECCV 2024
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
This paper presents ShapeLLM, the first 3D Multimodal Large Language Model (LLM) designed for embodied interaction, exploring a universal 3D object understanding with 3D point clouds and languages. ShapeLLM is built upon an improved 3D encoder by extending ReCon to ReCon++ that benefits from multi-view image distillation for enhanced geometry understanding. By utilizing ReCon++ as the 3D point cloud input encoder for LLMs, ShapeLLM is trained on constructed instruction-following data and tested on our newly human-curated benchmark, 3D MM-Vet. ReCon++ and ShapeLLM achieve state-of-the-art performance in 3D geometry understanding and language-unified 3D interaction tasks, such as embodied visual grounding. Project page: https://qizekun.github.io/shapellm/
[ { "created": "Tue, 27 Feb 2024 18:57:12 GMT", "version": "v1" }, { "created": "Wed, 6 Mar 2024 15:11:37 GMT", "version": "v2" }, { "created": "Fri, 12 Jul 2024 15:36:15 GMT", "version": "v3" } ]
2024-07-15
[ [ "Qi", "Zekun", "" ], [ "Dong", "Runpei", "" ], [ "Zhang", "Shaochen", "" ], [ "Geng", "Haoran", "" ], [ "Han", "Chunrui", "" ], [ "Ge", "Zheng", "" ], [ "Yi", "Li", "" ], [ "Ma", "Kaisheng", "" ] ]
This paper presents ShapeLLM, the first 3D Multimodal Large Language Model (LLM) designed for embodied interaction, exploring a universal 3D object understanding with 3D point clouds and languages. ShapeLLM is built upon an improved 3D encoder by extending ReCon to ReCon++ that benefits from multi-view image distillation for enhanced geometry understanding. By utilizing ReCon++ as the 3D point cloud input encoder for LLMs, ShapeLLM is trained on constructed instruction-following data and tested on our newly human-curated benchmark, 3D MM-Vet. ReCon++ and ShapeLLM achieve state-of-the-art performance in 3D geometry understanding and language-unified 3D interaction tasks, such as embodied visual grounding. Project page: https://qizekun.github.io/shapellm/
2308.03774
Nick Zhang
Nick Zhang
Knowledge Consilience: One Culture, Two Cultures or Many Cultures?
null
null
null
null
cs.DL
http://creativecommons.org/licenses/by/4.0/
The hostility between the two cultures, scientific and literary, was framed by C.P. Snow in 1959 and later by others. The scientific culture is nowadays often identified with STEM (Science, Technology, Engineering and Mathematics) whereas the literary culture generally refers to humanities and social sciences. Wilson expressed the wish for the unity of knowledge. We put forward the notions of knowledge distance and knowledge consilience threshold to quantitatively measure distance and coupling process between different branches of knowledge. Our findings suggest that the gulf between the two cultures is widening.
[ { "created": "Sun, 30 Jul 2023 11:26:32 GMT", "version": "v1" } ]
2023-08-09
[ [ "Zhang", "Nick", "" ] ]
The hostility between the two cultures, scientific and literary, was framed by C.P. Snow in 1959 and later by others. The scientific culture is nowadays often identified with STEM (Science, Technology, Engineering and Mathematics) whereas the literary culture generally refers to humanities and social sciences. Wilson expressed the wish for the unity of knowledge. We put forward the notions of knowledge distance and knowledge consilience threshold to quantitatively measure distance and coupling process between different branches of knowledge. Our findings suggest that the gulf between the two cultures is widening.
2404.16587
Zhihao Zhu
Zhihao Zhu, Ninglu Shao, Defu Lian, Chenwang Wu, Zheng Liu, Yi Yang, Enhong Chen
Understanding Privacy Risks of Embeddings Induced by Large Language Models
null
null
null
null
cs.CL cs.AI
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Large language models (LLMs) show early signs of artificial general intelligence but struggle with hallucinations. One promising solution to mitigate these hallucinations is to store external knowledge as embeddings, aiding LLMs in retrieval-augmented generation. However, such a solution risks compromising privacy, as recent studies experimentally showed that the original text can be partially reconstructed from text embeddings by pre-trained language models. The significant advantage of LLMs over traditional pre-trained models may exacerbate these concerns. To this end, we investigate the effectiveness of reconstructing original knowledge and predicting entity attributes from these embeddings when LLMs are employed. Empirical findings indicate that LLMs significantly improve the accuracy of two evaluated tasks over those from pre-trained models, regardless of whether the texts are in-distribution or out-of-distribution. This underscores a heightened potential for LLMs to jeopardize user privacy, highlighting the negative consequences of their widespread use. We further discuss preliminary strategies to mitigate this risk.
[ { "created": "Thu, 25 Apr 2024 13:10:48 GMT", "version": "v1" } ]
2024-04-26
[ [ "Zhu", "Zhihao", "" ], [ "Shao", "Ninglu", "" ], [ "Lian", "Defu", "" ], [ "Wu", "Chenwang", "" ], [ "Liu", "Zheng", "" ], [ "Yang", "Yi", "" ], [ "Chen", "Enhong", "" ] ]
Large language models (LLMs) show early signs of artificial general intelligence but struggle with hallucinations. One promising solution to mitigate these hallucinations is to store external knowledge as embeddings, aiding LLMs in retrieval-augmented generation. However, such a solution risks compromising privacy, as recent studies experimentally showed that the original text can be partially reconstructed from text embeddings by pre-trained language models. The significant advantage of LLMs over traditional pre-trained models may exacerbate these concerns. To this end, we investigate the effectiveness of reconstructing original knowledge and predicting entity attributes from these embeddings when LLMs are employed. Empirical findings indicate that LLMs significantly improve the accuracy of two evaluated tasks over those from pre-trained models, regardless of whether the texts are in-distribution or out-of-distribution. This underscores a heightened potential for LLMs to jeopardize user privacy, highlighting the negative consequences of their widespread use. We further discuss preliminary strategies to mitigate this risk.
2406.13474
Junhan Kim
Junhan Kim, Ho-young Kim, Eulrang Cho, Chungman Lee, Joonyoung Kim, Yongkweon Jeon
Attention-aware Post-training Quantization without Backpropagation
20 pages, under review
null
null
null
cs.LG cs.AI
http://creativecommons.org/licenses/by-nc-nd/4.0/
Quantization is a promising solution for deploying large-scale language models (LLMs) on resource-constrained devices. Existing quantization approaches, however, rely on gradient-based optimization, regardless of it being post-training quantization (PTQ) or quantization-aware training (QAT), which becomes problematic for hyper-scale LLMs with billions of parameters. This overhead can be alleviated via recently proposed backpropagation-free PTQ methods; however, their performance is somewhat limited by their lack of consideration of inter-layer dependencies. In this paper, we thus propose a novel PTQ algorithm that considers inter-layer dependencies without relying on backpropagation. The fundamental concept involved is the development of attention-aware Hessian matrices, which facilitates the consideration of inter-layer dependencies within the attention module. Extensive experiments demonstrate that the proposed algorithm significantly outperforms conventional PTQ methods, particularly for low bit-widths.
[ { "created": "Wed, 19 Jun 2024 11:53:21 GMT", "version": "v1" } ]
2024-06-21
[ [ "Kim", "Junhan", "" ], [ "Kim", "Ho-young", "" ], [ "Cho", "Eulrang", "" ], [ "Lee", "Chungman", "" ], [ "Kim", "Joonyoung", "" ], [ "Jeon", "Yongkweon", "" ] ]
Quantization is a promising solution for deploying large-scale language models (LLMs) on resource-constrained devices. Existing quantization approaches, however, rely on gradient-based optimization, regardless of it being post-training quantization (PTQ) or quantization-aware training (QAT), which becomes problematic for hyper-scale LLMs with billions of parameters. This overhead can be alleviated via recently proposed backpropagation-free PTQ methods; however, their performance is somewhat limited by their lack of consideration of inter-layer dependencies. In this paper, we thus propose a novel PTQ algorithm that considers inter-layer dependencies without relying on backpropagation. The fundamental concept involved is the development of attention-aware Hessian matrices, which facilitates the consideration of inter-layer dependencies within the attention module. Extensive experiments demonstrate that the proposed algorithm significantly outperforms conventional PTQ methods, particularly for low bit-widths.
2405.11471
Kento Uchida
Kento Uchida, Kenta Nishihara, Shinichi Shirakawa
CMA-ES with Adaptive Reevaluation for Multiplicative Noise
This paper has been accepted as a full paper at GECCO2024
null
null
null
cs.NE
http://creativecommons.org/licenses/by-sa/4.0/
The covariance matrix adaptation evolution strategy (CMA-ES) is a powerful optimization method for continuous black-box optimization problems. Several noise-handling methods have been proposed to bring out the optimization performance of the CMA-ES on noisy objective functions. The adaptations of the population size and the learning rate are two major approaches that perform well under additive Gaussian noise. The reevaluation technique is another technique that evaluates each solution multiple times. In this paper, we discuss the difference between those methods from the perspective of stochastic relaxation that considers the maximization of the expected utility function. We derive that the set of maximizers of the noise-independent utility, which is used in the reevaluation technique, certainly contains the optimal solution, while the noise-dependent utility, which is used in the population size and leaning rate adaptations, does not satisfy it under multiplicative noise. Based on the discussion, we develop the reevaluation adaptation CMA-ES (RA-CMA-ES), which computes two update directions using half of the evaluations and adapts the number of reevaluations based on the estimated correlation of those two update directions. The numerical simulation shows that the RA-CMA-ES outperforms the comparative method under multiplicative noise, maintaining competitive performance under additive noise.
[ { "created": "Sun, 19 May 2024 07:42:10 GMT", "version": "v1" } ]
2024-05-21
[ [ "Uchida", "Kento", "" ], [ "Nishihara", "Kenta", "" ], [ "Shirakawa", "Shinichi", "" ] ]
The covariance matrix adaptation evolution strategy (CMA-ES) is a powerful optimization method for continuous black-box optimization problems. Several noise-handling methods have been proposed to bring out the optimization performance of the CMA-ES on noisy objective functions. The adaptations of the population size and the learning rate are two major approaches that perform well under additive Gaussian noise. The reevaluation technique is another technique that evaluates each solution multiple times. In this paper, we discuss the difference between those methods from the perspective of stochastic relaxation that considers the maximization of the expected utility function. We derive that the set of maximizers of the noise-independent utility, which is used in the reevaluation technique, certainly contains the optimal solution, while the noise-dependent utility, which is used in the population size and leaning rate adaptations, does not satisfy it under multiplicative noise. Based on the discussion, we develop the reevaluation adaptation CMA-ES (RA-CMA-ES), which computes two update directions using half of the evaluations and adapts the number of reevaluations based on the estimated correlation of those two update directions. The numerical simulation shows that the RA-CMA-ES outperforms the comparative method under multiplicative noise, maintaining competitive performance under additive noise.
2106.06272
Stefano Maria Nicoletti
Stefano M. Nicoletti, Marijn Peppelman, Christina Kolb, Mari\"elle Stoelinga
Model-based Joint Analysis of Safety and Security: Survey and Identification of Gaps
null
null
null
null
cs.CR
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
We survey the state-of-the-art on model-based formalisms for safety and security joint analysis, where safety refers to the absence of unintended failures, and security to absence of malicious attacks. We conduct a thorough literature review and - as a result - we consider fourteen model-based formalisms and compare them with respect to several criteria: (1) Modelling capabilities and Expressiveness: which phenomena can be expressed in these formalisms? To which extent can they capture safety-security interactions? (2) Analytical capabilities: which analysis types are supported? (3) Practical applicability: to what extent have the formalisms been used to analyze small or larger case studies? Furthermore, (1) we present more precise definitions for safety-security dependencies in tree-like formalisms; (2) we showcase the potential of each formalism by modelling the same toy example from the literature and (3) we present our findings and reflect on possible ways to narrow highlighted gaps. In summary, our key findings are the following: (1) the majority of approaches combine tree-like formal models; (2) the exact nature of safety-security interaction is still ill-understood and (3) diverse formalisms can capture different interactions; (4) analyzed formalisms merge modelling constructs from existing safety- and security-specific formalisms, without introducing ad hoc constructs to model safety-security interactions, or (5) metrics to analyze trade offs. Moreover, (6) large case studies representing safety-security interactions are still missing.
[ { "created": "Fri, 11 Jun 2021 09:38:23 GMT", "version": "v1" }, { "created": "Wed, 19 Jan 2022 11:02:54 GMT", "version": "v2" }, { "created": "Mon, 23 Oct 2023 09:34:05 GMT", "version": "v3" } ]
2023-10-24
[ [ "Nicoletti", "Stefano M.", "" ], [ "Peppelman", "Marijn", "" ], [ "Kolb", "Christina", "" ], [ "Stoelinga", "Mariëlle", "" ] ]
We survey the state-of-the-art on model-based formalisms for safety and security joint analysis, where safety refers to the absence of unintended failures, and security to absence of malicious attacks. We conduct a thorough literature review and - as a result - we consider fourteen model-based formalisms and compare them with respect to several criteria: (1) Modelling capabilities and Expressiveness: which phenomena can be expressed in these formalisms? To which extent can they capture safety-security interactions? (2) Analytical capabilities: which analysis types are supported? (3) Practical applicability: to what extent have the formalisms been used to analyze small or larger case studies? Furthermore, (1) we present more precise definitions for safety-security dependencies in tree-like formalisms; (2) we showcase the potential of each formalism by modelling the same toy example from the literature and (3) we present our findings and reflect on possible ways to narrow highlighted gaps. In summary, our key findings are the following: (1) the majority of approaches combine tree-like formal models; (2) the exact nature of safety-security interaction is still ill-understood and (3) diverse formalisms can capture different interactions; (4) analyzed formalisms merge modelling constructs from existing safety- and security-specific formalisms, without introducing ad hoc constructs to model safety-security interactions, or (5) metrics to analyze trade offs. Moreover, (6) large case studies representing safety-security interactions are still missing.
2402.01296
Man-Jie Yuan
Man-Jie Yuan, Zheng Zou, Wei Gao
Bi-CryptoNets: Leveraging Different-Level Privacy for Encrypted Inference
null
null
null
null
cs.LG cs.CR cs.CV
http://creativecommons.org/licenses/by/4.0/
Privacy-preserving neural networks have attracted increasing attention in recent years, and various algorithms have been developed to keep the balance between accuracy, computational complexity and information security from the cryptographic view. This work takes a different view from the input data and structure of neural networks. We decompose the input data (e.g., some images) into sensitive and insensitive segments according to importance and privacy. The sensitive segment includes some important and private information such as human faces and we take strong homomorphic encryption to keep security, whereas the insensitive one contains some background and we add perturbations. We propose the bi-CryptoNets, i.e., plaintext and ciphertext branches, to deal with two segments, respectively, and ciphertext branch could utilize the information from plaintext branch by unidirectional connections. We adopt knowledge distillation for our bi-CryptoNets by transferring representations from a well-trained teacher neural network. Empirical studies show the effectiveness and decrease of inference latency for our bi-CryptoNets.
[ { "created": "Fri, 2 Feb 2024 10:35:05 GMT", "version": "v1" } ]
2024-02-05
[ [ "Yuan", "Man-Jie", "" ], [ "Zou", "Zheng", "" ], [ "Gao", "Wei", "" ] ]
Privacy-preserving neural networks have attracted increasing attention in recent years, and various algorithms have been developed to keep the balance between accuracy, computational complexity and information security from the cryptographic view. This work takes a different view from the input data and structure of neural networks. We decompose the input data (e.g., some images) into sensitive and insensitive segments according to importance and privacy. The sensitive segment includes some important and private information such as human faces and we take strong homomorphic encryption to keep security, whereas the insensitive one contains some background and we add perturbations. We propose the bi-CryptoNets, i.e., plaintext and ciphertext branches, to deal with two segments, respectively, and ciphertext branch could utilize the information from plaintext branch by unidirectional connections. We adopt knowledge distillation for our bi-CryptoNets by transferring representations from a well-trained teacher neural network. Empirical studies show the effectiveness and decrease of inference latency for our bi-CryptoNets.
0805.4219
Grenville Croll
Andrew Hawker
Building Financial Accuracy into Spreadsheets
6 Pages
Proc. European Spreadsheet Risks Int. Grp. (EuSpRIG) 2000 35-40 ISBN:1 86166 158 4
null
null
cs.SE
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Students learning how to apply spreadsheets to accounting problems are not always well served by the built-in financial functions. Problems can arise because of differences between UK and US practice, through anomalies in the functions themselves, and because the promptings of Wizards' engender an attitude of filling in the blanks on the screen, and hoping for the best. Some examples of these problems are described, and suggestions are presented for ways of improving the situation. Principally, it is suggested that spreadsheet prompts and 'Help' screens should offer integrated guidance, covering some aspects of financial practice, as well as matters of spreadsheet technique.
[ { "created": "Tue, 27 May 2008 21:11:48 GMT", "version": "v1" } ]
2008-05-29
[ [ "Hawker", "Andrew", "" ] ]
Students learning how to apply spreadsheets to accounting problems are not always well served by the built-in financial functions. Problems can arise because of differences between UK and US practice, through anomalies in the functions themselves, and because the promptings of Wizards' engender an attitude of filling in the blanks on the screen, and hoping for the best. Some examples of these problems are described, and suggestions are presented for ways of improving the situation. Principally, it is suggested that spreadsheet prompts and 'Help' screens should offer integrated guidance, covering some aspects of financial practice, as well as matters of spreadsheet technique.
2209.07749
Diana Negoescu
Diana M. Negoescu, Pasha Khosravi, Shadow Zhao, Nanyu Chen, Parvez Ahammad, Humberto Gonzalez
Sales Channel Optimization via Simulations Based on Observational Data with Delayed Rewards: A Case Study at LinkedIn
Accepted at REVEAL'22 Workshop (16th ACM Conference on Recommender Systems - RecSys 2022)
null
null
null
cs.LG
http://creativecommons.org/licenses/by/4.0/
Training models on data obtained from randomized experiments is ideal for making good decisions. However, randomized experiments are often time-consuming, expensive, risky, infeasible or unethical to perform, leaving decision makers little choice but to rely on observational data collected under historical policies when training models. This opens questions regarding not only which decision-making policies would perform best in practice, but also regarding the impact of different data collection protocols on the performance of various policies trained on the data, or the robustness of policy performance with respect to changes in problem characteristics such as action- or reward- specific delays in observing outcomes. We aim to answer such questions for the problem of optimizing sales channel allocations at LinkedIn, where sales accounts (leads) need to be allocated to one of three channels, with the goal of maximizing the number of successful conversions over a period of time. A key problem feature constitutes the presence of stochastic delays in observing allocation outcomes, whose distribution is both channel- and outcome- dependent. We built a discrete-time simulation that can handle our problem features and used it to evaluate: a) a historical rule-based policy; b) a supervised machine learning policy (XGBoost); and c) multi-armed bandit (MAB) policies, under different scenarios involving: i) data collection used for training (observational vs randomized); ii) lead conversion scenarios; iii) delay distributions. Our simulation results indicate that LinUCB, a simple MAB policy, consistently outperforms the other policies, achieving a 18-47% lift relative to a rule-based policy
[ { "created": "Fri, 16 Sep 2022 07:08:37 GMT", "version": "v1" } ]
2022-09-19
[ [ "Negoescu", "Diana M.", "" ], [ "Khosravi", "Pasha", "" ], [ "Zhao", "Shadow", "" ], [ "Chen", "Nanyu", "" ], [ "Ahammad", "Parvez", "" ], [ "Gonzalez", "Humberto", "" ] ]
Training models on data obtained from randomized experiments is ideal for making good decisions. However, randomized experiments are often time-consuming, expensive, risky, infeasible or unethical to perform, leaving decision makers little choice but to rely on observational data collected under historical policies when training models. This opens questions regarding not only which decision-making policies would perform best in practice, but also regarding the impact of different data collection protocols on the performance of various policies trained on the data, or the robustness of policy performance with respect to changes in problem characteristics such as action- or reward- specific delays in observing outcomes. We aim to answer such questions for the problem of optimizing sales channel allocations at LinkedIn, where sales accounts (leads) need to be allocated to one of three channels, with the goal of maximizing the number of successful conversions over a period of time. A key problem feature constitutes the presence of stochastic delays in observing allocation outcomes, whose distribution is both channel- and outcome- dependent. We built a discrete-time simulation that can handle our problem features and used it to evaluate: a) a historical rule-based policy; b) a supervised machine learning policy (XGBoost); and c) multi-armed bandit (MAB) policies, under different scenarios involving: i) data collection used for training (observational vs randomized); ii) lead conversion scenarios; iii) delay distributions. Our simulation results indicate that LinUCB, a simple MAB policy, consistently outperforms the other policies, achieving a 18-47% lift relative to a rule-based policy
2210.05791
Shalaleh Rismani
Renee Shelby, Shalaleh Rismani, Kathryn Henne, AJung Moon, Negar Rostamzadeh, Paul Nicholas, N'Mah Yilla, Jess Gallegos, Andrew Smart, Emilio Garcia, Gurleen Virk
Sociotechnical Harms of Algorithmic Systems: Scoping a Taxonomy for Harm Reduction
null
null
null
null
cs.HC cs.GL
http://creativecommons.org/licenses/by/4.0/
Understanding the landscape of potential harms from algorithmic systems enables practitioners to better anticipate consequences of the systems they build. It also supports the prospect of incorporating controls to help minimize harms that emerge from the interplay of technologies and social and cultural dynamics. A growing body of scholarship has identified a wide range of harms across different algorithmic technologies. However, computing research and practitioners lack a high level and synthesized overview of harms from algorithmic systems. Based on a scoping review of computing research $(n=172)$, we present an applied taxonomy of sociotechnical harms to support a more systematic surfacing of potential harms in algorithmic systems. The final taxonomy builds on and refers to existing taxonomies, classifications, and terminologies. Five major themes related to sociotechnical harms - representational, allocative, quality-of-service, interpersonal harms, and social system/societal harms - and sub-themes are presented along with a description of these categories. We conclude with a discussion of challenges and opportunities for future research.
[ { "created": "Tue, 11 Oct 2022 21:22:30 GMT", "version": "v1" }, { "created": "Thu, 9 Feb 2023 03:31:48 GMT", "version": "v2" }, { "created": "Wed, 19 Jul 2023 02:56:32 GMT", "version": "v3" } ]
2023-07-21
[ [ "Shelby", "Renee", "" ], [ "Rismani", "Shalaleh", "" ], [ "Henne", "Kathryn", "" ], [ "Moon", "AJung", "" ], [ "Rostamzadeh", "Negar", "" ], [ "Nicholas", "Paul", "" ], [ "Yilla", "N'Mah", "" ], [ "Gallegos", "Jess", "" ], [ "Smart", "Andrew", "" ], [ "Garcia", "Emilio", "" ], [ "Virk", "Gurleen", "" ] ]
Understanding the landscape of potential harms from algorithmic systems enables practitioners to better anticipate consequences of the systems they build. It also supports the prospect of incorporating controls to help minimize harms that emerge from the interplay of technologies and social and cultural dynamics. A growing body of scholarship has identified a wide range of harms across different algorithmic technologies. However, computing research and practitioners lack a high level and synthesized overview of harms from algorithmic systems. Based on a scoping review of computing research $(n=172)$, we present an applied taxonomy of sociotechnical harms to support a more systematic surfacing of potential harms in algorithmic systems. The final taxonomy builds on and refers to existing taxonomies, classifications, and terminologies. Five major themes related to sociotechnical harms - representational, allocative, quality-of-service, interpersonal harms, and social system/societal harms - and sub-themes are presented along with a description of these categories. We conclude with a discussion of challenges and opportunities for future research.
2112.00270
Jamison Ebert
Vamsi K. Amalladinne, Jamison R. Ebert, Jean-Francois Chamberland, and Krishna R. Narayanan
An Enhanced Decoding Algorithm for Coded Compressed Sensing with Applications to Unsourced Random Access
Submitted to MDPI Sensors
null
null
null
cs.IT math.IT
http://creativecommons.org/licenses/by-nc-nd/4.0/
Unsourced random access (URA) has emerged as a pragmatic framework for next-generation distributed sensor networks. Within URA, concatenated coding structures are often employed to ensure that the central base station can accurately recover the set of sent codewords during a given transmission period. Many URA algorithms employ independent inner and outer decoders, which can help reduce computational complexity at the expense of a decay in performance. In this article, an enhanced decoding algorithm is presented for a concatenated coding structure consisting of a wide range of inner codes and an outer tree-based code. It is shown that this algorithmic enhancement has the potential to simultaneously improve error performance and decrease the computational complexity of the decoder. This enhanced decoding algorithm is applied to two existing URA algorithms and the performance benefits of the algorithm are characterized. Findings are supported by numerical simulations.
[ { "created": "Wed, 1 Dec 2021 04:30:30 GMT", "version": "v1" } ]
2021-12-02
[ [ "Amalladinne", "Vamsi K.", "" ], [ "Ebert", "Jamison R.", "" ], [ "Chamberland", "Jean-Francois", "" ], [ "Narayanan", "Krishna R.", "" ] ]
Unsourced random access (URA) has emerged as a pragmatic framework for next-generation distributed sensor networks. Within URA, concatenated coding structures are often employed to ensure that the central base station can accurately recover the set of sent codewords during a given transmission period. Many URA algorithms employ independent inner and outer decoders, which can help reduce computational complexity at the expense of a decay in performance. In this article, an enhanced decoding algorithm is presented for a concatenated coding structure consisting of a wide range of inner codes and an outer tree-based code. It is shown that this algorithmic enhancement has the potential to simultaneously improve error performance and decrease the computational complexity of the decoder. This enhanced decoding algorithm is applied to two existing URA algorithms and the performance benefits of the algorithm are characterized. Findings are supported by numerical simulations.
1102.4121
EPTCS
Laurent Doyen (LSV, ENS Cachan & CNRS, France), Thierry Massart (Universit\'e Libre de Bruxelles, Belgium), Mahsa Shirmohammadi (Universit\'e Libre de Bruxelles, Belgium)
Synchronizing Objectives for Markov Decision Processes
In Proceedings iWIGP 2011, arXiv:1102.3741
EPTCS 50, 2011, pp. 61-75
10.4204/EPTCS.50.5
null
cs.LO cs.CC
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
We introduce synchronizing objectives for Markov decision processes (MDP). Intuitively, a synchronizing objective requires that eventually, at every step there is a state which concentrates almost all the probability mass. In particular, it implies that the probabilistic system behaves in the long run like a deterministic system: eventually, the current state of the MDP can be identified with almost certainty. We study the problem of deciding the existence of a strategy to enforce a synchronizing objective in MDPs. We show that the problem is decidable for general strategies, as well as for blind strategies where the player cannot observe the current state of the MDP. We also show that pure strategies are sufficient, but memory may be necessary.
[ { "created": "Mon, 21 Feb 2011 02:30:55 GMT", "version": "v1" } ]
2011-02-22
[ [ "Doyen", "Laurent", "", "LSV, ENS Cachan & CNRS, France" ], [ "Massart", "Thierry", "", "Université Libre de Bruxelles, Belgium" ], [ "Shirmohammadi", "Mahsa", "", "Université\n Libre de Bruxelles, Belgium" ] ]
We introduce synchronizing objectives for Markov decision processes (MDP). Intuitively, a synchronizing objective requires that eventually, at every step there is a state which concentrates almost all the probability mass. In particular, it implies that the probabilistic system behaves in the long run like a deterministic system: eventually, the current state of the MDP can be identified with almost certainty. We study the problem of deciding the existence of a strategy to enforce a synchronizing objective in MDPs. We show that the problem is decidable for general strategies, as well as for blind strategies where the player cannot observe the current state of the MDP. We also show that pure strategies are sufficient, but memory may be necessary.
cs/0510044
Andrea Montanari
Andrea Montanari, Balaji Prabhakar, David Tse
Belief Propagation Based Multi--User Detection
9 pages, 4 eps figures. Forty-third Allerton Conference on Communications, Control and Computing, invited paper
null
null
null
cs.IT math.IT
null
We apply belief propagation (BP) to multi--user detection in a spread spectrum system, under the assumption of Gaussian symbols. We prove that BP is both convergent and allows to estimate the correct conditional expectation of the input symbols. It is therefore an optimal --minimum mean square error-- detection algorithm. This suggests the possibility of designing BP detection algorithms for more general systems. As a byproduct we rederive the Tse-Hanly formula for minimum mean square error without any recourse to random matrix theory.
[ { "created": "Sun, 16 Oct 2005 16:05:31 GMT", "version": "v1" }, { "created": "Mon, 22 May 2006 10:56:18 GMT", "version": "v2" } ]
2007-07-13
[ [ "Montanari", "Andrea", "" ], [ "Prabhakar", "Balaji", "" ], [ "Tse", "David", "" ] ]
We apply belief propagation (BP) to multi--user detection in a spread spectrum system, under the assumption of Gaussian symbols. We prove that BP is both convergent and allows to estimate the correct conditional expectation of the input symbols. It is therefore an optimal --minimum mean square error-- detection algorithm. This suggests the possibility of designing BP detection algorithms for more general systems. As a byproduct we rederive the Tse-Hanly formula for minimum mean square error without any recourse to random matrix theory.
2003.12299
Youngjae Yu
Youngjae Yu, Seunghwan Lee, Yuncheol Choi, Gunhee Kim
CurlingNet: Compositional Learning between Images and Text for Fashion IQ Data
4 pages, 4 figures, ICCV 2019 Linguistics Meets image and video retrieval workshop, Fashion IQ challenge
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
We present an approach named CurlingNet that can measure the semantic distance of composition of image-text embedding. In order to learn an effective image-text composition for the data in the fashion domain, our model proposes two key components as follows. First, the Delivery makes the transition of a source image in an embedding space. Second, the Sweeping emphasizes query-related components of fashion images in the embedding space. We utilize a channel-wise gating mechanism to make it possible. Our single model outperforms previous state-of-the-art image-text composition models including TIRG and FiLM. We participate in the first fashion-IQ challenge in ICCV 2019, for which ensemble of our model achieves one of the best performances.
[ { "created": "Fri, 27 Mar 2020 09:36:32 GMT", "version": "v1" }, { "created": "Mon, 30 Mar 2020 04:35:16 GMT", "version": "v2" } ]
2020-03-31
[ [ "Yu", "Youngjae", "" ], [ "Lee", "Seunghwan", "" ], [ "Choi", "Yuncheol", "" ], [ "Kim", "Gunhee", "" ] ]
We present an approach named CurlingNet that can measure the semantic distance of composition of image-text embedding. In order to learn an effective image-text composition for the data in the fashion domain, our model proposes two key components as follows. First, the Delivery makes the transition of a source image in an embedding space. Second, the Sweeping emphasizes query-related components of fashion images in the embedding space. We utilize a channel-wise gating mechanism to make it possible. Our single model outperforms previous state-of-the-art image-text composition models including TIRG and FiLM. We participate in the first fashion-IQ challenge in ICCV 2019, for which ensemble of our model achieves one of the best performances.
1309.4035
Peter Turney
Peter D. Turney
Domain and Function: A Dual-Space Model of Semantic Relations and Compositions
null
Journal of Artificial Intelligence Research (JAIR), (2012), 44, 533-585
10.1613/jair.3640
null
cs.CL cs.AI cs.LG
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Given appropriate representations of the semantic relations between carpenter and wood and between mason and stone (for example, vectors in a vector space model), a suitable algorithm should be able to recognize that these relations are highly similar (carpenter is to wood as mason is to stone; the relations are analogous). Likewise, with representations of dog, house, and kennel, an algorithm should be able to recognize that the semantic composition of dog and house, dog house, is highly similar to kennel (dog house and kennel are synonymous). It seems that these two tasks, recognizing relations and compositions, are closely connected. However, up to now, the best models for relations are significantly different from the best models for compositions. In this paper, we introduce a dual-space model that unifies these two tasks. This model matches the performance of the best previous models for relations and compositions. The dual-space model consists of a space for measuring domain similarity and a space for measuring function similarity. Carpenter and wood share the same domain, the domain of carpentry. Mason and stone share the same domain, the domain of masonry. Carpenter and mason share the same function, the function of artisans. Wood and stone share the same function, the function of materials. In the composition dog house, kennel has some domain overlap with both dog and house (the domains of pets and buildings). The function of kennel is similar to the function of house (the function of shelters). By combining domain and function similarities in various ways, we can model relations, compositions, and other aspects of semantics.
[ { "created": "Mon, 16 Sep 2013 16:51:02 GMT", "version": "v1" } ]
2013-09-17
[ [ "Turney", "Peter D.", "" ] ]
Given appropriate representations of the semantic relations between carpenter and wood and between mason and stone (for example, vectors in a vector space model), a suitable algorithm should be able to recognize that these relations are highly similar (carpenter is to wood as mason is to stone; the relations are analogous). Likewise, with representations of dog, house, and kennel, an algorithm should be able to recognize that the semantic composition of dog and house, dog house, is highly similar to kennel (dog house and kennel are synonymous). It seems that these two tasks, recognizing relations and compositions, are closely connected. However, up to now, the best models for relations are significantly different from the best models for compositions. In this paper, we introduce a dual-space model that unifies these two tasks. This model matches the performance of the best previous models for relations and compositions. The dual-space model consists of a space for measuring domain similarity and a space for measuring function similarity. Carpenter and wood share the same domain, the domain of carpentry. Mason and stone share the same domain, the domain of masonry. Carpenter and mason share the same function, the function of artisans. Wood and stone share the same function, the function of materials. In the composition dog house, kennel has some domain overlap with both dog and house (the domains of pets and buildings). The function of kennel is similar to the function of house (the function of shelters). By combining domain and function similarities in various ways, we can model relations, compositions, and other aspects of semantics.
2004.00794
Zhonghao Wang
Zhonghao Wang, Yunchao Wei, Rogerior Feris, Jinjun Xiong, Wen-Mei Hwu, Thomas S. Huang, Humphrey Shi
Alleviating Semantic-level Shift: A Semi-supervised Domain Adaptation Method for Semantic Segmentation
CVPRW 2020
null
null
null
cs.CV cs.LG
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Learning segmentation from synthetic data and adapting to real data can significantly relieve human efforts in labelling pixel-level masks. A key challenge of this task is how to alleviate the data distribution discrepancy between the source and target domains, i.e. reducing domain shift. The common approach to this problem is to minimize the discrepancy between feature distributions from different domains through adversarial training. However, directly aligning the feature distribution globally cannot guarantee consistency from a local view (i.e. semantic-level), which prevents certain semantic knowledge learned on the source domain from being applied to the target domain. To tackle this issue, we propose a semi-supervised approach named Alleviating Semantic-level Shift (ASS), which can successfully promote the distribution consistency from both global and local views. Specifically, leveraging a small number of labeled data from the target domain, we directly extract semantic-level feature representations from both the source and the target domains by averaging the features corresponding to same categories advised by pixel-level masks. We then feed the produced features to the discriminator to conduct semantic-level adversarial learning, which collaborates with the adversarial learning from the global view to better alleviate the domain shift. We apply our ASS to two domain adaptation tasks, from GTA5 to Cityscapes and from Synthia to Cityscapes. Extensive experiments demonstrate that: (1) ASS can significantly outperform the current unsupervised state-of-the-arts by employing a small number of annotated samples from the target domain; (2) ASS can beat the oracle model trained on the whole target dataset by over 3 points by augmenting the synthetic source data with annotated samples from the target domain without suffering from the prevalent problem of overfitting to the source domain.
[ { "created": "Thu, 2 Apr 2020 03:25:05 GMT", "version": "v1" }, { "created": "Tue, 9 Jun 2020 22:38:27 GMT", "version": "v2" } ]
2020-06-11
[ [ "Wang", "Zhonghao", "" ], [ "Wei", "Yunchao", "" ], [ "Feris", "Rogerior", "" ], [ "Xiong", "Jinjun", "" ], [ "Hwu", "Wen-Mei", "" ], [ "Huang", "Thomas S.", "" ], [ "Shi", "Humphrey", "" ] ]
Learning segmentation from synthetic data and adapting to real data can significantly relieve human efforts in labelling pixel-level masks. A key challenge of this task is how to alleviate the data distribution discrepancy between the source and target domains, i.e. reducing domain shift. The common approach to this problem is to minimize the discrepancy between feature distributions from different domains through adversarial training. However, directly aligning the feature distribution globally cannot guarantee consistency from a local view (i.e. semantic-level), which prevents certain semantic knowledge learned on the source domain from being applied to the target domain. To tackle this issue, we propose a semi-supervised approach named Alleviating Semantic-level Shift (ASS), which can successfully promote the distribution consistency from both global and local views. Specifically, leveraging a small number of labeled data from the target domain, we directly extract semantic-level feature representations from both the source and the target domains by averaging the features corresponding to same categories advised by pixel-level masks. We then feed the produced features to the discriminator to conduct semantic-level adversarial learning, which collaborates with the adversarial learning from the global view to better alleviate the domain shift. We apply our ASS to two domain adaptation tasks, from GTA5 to Cityscapes and from Synthia to Cityscapes. Extensive experiments demonstrate that: (1) ASS can significantly outperform the current unsupervised state-of-the-arts by employing a small number of annotated samples from the target domain; (2) ASS can beat the oracle model trained on the whole target dataset by over 3 points by augmenting the synthetic source data with annotated samples from the target domain without suffering from the prevalent problem of overfitting to the source domain.
1707.03642
Longfei Zhou
L. Zhou, L. Zheng, X. Wang, W. Jiang, and W. Luo
Coordinated Multicell Multicast Beamforming Based on Manifold Optimization
This paper is already available in IEEE Commmunication Letter. See http://ieeexplore.ieee.org/document/7898517/
null
10.1109/LCOMM.2017.2693374
null
cs.IT math.IT
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Multicast beamforming is a key technology for next-generation wireless cellular networks to support high-rate content distribution services. In this paper, the coordinated downlink multicast beamforming design in multicell networks is considered. The goal is to maximize the minimum signal-to-interference-plus-noise ratio of all users under individual base station power constraints. We exploit the fractional form of the objective function and geometric properties of the con-straints to reformulate the problem as a parametric manifold optimization program. Afterwards we propose a low-complexity Dinkelbach-type algorithm combined with adaptive exponential smoothing and Riemannian conjugate gradient iteration, which is guaranteed to converge. Numerical experiments show that the proposed algorithm outperforms the existing SDP-based method and DC-programming-based method and achieves near-optimal performance.
[ { "created": "Wed, 12 Jul 2017 10:56:50 GMT", "version": "v1" } ]
2017-07-13
[ [ "Zhou", "L.", "" ], [ "Zheng", "L.", "" ], [ "Wang", "X.", "" ], [ "Jiang", "W.", "" ], [ "Luo", "W.", "" ] ]
Multicast beamforming is a key technology for next-generation wireless cellular networks to support high-rate content distribution services. In this paper, the coordinated downlink multicast beamforming design in multicell networks is considered. The goal is to maximize the minimum signal-to-interference-plus-noise ratio of all users under individual base station power constraints. We exploit the fractional form of the objective function and geometric properties of the con-straints to reformulate the problem as a parametric manifold optimization program. Afterwards we propose a low-complexity Dinkelbach-type algorithm combined with adaptive exponential smoothing and Riemannian conjugate gradient iteration, which is guaranteed to converge. Numerical experiments show that the proposed algorithm outperforms the existing SDP-based method and DC-programming-based method and achieves near-optimal performance.
1703.07384
Vishal Jain
Vishal Jain and Dr. Mayank Singh
Ontology Based Pivoted normalization using Vector Based Approach for information Retrieval
null
7th International Conference on Advanced Computing and Communication Technologies, 16th November, 2013
null
null
cs.IR cs.AI
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
The proposed methodology is procedural i.e. it follows finite number of steps that extracts relevant documents according to users query. It is based on principles of Data Mining for analyzing web data. Data Mining first adapts integration of data to generate warehouse. Then, it extracts useful information with the help of algorithm. The task of representing extracted documents is done by using Vector Based Statistical Approach that represents each document in set of Terms.
[ { "created": "Tue, 21 Mar 2017 18:34:34 GMT", "version": "v1" } ]
2017-03-23
[ [ "Jain", "Vishal", "" ], [ "Singh", "Dr. Mayank", "" ] ]
The proposed methodology is procedural i.e. it follows finite number of steps that extracts relevant documents according to users query. It is based on principles of Data Mining for analyzing web data. Data Mining first adapts integration of data to generate warehouse. Then, it extracts useful information with the help of algorithm. The task of representing extracted documents is done by using Vector Based Statistical Approach that represents each document in set of Terms.
2312.10273
Rui Jin
Rui Jin, Yong Liao, and Pengyuan Zhou
User Authentication and Identity Inconsistency Detection via Mouse-trajectory Similarity Measurement
null
null
null
null
cs.CR
http://creativecommons.org/licenses/by/4.0/
Completely Automated Public Turing Test To Tell Computers and Humans Apart (CAPTCHA) is a type of challenge-response test widely used in authentication systems. A well-known challenge it faces is the CAPTCHA farm, where workers are hired to solve CAPTCHAs manually. In this work, we propose to tackle this challenge from a novel perspective, converting CAPTCHA farm detection to identity inconsistency detection, which essentially becomes an authentication process. Specifically, we develop a novel embedding model, which measures the similarity between mouse trajectories collected during the session and when registering/solving CAPTCHA, to authenticate and detect identity inconsistency. Moreover, unlike most existing works that employ a separate mouse movement classifier for each individual user, which brings in considerable costs when serving a large number of users, our model performs detection tasks using only one classifier for all users, significantly reducing the cost. Experiment results validate the superiority of our method over the state-of-the-art time series classification methods, achieving 94.3% and 97.7% of AUC in identity and authentication inconsistency detection, respectively.
[ { "created": "Sat, 16 Dec 2023 00:28:10 GMT", "version": "v1" } ]
2023-12-19
[ [ "Jin", "Rui", "" ], [ "Liao", "Yong", "" ], [ "Zhou", "Pengyuan", "" ] ]
Completely Automated Public Turing Test To Tell Computers and Humans Apart (CAPTCHA) is a type of challenge-response test widely used in authentication systems. A well-known challenge it faces is the CAPTCHA farm, where workers are hired to solve CAPTCHAs manually. In this work, we propose to tackle this challenge from a novel perspective, converting CAPTCHA farm detection to identity inconsistency detection, which essentially becomes an authentication process. Specifically, we develop a novel embedding model, which measures the similarity between mouse trajectories collected during the session and when registering/solving CAPTCHA, to authenticate and detect identity inconsistency. Moreover, unlike most existing works that employ a separate mouse movement classifier for each individual user, which brings in considerable costs when serving a large number of users, our model performs detection tasks using only one classifier for all users, significantly reducing the cost. Experiment results validate the superiority of our method over the state-of-the-art time series classification methods, achieving 94.3% and 97.7% of AUC in identity and authentication inconsistency detection, respectively.
1410.8808
Paolo Pareti Mr.
Paolo Pareti, Ewan Klein and Adam Barker
A Semantic Web of Know-How: Linked Data for Community-Centric Tasks
6th International Workshop on Web Intelligence & Communities (WIC14), Proceedings of the companion publication of the 23rd International Conference on World Wide Web (WWW 2014)
null
null
null
cs.AI cs.CL
http://creativecommons.org/licenses/by-nc-sa/3.0/
This paper proposes a novel framework for representing community know-how on the Semantic Web. Procedural knowledge generated by web communities typically takes the form of natural language instructions or videos and is largely unstructured. The absence of semantic structure impedes the deployment of many useful applications, in particular the ability to discover and integrate know-how automatically. We discuss the characteristics of community know-how and argue that existing knowledge representation frameworks fail to represent it adequately. We present a novel framework for representing the semantic structure of community know-how and demonstrate the feasibility of our approach by providing a concrete implementation which includes a method for automatically acquiring procedural knowledge for real-world tasks.
[ { "created": "Wed, 29 Oct 2014 15:48:40 GMT", "version": "v1" } ]
2014-11-03
[ [ "Pareti", "Paolo", "" ], [ "Klein", "Ewan", "" ], [ "Barker", "Adam", "" ] ]
This paper proposes a novel framework for representing community know-how on the Semantic Web. Procedural knowledge generated by web communities typically takes the form of natural language instructions or videos and is largely unstructured. The absence of semantic structure impedes the deployment of many useful applications, in particular the ability to discover and integrate know-how automatically. We discuss the characteristics of community know-how and argue that existing knowledge representation frameworks fail to represent it adequately. We present a novel framework for representing the semantic structure of community know-how and demonstrate the feasibility of our approach by providing a concrete implementation which includes a method for automatically acquiring procedural knowledge for real-world tasks.
2210.02594
Jeongyeol Kwon
Jeongyeol Kwon, Yonathan Efroni, Constantine Caramanis, Shie Mannor
Reward-Mixing MDPs with a Few Latent Contexts are Learnable
null
null
null
null
cs.LG cs.IT math.IT stat.ML
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
We consider episodic reinforcement learning in reward-mixing Markov decision processes (RMMDPs): at the beginning of every episode nature randomly picks a latent reward model among $M$ candidates and an agent interacts with the MDP throughout the episode for $H$ time steps. Our goal is to learn a near-optimal policy that nearly maximizes the $H$ time-step cumulative rewards in such a model. Previous work established an upper bound for RMMDPs for $M=2$. In this work, we resolve several open questions remained for the RMMDP model. For an arbitrary $M\ge2$, we provide a sample-efficient algorithm--$\texttt{EM}^2$--that outputs an $\epsilon$-optimal policy using $\tilde{O} \left(\epsilon^{-2} \cdot S^d A^d \cdot \texttt{poly}(H, Z)^d \right)$ episodes, where $S, A$ are the number of states and actions respectively, $H$ is the time-horizon, $Z$ is the support size of reward distributions and $d=\min(2M-1,H)$. Our technique is a higher-order extension of the method-of-moments based approach, nevertheless, the design and analysis of the \algname algorithm requires several new ideas beyond existing techniques. We also provide a lower bound of $(SA)^{\Omega(\sqrt{M})} / \epsilon^{2}$ for a general instance of RMMDP, supporting that super-polynomial sample complexity in $M$ is necessary.
[ { "created": "Wed, 5 Oct 2022 22:52:00 GMT", "version": "v1" } ]
2022-10-07
[ [ "Kwon", "Jeongyeol", "" ], [ "Efroni", "Yonathan", "" ], [ "Caramanis", "Constantine", "" ], [ "Mannor", "Shie", "" ] ]
We consider episodic reinforcement learning in reward-mixing Markov decision processes (RMMDPs): at the beginning of every episode nature randomly picks a latent reward model among $M$ candidates and an agent interacts with the MDP throughout the episode for $H$ time steps. Our goal is to learn a near-optimal policy that nearly maximizes the $H$ time-step cumulative rewards in such a model. Previous work established an upper bound for RMMDPs for $M=2$. In this work, we resolve several open questions remained for the RMMDP model. For an arbitrary $M\ge2$, we provide a sample-efficient algorithm--$\texttt{EM}^2$--that outputs an $\epsilon$-optimal policy using $\tilde{O} \left(\epsilon^{-2} \cdot S^d A^d \cdot \texttt{poly}(H, Z)^d \right)$ episodes, where $S, A$ are the number of states and actions respectively, $H$ is the time-horizon, $Z$ is the support size of reward distributions and $d=\min(2M-1,H)$. Our technique is a higher-order extension of the method-of-moments based approach, nevertheless, the design and analysis of the \algname algorithm requires several new ideas beyond existing techniques. We also provide a lower bound of $(SA)^{\Omega(\sqrt{M})} / \epsilon^{2}$ for a general instance of RMMDP, supporting that super-polynomial sample complexity in $M$ is necessary.
2206.03761
Yifan Wang
Yifan Wang, Weizhi Ma, Min Zhang, Yiqun Liu, and Shaoping Ma
A Survey on the Fairness of Recommender Systems
Submitted to the Special Section on Trustworthy Recommendation and Search of ACM TOIS on March 27, 2022 and accepted on June 6
null
10.1145/3547333
null
cs.IR
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Recommender systems are an essential tool to relieve the information overload challenge and play an important role in people's daily lives. Since recommendations involve allocations of social resources (e.g., job recommendation), an important issue is whether recommendations are fair. Unfair recommendations are not only unethical but also harm the long-term interests of the recommender system itself. As a result, fairness issues in recommender systems have recently attracted increasing attention. However, due to multiple complex resource allocation processes and various fairness definitions, the research on fairness in recommendation is scattered. To fill this gap, we review over 60 papers published in top conferences/journals, including TOIS, SIGIR, and WWW. First, we summarize fairness definitions in the recommendation and provide several views to classify fairness issues. Then, we review recommendation datasets and measurements in fairness studies and provide an elaborate taxonomy of fairness methods in the recommendation. Finally, we conclude this survey by outlining some promising future directions.
[ { "created": "Wed, 8 Jun 2022 09:15:08 GMT", "version": "v1" }, { "created": "Sun, 19 Jun 2022 16:14:06 GMT", "version": "v2" } ]
2022-07-12
[ [ "Wang", "Yifan", "" ], [ "Ma", "Weizhi", "" ], [ "Zhang", "Min", "" ], [ "Liu", "Yiqun", "" ], [ "Ma", "Shaoping", "" ] ]
Recommender systems are an essential tool to relieve the information overload challenge and play an important role in people's daily lives. Since recommendations involve allocations of social resources (e.g., job recommendation), an important issue is whether recommendations are fair. Unfair recommendations are not only unethical but also harm the long-term interests of the recommender system itself. As a result, fairness issues in recommender systems have recently attracted increasing attention. However, due to multiple complex resource allocation processes and various fairness definitions, the research on fairness in recommendation is scattered. To fill this gap, we review over 60 papers published in top conferences/journals, including TOIS, SIGIR, and WWW. First, we summarize fairness definitions in the recommendation and provide several views to classify fairness issues. Then, we review recommendation datasets and measurements in fairness studies and provide an elaborate taxonomy of fairness methods in the recommendation. Finally, we conclude this survey by outlining some promising future directions.
1402.5078
Kri\v{s}j\=anis Pr\=usis
Andris Ambainis and Kri\v{s}j\=anis Pr\=usis
A Tight Lower Bound on Certificate Complexity in Terms of Block Sensitivity and Sensitivity
12 pages
null
10.1007/978-3-662-44465-8_4
null
cs.CC
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Sensitivity, certificate complexity and block sensitivity are widely used Boolean function complexity measures. A longstanding open problem, proposed by Nisan and Szegedy, is whether sensitivity and block sensitivity are polynomially related. Motivated by the constructions of functions which achieve the largest known separations, we study the relation between 1-certificate complexity and 0-sensitivity and 0-block sensitivity. Previously the best known lower bound was $C_1(f)\geq \frac{bs_0(f)}{2 s_0(f)}$, achieved by Kenyon and Kutin. We improve this to $C_1(f)\geq \frac{3 bs_0(f)}{2 s_0(f)}$. While this improvement is only by a constant factor, this is quite important, as it precludes achieving a superquadratic separation between $bs(f)$ and $s(f)$ by iterating functions which reach this bound. In addition, this bound is tight, as it matches the construction of Ambainis and Sun up to an additive constant.
[ { "created": "Thu, 20 Feb 2014 17:16:23 GMT", "version": "v1" }, { "created": "Thu, 31 Jul 2014 11:55:55 GMT", "version": "v2" } ]
2015-03-27
[ [ "Ambainis", "Andris", "" ], [ "Prūsis", "Krišjānis", "" ] ]
Sensitivity, certificate complexity and block sensitivity are widely used Boolean function complexity measures. A longstanding open problem, proposed by Nisan and Szegedy, is whether sensitivity and block sensitivity are polynomially related. Motivated by the constructions of functions which achieve the largest known separations, we study the relation between 1-certificate complexity and 0-sensitivity and 0-block sensitivity. Previously the best known lower bound was $C_1(f)\geq \frac{bs_0(f)}{2 s_0(f)}$, achieved by Kenyon and Kutin. We improve this to $C_1(f)\geq \frac{3 bs_0(f)}{2 s_0(f)}$. While this improvement is only by a constant factor, this is quite important, as it precludes achieving a superquadratic separation between $bs(f)$ and $s(f)$ by iterating functions which reach this bound. In addition, this bound is tight, as it matches the construction of Ambainis and Sun up to an additive constant.
2109.06853
Naoya Inoue
Naoya Inoue, Harsh Trivedi, Steven Sinha, Niranjan Balasubramanian and Kentaro Inui
Summarize-then-Answer: Generating Concise Explanations for Multi-hop Reading Comprehension
Accepted to EMNLP2021 Long Paper (Main Track)
null
null
null
cs.CL
http://creativecommons.org/licenses/by/4.0/
How can we generate concise explanations for multi-hop Reading Comprehension (RC)? The current strategies of identifying supporting sentences can be seen as an extractive question-focused summarization of the input text. However, these extractive explanations are not necessarily concise i.e. not minimally sufficient for answering a question. Instead, we advocate for an abstractive approach, where we propose to generate a question-focused, abstractive summary of input paragraphs and then feed it to an RC system. Given a limited amount of human-annotated abstractive explanations, we train the abstractive explainer in a semi-supervised manner, where we start from the supervised model and then train it further through trial and error maximizing a conciseness-promoted reward function. Our experiments demonstrate that the proposed abstractive explainer can generate more compact explanations than an extractive explainer with limited supervision (only 2k instances) while maintaining sufficiency.
[ { "created": "Tue, 14 Sep 2021 17:44:34 GMT", "version": "v1" } ]
2021-09-15
[ [ "Inoue", "Naoya", "" ], [ "Trivedi", "Harsh", "" ], [ "Sinha", "Steven", "" ], [ "Balasubramanian", "Niranjan", "" ], [ "Inui", "Kentaro", "" ] ]
How can we generate concise explanations for multi-hop Reading Comprehension (RC)? The current strategies of identifying supporting sentences can be seen as an extractive question-focused summarization of the input text. However, these extractive explanations are not necessarily concise i.e. not minimally sufficient for answering a question. Instead, we advocate for an abstractive approach, where we propose to generate a question-focused, abstractive summary of input paragraphs and then feed it to an RC system. Given a limited amount of human-annotated abstractive explanations, we train the abstractive explainer in a semi-supervised manner, where we start from the supervised model and then train it further through trial and error maximizing a conciseness-promoted reward function. Our experiments demonstrate that the proposed abstractive explainer can generate more compact explanations than an extractive explainer with limited supervision (only 2k instances) while maintaining sufficiency.
2102.00321
Orestis Papadigenopoulos
Orestis Papadigenopoulos and Constantine Caramanis
Recurrent Submodular Welfare and Matroid Blocking Bandits
Corrected Remark 3.2
null
null
null
cs.LG cs.DS
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
A recent line of research focuses on the study of the stochastic multi-armed bandits problem (MAB), in the case where temporal correlations of specific structure are imposed between the player's actions and the reward distributions of the arms (Kleinberg and Immorlica [FOCS18], Basu et al. [NeurIPS19]). As opposed to the standard MAB setting, where the optimal solution in hindsight can be trivially characterized, these correlations lead to (sub-)optimal solutions that exhibit interesting dynamical patterns -- a phenomenon that yields new challenges both from an algorithmic as well as a learning perspective. In this work, we extend the above direction to a combinatorial bandit setting and study a variant of stochastic MAB, where arms are subject to matroid constraints and each arm becomes unavailable (blocked) for a fixed number of rounds after each play. A natural common generalization of the state-of-the-art for blocking bandits, and that for matroid bandits, yields a $(1-\frac{1}{e})$-approximation for partition matroids, yet it only guarantees a $\frac{1}{2}$-approximation for general matroids. In this paper we develop new algorithmic ideas that allow us to obtain a polynomial-time $(1 - \frac{1}{e})$-approximation algorithm (asymptotically and in expectation) for any matroid, and thus to control the $(1-\frac{1}{e})$-approximate regret. A key ingredient is the technique of correlated (interleaved) scheduling. Along the way, we discover an interesting connection to a variant of Submodular Welfare Maximization, for which we provide (asymptotically) matching upper and lower approximability bounds.
[ { "created": "Sat, 30 Jan 2021 21:51:47 GMT", "version": "v1" }, { "created": "Tue, 16 Feb 2021 06:30:35 GMT", "version": "v2" }, { "created": "Sun, 28 Feb 2021 03:34:19 GMT", "version": "v3" } ]
2021-03-02
[ [ "Papadigenopoulos", "Orestis", "" ], [ "Caramanis", "Constantine", "" ] ]
A recent line of research focuses on the study of the stochastic multi-armed bandits problem (MAB), in the case where temporal correlations of specific structure are imposed between the player's actions and the reward distributions of the arms (Kleinberg and Immorlica [FOCS18], Basu et al. [NeurIPS19]). As opposed to the standard MAB setting, where the optimal solution in hindsight can be trivially characterized, these correlations lead to (sub-)optimal solutions that exhibit interesting dynamical patterns -- a phenomenon that yields new challenges both from an algorithmic as well as a learning perspective. In this work, we extend the above direction to a combinatorial bandit setting and study a variant of stochastic MAB, where arms are subject to matroid constraints and each arm becomes unavailable (blocked) for a fixed number of rounds after each play. A natural common generalization of the state-of-the-art for blocking bandits, and that for matroid bandits, yields a $(1-\frac{1}{e})$-approximation for partition matroids, yet it only guarantees a $\frac{1}{2}$-approximation for general matroids. In this paper we develop new algorithmic ideas that allow us to obtain a polynomial-time $(1 - \frac{1}{e})$-approximation algorithm (asymptotically and in expectation) for any matroid, and thus to control the $(1-\frac{1}{e})$-approximate regret. A key ingredient is the technique of correlated (interleaved) scheduling. Along the way, we discover an interesting connection to a variant of Submodular Welfare Maximization, for which we provide (asymptotically) matching upper and lower approximability bounds.
2404.05587
Wolfgang Otto
Wolfgang Otto, Sharmila Upadhyaya, Stefan Dietze
Enhancing Software-Related Information Extraction via Single-Choice Question Answering with Large Language Models
Accepted at: 1st Workshop on Natural Scientific Language Processing and Research Knowledge Graphs (NSLP 2024) Co-located with Extended Semantic Web Conference (ESWC 2024)
null
null
null
cs.CL
http://creativecommons.org/licenses/by-sa/4.0/
This paper describes our participation in the Shared Task on Software Mentions Disambiguation (SOMD), with a focus on improving relation extraction in scholarly texts through generative Large Language Models (LLMs) using single-choice question-answering. The methodology prioritises the use of in-context learning capabilities of GLMs to extract software-related entities and their descriptive attributes, such as distributive information. Our approach uses Retrieval-Augmented Generation (RAG) techniques and GLMs for Named Entity Recognition (NER) and Attributive NER to identify relationships between extracted software entities, providing a structured solution for analysing software citations in academic literature. The paper provides a detailed description of our approach, demonstrating how using GLMs in a single-choice QA paradigm can greatly enhance IE methodologies. Our participation in the SOMD shared task highlights the importance of precise software citation practices and showcases our system's ability to overcome the challenges of disambiguating and extracting relationships between software mentions. This sets the groundwork for future research and development in this field.
[ { "created": "Mon, 8 Apr 2024 15:00:36 GMT", "version": "v1" }, { "created": "Fri, 19 Apr 2024 23:19:17 GMT", "version": "v2" } ]
2024-04-23
[ [ "Otto", "Wolfgang", "" ], [ "Upadhyaya", "Sharmila", "" ], [ "Dietze", "Stefan", "" ] ]
This paper describes our participation in the Shared Task on Software Mentions Disambiguation (SOMD), with a focus on improving relation extraction in scholarly texts through generative Large Language Models (LLMs) using single-choice question-answering. The methodology prioritises the use of in-context learning capabilities of GLMs to extract software-related entities and their descriptive attributes, such as distributive information. Our approach uses Retrieval-Augmented Generation (RAG) techniques and GLMs for Named Entity Recognition (NER) and Attributive NER to identify relationships between extracted software entities, providing a structured solution for analysing software citations in academic literature. The paper provides a detailed description of our approach, demonstrating how using GLMs in a single-choice QA paradigm can greatly enhance IE methodologies. Our participation in the SOMD shared task highlights the importance of precise software citation practices and showcases our system's ability to overcome the challenges of disambiguating and extracting relationships between software mentions. This sets the groundwork for future research and development in this field.
1909.00440
Abir De
Abir De, Adish Singla, Utkarsh Upadhyay, Manuel Gomez-Rodriguez
Can A User Anticipate What Her Followers Want?
Fixed some typos
null
null
null
cs.SI cs.LG stat.ML
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Whenever a social media user decides to share a story, she is typically pleased to receive likes, comments, shares, or, more generally, feedback from her followers. As a result, she may feel compelled to use the feedback she receives to (re-)estimate her followers' preferences and decides which stories to share next to receive more (positive) feedback. Under which conditions can she succeed? In this work, we first look into this problem from a theoretical perspective and then provide a set of practical algorithms to identify and characterize such behavior in social media. More specifically, we address the above problem from the viewpoint of sequential decision making and utility maximization. For a wide variety of utility functions, we first show that, to succeed, a user needs to actively trade off exploitation-- sharing stories which lead to more (positive) feedback--and exploration-- sharing stories to learn about her followers' preferences. However, exploration is not necessary if a user utilizes the feedback her followers provide to other users in addition to the feedback she receives. Then, we develop a utility estimation framework for observation data, which relies on statistical hypothesis testing to determine whether a user utilizes the feedback she receives from each of her followers to decide what to post next. Experiments on synthetic data illustrate our theoretical findings and show that our estimation framework is able to accurately recover users' underlying utility functions. Experiments on several real datasets gathered from Twitter and Reddit reveal that up to 82% (43%) of the Twitter (Reddit) users in our datasets do use the feedback they receive to decide what to post next.
[ { "created": "Sun, 1 Sep 2019 18:19:51 GMT", "version": "v1" }, { "created": "Thu, 19 Sep 2019 12:28:39 GMT", "version": "v2" } ]
2019-09-20
[ [ "De", "Abir", "" ], [ "Singla", "Adish", "" ], [ "Upadhyay", "Utkarsh", "" ], [ "Gomez-Rodriguez", "Manuel", "" ] ]
Whenever a social media user decides to share a story, she is typically pleased to receive likes, comments, shares, or, more generally, feedback from her followers. As a result, she may feel compelled to use the feedback she receives to (re-)estimate her followers' preferences and decides which stories to share next to receive more (positive) feedback. Under which conditions can she succeed? In this work, we first look into this problem from a theoretical perspective and then provide a set of practical algorithms to identify and characterize such behavior in social media. More specifically, we address the above problem from the viewpoint of sequential decision making and utility maximization. For a wide variety of utility functions, we first show that, to succeed, a user needs to actively trade off exploitation-- sharing stories which lead to more (positive) feedback--and exploration-- sharing stories to learn about her followers' preferences. However, exploration is not necessary if a user utilizes the feedback her followers provide to other users in addition to the feedback she receives. Then, we develop a utility estimation framework for observation data, which relies on statistical hypothesis testing to determine whether a user utilizes the feedback she receives from each of her followers to decide what to post next. Experiments on synthetic data illustrate our theoretical findings and show that our estimation framework is able to accurately recover users' underlying utility functions. Experiments on several real datasets gathered from Twitter and Reddit reveal that up to 82% (43%) of the Twitter (Reddit) users in our datasets do use the feedback they receive to decide what to post next.
2011.03659
Jingnan Shi
Jingnan Shi, Heng Yang, Luca Carlone
ROBIN: a Graph-Theoretic Approach to Reject Outliers in Robust Estimation using Invariants
null
null
null
null
cs.CV cs.RO
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Many estimation problems in robotics, computer vision, and learning require estimating unknown quantities in the face of outliers. Outliers are typically the result of incorrect data association or feature matching, and it is common to have problems where more than 90% of the measurements used for estimation are outliers. While current approaches for robust estimation are able to deal with moderate amounts of outliers, they fail to produce accurate estimates in the presence of many outliers. This paper develops an approach to prune outliers. First, we develop a theory of invariance that allows us to quickly check if a subset of measurements are mutually compatible without explicitly solving the estimation problem. Second, we develop a graph-theoretic framework, where measurements are modeled as vertices and mutual compatibility is captured by edges. We generalize existing results showing that the inliers form a clique in this graph and typically belong to the maximum clique. We also show that in practice the maximum k-core of the compatibility graph provides an approximation of the maximum clique, while being faster to compute in large problems. These two contributions leads to ROBIN, our approach to Reject Outliers Based on INvariants, which allows us to quickly prune outliers in generic estimation problems. We demonstrate ROBIN in four geometric perception problems and show it boosts robustness of existing solvers while running in milliseconds in large problems.
[ { "created": "Sat, 7 Nov 2020 02:09:33 GMT", "version": "v1" }, { "created": "Tue, 23 Mar 2021 20:02:00 GMT", "version": "v2" } ]
2021-03-25
[ [ "Shi", "Jingnan", "" ], [ "Yang", "Heng", "" ], [ "Carlone", "Luca", "" ] ]
Many estimation problems in robotics, computer vision, and learning require estimating unknown quantities in the face of outliers. Outliers are typically the result of incorrect data association or feature matching, and it is common to have problems where more than 90% of the measurements used for estimation are outliers. While current approaches for robust estimation are able to deal with moderate amounts of outliers, they fail to produce accurate estimates in the presence of many outliers. This paper develops an approach to prune outliers. First, we develop a theory of invariance that allows us to quickly check if a subset of measurements are mutually compatible without explicitly solving the estimation problem. Second, we develop a graph-theoretic framework, where measurements are modeled as vertices and mutual compatibility is captured by edges. We generalize existing results showing that the inliers form a clique in this graph and typically belong to the maximum clique. We also show that in practice the maximum k-core of the compatibility graph provides an approximation of the maximum clique, while being faster to compute in large problems. These two contributions leads to ROBIN, our approach to Reject Outliers Based on INvariants, which allows us to quickly prune outliers in generic estimation problems. We demonstrate ROBIN in four geometric perception problems and show it boosts robustness of existing solvers while running in milliseconds in large problems.
1609.01755
Ulf R\"uegg
Adalat Jabrayilov, Sven Mallach, Petra Mutzel, Ulf R\"uegg, and Reinhard von Hanxleden
Compact Layered Drawings of General Directed Graphs
Appears in the Proceedings of the 24th International Symposium on Graph Drawing and Network Visualization (GD 2016)
null
null
null
cs.DS
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
We consider the problem of layering general directed graphs under height and possibly also width constraints. Given a directed graph G = (V,A) and a maximal height, we propose a layering approach that minimizes a weighted sum of the number of reversed arcs, the arc lengths, and the width of the drawing. We call this the Compact Generalized Layering Problem (CGLP). Here, the width of a drawing is defined as the maximum sum of the number of vertices placed on a layer and the number of dummy vertices caused by arcs traversing the layer. The CGLP is NP-hard. We present two MIP models for this problem. The first one (EXT) is our extension of a natural formulation for directed acyclic graphs as suggested by Healy and Nikolov. The second one (CGL) is a new formulation based on partial orderings. Our computational experiments on two benchmark sets show that the CGL formulation can be solved much faster than EXT using standard commercial MIP solvers. Moreover, we suggest a variant of CGL, called MML, that can be seen as a heuristic approach. In our experiments, MML clearly improves on CGL in terms of running time while it does not considerably increase the average arc lengths and widths of the layouts although it solves a slightly different problem where the dummy vertices are not taken into account.
[ { "created": "Mon, 29 Aug 2016 22:09:37 GMT", "version": "v1" } ]
2016-09-08
[ [ "Jabrayilov", "Adalat", "" ], [ "Mallach", "Sven", "" ], [ "Mutzel", "Petra", "" ], [ "Rüegg", "Ulf", "" ], [ "von Hanxleden", "Reinhard", "" ] ]
We consider the problem of layering general directed graphs under height and possibly also width constraints. Given a directed graph G = (V,A) and a maximal height, we propose a layering approach that minimizes a weighted sum of the number of reversed arcs, the arc lengths, and the width of the drawing. We call this the Compact Generalized Layering Problem (CGLP). Here, the width of a drawing is defined as the maximum sum of the number of vertices placed on a layer and the number of dummy vertices caused by arcs traversing the layer. The CGLP is NP-hard. We present two MIP models for this problem. The first one (EXT) is our extension of a natural formulation for directed acyclic graphs as suggested by Healy and Nikolov. The second one (CGL) is a new formulation based on partial orderings. Our computational experiments on two benchmark sets show that the CGL formulation can be solved much faster than EXT using standard commercial MIP solvers. Moreover, we suggest a variant of CGL, called MML, that can be seen as a heuristic approach. In our experiments, MML clearly improves on CGL in terms of running time while it does not considerably increase the average arc lengths and widths of the layouts although it solves a slightly different problem where the dummy vertices are not taken into account.
2204.09009
Ishay Haviv
Ishay Haviv
Fixed-Parameter Algorithms for the Kneser and Schrijver Problems
31 pages. This paper includes and extends the content of arXiv:2204.06761
null
null
null
cs.DS math.CO
http://creativecommons.org/licenses/by/4.0/
The Kneser graph $K(n,k)$ is defined for integers $n$ and $k$ with $n \geq 2k$ as the graph whose vertices are all the $k$-subsets of $[n]=\{1,2,\ldots,n\}$ where two such sets are adjacent if they are disjoint. The Schrijver graph $S(n,k)$ is defined as the subgraph of $K(n,k)$ induced by the collection of all $k$-subsets of $[n]$ that do not include two consecutive elements modulo $n$. It is known that the chromatic number of both $K(n,k)$ and $S(n,k)$ is $n-2k+2$. In the computational Kneser and Schrijver problems, we are given an access to a coloring with $n-2k+1$ colors of the vertices of $K(n,k)$ and $S(n,k)$ respectively, and the goal is to find a monochromatic edge. We prove that the problems admit randomized algorithms with running time $n^{O(1)} \cdot k^{O(k)}$, hence they are fixed-parameter tractable with respect to the parameter $k$. The analysis involves structural results on intersecting families and on induced subgraphs of Kneser and Schrijver graphs. We also study the Agreeable-Set problem of assigning a small subset of a set of $m$ items to a group of $\ell$ agents, so that all agents value the subset at least as much as its complement. As an application of our algorithm for the Kneser problem, we obtain a randomized polynomial-time algorithm for the Agreeable-Set problem for instances with $\ell \geq m - O(\frac{\log m}{\log \log m})$. We further show that the Agreeable-Set problem is at least as hard as a variant of the Kneser problem with an extended access to the input coloring.
[ { "created": "Tue, 19 Apr 2022 17:09:01 GMT", "version": "v1" }, { "created": "Sun, 1 May 2022 06:49:00 GMT", "version": "v2" }, { "created": "Tue, 13 Feb 2024 08:11:56 GMT", "version": "v3" } ]
2024-02-14
[ [ "Haviv", "Ishay", "" ] ]
The Kneser graph $K(n,k)$ is defined for integers $n$ and $k$ with $n \geq 2k$ as the graph whose vertices are all the $k$-subsets of $[n]=\{1,2,\ldots,n\}$ where two such sets are adjacent if they are disjoint. The Schrijver graph $S(n,k)$ is defined as the subgraph of $K(n,k)$ induced by the collection of all $k$-subsets of $[n]$ that do not include two consecutive elements modulo $n$. It is known that the chromatic number of both $K(n,k)$ and $S(n,k)$ is $n-2k+2$. In the computational Kneser and Schrijver problems, we are given an access to a coloring with $n-2k+1$ colors of the vertices of $K(n,k)$ and $S(n,k)$ respectively, and the goal is to find a monochromatic edge. We prove that the problems admit randomized algorithms with running time $n^{O(1)} \cdot k^{O(k)}$, hence they are fixed-parameter tractable with respect to the parameter $k$. The analysis involves structural results on intersecting families and on induced subgraphs of Kneser and Schrijver graphs. We also study the Agreeable-Set problem of assigning a small subset of a set of $m$ items to a group of $\ell$ agents, so that all agents value the subset at least as much as its complement. As an application of our algorithm for the Kneser problem, we obtain a randomized polynomial-time algorithm for the Agreeable-Set problem for instances with $\ell \geq m - O(\frac{\log m}{\log \log m})$. We further show that the Agreeable-Set problem is at least as hard as a variant of the Kneser problem with an extended access to the input coloring.
2402.09872
Arman Isajanyan
Arman Isajanyan, Artur Shatveryan, David Kocharyan, Zhangyang Wang, Humphrey Shi
Social Reward: Evaluating and Enhancing Generative AI through Million-User Feedback from an Online Creative Community
16 pages with 10 figures, accepted at ICLR 2024 as a spotlight, codes can be accessed at https://github.com/Picsart-AI-Research/Social-Reward
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Social reward as a form of community recognition provides a strong source of motivation for users of online platforms to engage and contribute with content. The recent progress of text-conditioned image synthesis has ushered in a collaborative era where AI empowers users to craft original visual artworks seeking community validation. Nevertheless, assessing these models in the context of collective community preference introduces distinct challenges. Existing evaluation methods predominantly center on limited size user studies guided by image quality and prompt alignment. This work pioneers a paradigm shift, unveiling Social Reward - an innovative reward modeling framework that leverages implicit feedback from social network users engaged in creative editing of generated images. We embark on an extensive journey of dataset curation and refinement, drawing from Picsart: an online visual creation and editing platform, yielding a first million-user-scale dataset of implicit human preferences for user-generated visual art named Picsart Image-Social. Our analysis exposes the shortcomings of current metrics in modeling community creative preference of text-to-image models' outputs, compelling us to introduce a novel predictive model explicitly tailored to address these limitations. Rigorous quantitative experiments and user study show that our Social Reward model aligns better with social popularity than existing metrics. Furthermore, we utilize Social Reward to fine-tune text-to-image models, yielding images that are more favored by not only Social Reward, but also other established metrics. These findings highlight the relevance and effectiveness of Social Reward in assessing community appreciation for AI-generated artworks, establishing a closer alignment with users' creative goals: creating popular visual art. Codes can be accessed at https://github.com/Picsart-AI-Research/Social-Reward
[ { "created": "Thu, 15 Feb 2024 10:56:31 GMT", "version": "v1" } ]
2024-02-16
[ [ "Isajanyan", "Arman", "" ], [ "Shatveryan", "Artur", "" ], [ "Kocharyan", "David", "" ], [ "Wang", "Zhangyang", "" ], [ "Shi", "Humphrey", "" ] ]
Social reward as a form of community recognition provides a strong source of motivation for users of online platforms to engage and contribute with content. The recent progress of text-conditioned image synthesis has ushered in a collaborative era where AI empowers users to craft original visual artworks seeking community validation. Nevertheless, assessing these models in the context of collective community preference introduces distinct challenges. Existing evaluation methods predominantly center on limited size user studies guided by image quality and prompt alignment. This work pioneers a paradigm shift, unveiling Social Reward - an innovative reward modeling framework that leverages implicit feedback from social network users engaged in creative editing of generated images. We embark on an extensive journey of dataset curation and refinement, drawing from Picsart: an online visual creation and editing platform, yielding a first million-user-scale dataset of implicit human preferences for user-generated visual art named Picsart Image-Social. Our analysis exposes the shortcomings of current metrics in modeling community creative preference of text-to-image models' outputs, compelling us to introduce a novel predictive model explicitly tailored to address these limitations. Rigorous quantitative experiments and user study show that our Social Reward model aligns better with social popularity than existing metrics. Furthermore, we utilize Social Reward to fine-tune text-to-image models, yielding images that are more favored by not only Social Reward, but also other established metrics. These findings highlight the relevance and effectiveness of Social Reward in assessing community appreciation for AI-generated artworks, establishing a closer alignment with users' creative goals: creating popular visual art. Codes can be accessed at https://github.com/Picsart-AI-Research/Social-Reward
2108.11903
Grischa Liebel
Rodi Jolak and Andreas Wortmann and Grischa Liebel and Eric Umuhoza and Michel R.V. Chaudron
Design Thinking and Creativity of Co-located vs. Globally Distributed Software Developers
This is a pre-peer-review version of an article published in Wiley Journal of Software: Evolution and Process. The final version is available via https://dx.doi.org/10.1002/smr.2377
null
10.1002/smr.2377
null
cs.SE
http://creativecommons.org/licenses/by/4.0/
Context: Designing software is an activity in which software developers think and make design decisions that shape the structure and behavior of software products. Designing software is one of the least understood software engineering activities. In a collaborative design setting, various types of distances can lead to challenges and effects that potentially affect how software is designed. Objective: To contribute to a better understanding of collaborative software design, we investigate how geographic distance affects its design thinking and the creativity of its discussions. Method: To this end, we conducted a multiple-case study exploring the design thinking and creativity of co-located and distributed software developers in a collaborative design setting. Results: Compared to co-located developers, distributed developers spend less time on exploring the problem space, which could be related to different socio-technical challenges, such as lack of awareness and common understanding. Distributed development does not seem to affect the creativity of their activities. Conclusion: Developers engaging in collaborative design need to be aware that problem space exploration is reduced in a distributed setting. Unless distributed teams take compensatory measures, this could adversely affect the development. Regarding the effect distance has on creativity, our results are inconclusive and further studies are needed.
[ { "created": "Thu, 26 Aug 2021 16:50:31 GMT", "version": "v1" } ]
2021-08-27
[ [ "Jolak", "Rodi", "" ], [ "Wortmann", "Andreas", "" ], [ "Liebel", "Grischa", "" ], [ "Umuhoza", "Eric", "" ], [ "Chaudron", "Michel R. V.", "" ] ]
Context: Designing software is an activity in which software developers think and make design decisions that shape the structure and behavior of software products. Designing software is one of the least understood software engineering activities. In a collaborative design setting, various types of distances can lead to challenges and effects that potentially affect how software is designed. Objective: To contribute to a better understanding of collaborative software design, we investigate how geographic distance affects its design thinking and the creativity of its discussions. Method: To this end, we conducted a multiple-case study exploring the design thinking and creativity of co-located and distributed software developers in a collaborative design setting. Results: Compared to co-located developers, distributed developers spend less time on exploring the problem space, which could be related to different socio-technical challenges, such as lack of awareness and common understanding. Distributed development does not seem to affect the creativity of their activities. Conclusion: Developers engaging in collaborative design need to be aware that problem space exploration is reduced in a distributed setting. Unless distributed teams take compensatory measures, this could adversely affect the development. Regarding the effect distance has on creativity, our results are inconclusive and further studies are needed.
2401.13150
Onur Cankur
Onur Cankur, Aditya Tomar, Daniel Nichols, Connor Scully-Allison, Katherine E. Isaacs, Abhinav Bhatele
Automated Programmatic Performance Analysis of Parallel Programs
null
null
null
null
cs.DC cs.PF
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Developing efficient parallel applications is critical to advancing scientific development but requires significant performance analysis and optimization. Performance analysis tools help developers manage the increasing complexity and scale of performance data, but often rely on the user to manually explore low-level data and are rigid in how the data can be manipulated. We propose a Python-based API, Chopper, which provides high-level and flexible performance analysis for both single and multiple executions of parallel applications. Chopper facilitates performance analysis and reduces developer effort by providing configurable high-level methods for common performance analysis tasks such as calculating load imbalance, hot paths, scalability bottlenecks, correlation between metrics and CCT nodes, and causes of performance variability within a robust and mature Python environment that provides fluid access to lower-level data manipulations. We demonstrate how Chopper allows developers to quickly and succinctly explore performance and identify issues across applications such as AMG, Laghos, LULESH, Quicksilver and Tortuga.
[ { "created": "Tue, 23 Jan 2024 23:52:48 GMT", "version": "v1" } ]
2024-01-25
[ [ "Cankur", "Onur", "" ], [ "Tomar", "Aditya", "" ], [ "Nichols", "Daniel", "" ], [ "Scully-Allison", "Connor", "" ], [ "Isaacs", "Katherine E.", "" ], [ "Bhatele", "Abhinav", "" ] ]
Developing efficient parallel applications is critical to advancing scientific development but requires significant performance analysis and optimization. Performance analysis tools help developers manage the increasing complexity and scale of performance data, but often rely on the user to manually explore low-level data and are rigid in how the data can be manipulated. We propose a Python-based API, Chopper, which provides high-level and flexible performance analysis for both single and multiple executions of parallel applications. Chopper facilitates performance analysis and reduces developer effort by providing configurable high-level methods for common performance analysis tasks such as calculating load imbalance, hot paths, scalability bottlenecks, correlation between metrics and CCT nodes, and causes of performance variability within a robust and mature Python environment that provides fluid access to lower-level data manipulations. We demonstrate how Chopper allows developers to quickly and succinctly explore performance and identify issues across applications such as AMG, Laghos, LULESH, Quicksilver and Tortuga.
2207.12909
Zerui Chen
Zerui Chen, Yana Hasson, Cordelia Schmid, Ivan Laptev
AlignSDF: Pose-Aligned Signed Distance Fields for Hand-Object Reconstruction
Accepted by ECCV 2022. Project Page: https://zerchen.github.io/projects/alignsdf.html
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Recent work achieved impressive progress towards joint reconstruction of hands and manipulated objects from monocular color images. Existing methods focus on two alternative representations in terms of either parametric meshes or signed distance fields (SDFs). On one side, parametric models can benefit from prior knowledge at the cost of limited shape deformations and mesh resolutions. Mesh models, hence, may fail to precisely reconstruct details such as contact surfaces of hands and objects. SDF-based methods, on the other side, can represent arbitrary details but are lacking explicit priors. In this work we aim to improve SDF models using priors provided by parametric representations. In particular, we propose a joint learning framework that disentangles the pose and the shape. We obtain hand and object poses from parametric models and use them to align SDFs in 3D space. We show that such aligned SDFs better focus on reconstructing shape details and improve reconstruction accuracy both for hands and objects. We evaluate our method and demonstrate significant improvements over the state of the art on the challenging ObMan and DexYCB benchmarks.
[ { "created": "Tue, 26 Jul 2022 13:58:59 GMT", "version": "v1" } ]
2022-07-27
[ [ "Chen", "Zerui", "" ], [ "Hasson", "Yana", "" ], [ "Schmid", "Cordelia", "" ], [ "Laptev", "Ivan", "" ] ]
Recent work achieved impressive progress towards joint reconstruction of hands and manipulated objects from monocular color images. Existing methods focus on two alternative representations in terms of either parametric meshes or signed distance fields (SDFs). On one side, parametric models can benefit from prior knowledge at the cost of limited shape deformations and mesh resolutions. Mesh models, hence, may fail to precisely reconstruct details such as contact surfaces of hands and objects. SDF-based methods, on the other side, can represent arbitrary details but are lacking explicit priors. In this work we aim to improve SDF models using priors provided by parametric representations. In particular, we propose a joint learning framework that disentangles the pose and the shape. We obtain hand and object poses from parametric models and use them to align SDFs in 3D space. We show that such aligned SDFs better focus on reconstructing shape details and improve reconstruction accuracy both for hands and objects. We evaluate our method and demonstrate significant improvements over the state of the art on the challenging ObMan and DexYCB benchmarks.
2403.07376
Bingqian Lin
Bingqian Lin, Yunshuang Nie, Ziming Wei, Jiaqi Chen, Shikui Ma, Jianhua Han, Hang Xu, Xiaojun Chang, Xiaodan Liang
NavCoT: Boosting LLM-Based Vision-and-Language Navigation via Learning Disentangled Reasoning
null
null
null
null
cs.CV cs.AI cs.CL cs.RO
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Vision-and-Language Navigation (VLN), as a crucial research problem of Embodied AI, requires an embodied agent to navigate through complex 3D environments following natural language instructions. Recent research has highlighted the promising capacity of large language models (LLMs) in VLN by improving navigational reasoning accuracy and interpretability. However, their predominant use in an offline manner usually suffers from substantial domain gap between the VLN task and the LLM training corpus. This paper introduces a novel strategy called Navigational Chain-of-Thought (NavCoT), where we fulfill parameter-efficient in-domain training to enable self-guided navigational decision, leading to a significant mitigation of the domain gap in a cost-effective manner. Specifically, at each timestep, the LLM is prompted to forecast the navigational chain-of-thought by: 1) acting as a world model to imagine the next observation according to the instruction, 2) selecting the candidate observation that best aligns with the imagination, and 3) determining the action based on the reasoning from the prior steps. Through constructing formalized labels for training, the LLM can learn to generate desired and reasonable chain-of-thought outputs for improving the action decision. Experimental results across various training settings and popular VLN benchmarks (e.g., Room-to-Room (R2R), Room-across-Room (RxR), Room-for-Room (R4R)) show the significant superiority of NavCoT over the direct action prediction variants. Through simple parameter-efficient finetuning, our NavCoT outperforms a recent GPT4-based approach with ~7% relative improvement on the R2R dataset. We believe that NavCoT will help unlock more task-adaptive and scalable LLM-based embodied agents, which are helpful for developing real-world robotics applications. Code is available at https://github.com/expectorlin/NavCoT.
[ { "created": "Tue, 12 Mar 2024 07:27:02 GMT", "version": "v1" } ]
2024-03-13
[ [ "Lin", "Bingqian", "" ], [ "Nie", "Yunshuang", "" ], [ "Wei", "Ziming", "" ], [ "Chen", "Jiaqi", "" ], [ "Ma", "Shikui", "" ], [ "Han", "Jianhua", "" ], [ "Xu", "Hang", "" ], [ "Chang", "Xiaojun", "" ], [ "Liang", "Xiaodan", "" ] ]
Vision-and-Language Navigation (VLN), as a crucial research problem of Embodied AI, requires an embodied agent to navigate through complex 3D environments following natural language instructions. Recent research has highlighted the promising capacity of large language models (LLMs) in VLN by improving navigational reasoning accuracy and interpretability. However, their predominant use in an offline manner usually suffers from substantial domain gap between the VLN task and the LLM training corpus. This paper introduces a novel strategy called Navigational Chain-of-Thought (NavCoT), where we fulfill parameter-efficient in-domain training to enable self-guided navigational decision, leading to a significant mitigation of the domain gap in a cost-effective manner. Specifically, at each timestep, the LLM is prompted to forecast the navigational chain-of-thought by: 1) acting as a world model to imagine the next observation according to the instruction, 2) selecting the candidate observation that best aligns with the imagination, and 3) determining the action based on the reasoning from the prior steps. Through constructing formalized labels for training, the LLM can learn to generate desired and reasonable chain-of-thought outputs for improving the action decision. Experimental results across various training settings and popular VLN benchmarks (e.g., Room-to-Room (R2R), Room-across-Room (RxR), Room-for-Room (R4R)) show the significant superiority of NavCoT over the direct action prediction variants. Through simple parameter-efficient finetuning, our NavCoT outperforms a recent GPT4-based approach with ~7% relative improvement on the R2R dataset. We believe that NavCoT will help unlock more task-adaptive and scalable LLM-based embodied agents, which are helpful for developing real-world robotics applications. Code is available at https://github.com/expectorlin/NavCoT.
2309.08638
Rajan Vivek
Rajan Vivek, Kawin Ethayarajh, Diyi Yang, Douwe Kiela
Anchor Points: Benchmarking Models with Much Fewer Examples
Accepted to EACL 2024 Main Conference. Code will be released at: https://github.com/rvivek3/AnchorPoints
null
null
null
cs.CL
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Modern language models often exhibit powerful but brittle behavior, leading to the development of larger and more diverse benchmarks to reliably assess their behavior. Here, we suggest that model performance can be benchmarked and elucidated with much smaller evaluation sets. We first show that in six popular language classification benchmarks, model confidence in the correct class on many pairs of points is strongly correlated across models. We build upon this phenomenon to propose Anchor Point Selection, a technique to select small subsets of datasets that capture model behavior across the entire dataset. Anchor points reliably rank models: across 87 diverse language model-prompt pairs, evaluating models using 1-30 anchor points outperforms uniform sampling and other baselines at accurately ranking models. Moreover, just several anchor points can be used to estimate model per-class predictions on all other points in a dataset with low mean absolute error, sufficient for gauging where the model is likely to fail. Lastly, we present Anchor Point Maps for visualizing these insights and facilitating comparisons of the performance of different models on various regions within the dataset distribution.
[ { "created": "Thu, 14 Sep 2023 17:45:51 GMT", "version": "v1" }, { "created": "Sun, 18 Feb 2024 21:37:47 GMT", "version": "v2" } ]
2024-02-20
[ [ "Vivek", "Rajan", "" ], [ "Ethayarajh", "Kawin", "" ], [ "Yang", "Diyi", "" ], [ "Kiela", "Douwe", "" ] ]
Modern language models often exhibit powerful but brittle behavior, leading to the development of larger and more diverse benchmarks to reliably assess their behavior. Here, we suggest that model performance can be benchmarked and elucidated with much smaller evaluation sets. We first show that in six popular language classification benchmarks, model confidence in the correct class on many pairs of points is strongly correlated across models. We build upon this phenomenon to propose Anchor Point Selection, a technique to select small subsets of datasets that capture model behavior across the entire dataset. Anchor points reliably rank models: across 87 diverse language model-prompt pairs, evaluating models using 1-30 anchor points outperforms uniform sampling and other baselines at accurately ranking models. Moreover, just several anchor points can be used to estimate model per-class predictions on all other points in a dataset with low mean absolute error, sufficient for gauging where the model is likely to fail. Lastly, we present Anchor Point Maps for visualizing these insights and facilitating comparisons of the performance of different models on various regions within the dataset distribution.
2112.13974
Akansha Singh Bansal
Akansha Singh Bansal, Trapit Bansal, David Irwin
A Moment in the Sun: Solar Nowcasting from Multispectral Satellite Data using Self-Supervised Learning
18 pages
null
null
null
cs.LG cs.AI cs.CV
http://creativecommons.org/licenses/by/4.0/
Solar energy is now the cheapest form of electricity in history. Unfortunately, significantly increasing the grid's fraction of solar energy remains challenging due to its variability, which makes balancing electricity's supply and demand more difficult. While thermal generators' ramp rate -- the maximum rate that they can change their output -- is finite, solar's ramp rate is essentially infinite. Thus, accurate near-term solar forecasting, or nowcasting, is important to provide advance warning to adjust thermal generator output in response to solar variations to ensure a balanced supply and demand. To address the problem, this paper develops a general model for solar nowcasting from abundant and readily available multispectral satellite data using self-supervised learning. Specifically, we develop deep auto-regressive models using convolutional neural networks (CNN) and long short-term memory networks (LSTM) that are globally trained across multiple locations to predict raw future observations of the spatio-temporal data collected by the recently launched GOES-R series of satellites. Our model estimates a location's future solar irradiance based on satellite observations, which we feed to a regression model trained on smaller site-specific solar data to provide near-term solar photovoltaic (PV) forecasts that account for site-specific characteristics. We evaluate our approach for different coverage areas and forecast horizons across 25 solar sites and show that our approach yields errors close to that of a model using ground-truth observations.
[ { "created": "Tue, 28 Dec 2021 03:13:44 GMT", "version": "v1" } ]
2021-12-30
[ [ "Bansal", "Akansha Singh", "" ], [ "Bansal", "Trapit", "" ], [ "Irwin", "David", "" ] ]
Solar energy is now the cheapest form of electricity in history. Unfortunately, significantly increasing the grid's fraction of solar energy remains challenging due to its variability, which makes balancing electricity's supply and demand more difficult. While thermal generators' ramp rate -- the maximum rate that they can change their output -- is finite, solar's ramp rate is essentially infinite. Thus, accurate near-term solar forecasting, or nowcasting, is important to provide advance warning to adjust thermal generator output in response to solar variations to ensure a balanced supply and demand. To address the problem, this paper develops a general model for solar nowcasting from abundant and readily available multispectral satellite data using self-supervised learning. Specifically, we develop deep auto-regressive models using convolutional neural networks (CNN) and long short-term memory networks (LSTM) that are globally trained across multiple locations to predict raw future observations of the spatio-temporal data collected by the recently launched GOES-R series of satellites. Our model estimates a location's future solar irradiance based on satellite observations, which we feed to a regression model trained on smaller site-specific solar data to provide near-term solar photovoltaic (PV) forecasts that account for site-specific characteristics. We evaluate our approach for different coverage areas and forecast horizons across 25 solar sites and show that our approach yields errors close to that of a model using ground-truth observations.
2101.08918
Tianming Feng
Tianming Feng, Shuo Shi, Shushi Gu, Ning Zhang, Wei Xiang, and Xuemai Gu
Performance Analysis for Cache-enabled Cellular Networks with Cooperative Transmission
arXiv admin note: text overlap with arXiv:2101.08669
null
null
null
cs.IT eess.SP math.IT
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
The large amount of deployed smart devices put tremendous traffic pressure on networks. Caching at the edge has been widely studied as a promising technique to solve this problem. To further improve the successful transmission probability (STP) of cache-enabled cellular networks (CEN), we combine the cooperative transmission technique with CEN and propose a novel transmission scheme. Local channel state information (CSI) is introduced at each cooperative base station (BS) to enhance the strength of the signal received by the user. A tight approximation for the STP of this scheme is derived using tools from stochastic geometry. The optimal content placement strategy of this scheme is obtained using a numerical method to maximize the STP. Simulation results demonstrate the optimal strategy achieves significant gains in STP over several comparative baselines with the proposed scheme.
[ { "created": "Fri, 22 Jan 2021 02:00:06 GMT", "version": "v1" } ]
2021-01-25
[ [ "Feng", "Tianming", "" ], [ "Shi", "Shuo", "" ], [ "Gu", "Shushi", "" ], [ "Zhang", "Ning", "" ], [ "Xiang", "Wei", "" ], [ "Gu", "Xuemai", "" ] ]
The large amount of deployed smart devices put tremendous traffic pressure on networks. Caching at the edge has been widely studied as a promising technique to solve this problem. To further improve the successful transmission probability (STP) of cache-enabled cellular networks (CEN), we combine the cooperative transmission technique with CEN and propose a novel transmission scheme. Local channel state information (CSI) is introduced at each cooperative base station (BS) to enhance the strength of the signal received by the user. A tight approximation for the STP of this scheme is derived using tools from stochastic geometry. The optimal content placement strategy of this scheme is obtained using a numerical method to maximize the STP. Simulation results demonstrate the optimal strategy achieves significant gains in STP over several comparative baselines with the proposed scheme.
2203.03704
Jeff Delaune
Jeff Delaune, Jacob Izraelevitz, Samuel Sirlin, David Sternberg, Louis Giersch, L. Phillipe Tosi, Evgeniy Skliyanskiy, Larry Young, Michael Mischna, Shannah Withrow-Maser, Juergen Mueller, Joshua Bowman, Mark S Wallace, Havard F. Grip, Larry Matthies, Wayne Johnson, Matthew Keennon, Benjamin Pipenberg, Harsh Patel, Christopher Lim, Aaron Schutte, Marcel Veismann, Haley Cummings, Sarah Conley, Jonathan Bapst, Theodore Tzanetos, Roland Brockers, Abhinandan Jain, David Bayard, Art Chmielewski, Olivier Toupet, Joel Burdick, Morteza Gharib and J. (Bob) Balaram
Mid-Air Helicopter Delivery at Mars Using a Jetpack
Accepted in 2022 IEEE Aerospace Conference
null
null
null
cs.RO
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Mid-Air Helicopter Delivery (MAHD) is a new Entry, Descent and Landing (EDL) architecture to enable in situ mobility for Mars science at lower cost than previous missions. It uses a jetpack to slow down a Mars Science Helicopter (MSH) after separation from the backshell, and reach aerodynamic conditions suitable for helicopter take-off in mid air. For given aeroshell dimensions, only MAHD's lander-free approach leaves enough room in the aeroshell to accommodate the largest rotor option for MSH. This drastically improves flight performance, notably allowing +150\% increased science payload mass. Compared to heritage EDL approaches, the simpler MAHD architecture is also likely to reduce cost, and enables access to more hazardous and higher-elevation terrains on Mars. This paper introduces a design for the MAHD system architecture and operations. We present a mechanical configuration that fits both MSH and the jetpack within the 2.65-m Mars heritage aeroshell, and a jetpack control architecture which fully leverages the available helicopter avionics. We discuss preliminary numerical models of the flow dynamics resulting from the interaction between the jets, the rotors and the side winds. We define a force-torque sensing architecture capable of handling the wind and trimming the rotors to prepare for safe take-off. Finally, we analyze the dynamic environment and closed-loop control simulation results to demonstrate the preliminary feasibility of MAHD.
[ { "created": "Mon, 7 Mar 2022 21:07:56 GMT", "version": "v1" } ]
2022-03-09
[ [ "Delaune", "Jeff", "", "Bob" ], [ "Izraelevitz", "Jacob", "", "Bob" ], [ "Sirlin", "Samuel", "", "Bob" ], [ "Sternberg", "David", "", "Bob" ], [ "Giersch", "Louis", "", "Bob" ], [ "Tosi", "L. Phillipe", "", "Bob" ], [ "Skliyanskiy", "Evgeniy", "", "Bob" ], [ "Young", "Larry", "", "Bob" ], [ "Mischna", "Michael", "", "Bob" ], [ "Withrow-Maser", "Shannah", "", "Bob" ], [ "Mueller", "Juergen", "", "Bob" ], [ "Bowman", "Joshua", "", "Bob" ], [ "Wallace", "Mark S", "", "Bob" ], [ "Grip", "Havard F.", "", "Bob" ], [ "Matthies", "Larry", "", "Bob" ], [ "Johnson", "Wayne", "", "Bob" ], [ "Keennon", "Matthew", "", "Bob" ], [ "Pipenberg", "Benjamin", "", "Bob" ], [ "Patel", "Harsh", "", "Bob" ], [ "Lim", "Christopher", "", "Bob" ], [ "Schutte", "Aaron", "", "Bob" ], [ "Veismann", "Marcel", "", "Bob" ], [ "Cummings", "Haley", "", "Bob" ], [ "Conley", "Sarah", "", "Bob" ], [ "Bapst", "Jonathan", "", "Bob" ], [ "Tzanetos", "Theodore", "", "Bob" ], [ "Brockers", "Roland", "", "Bob" ], [ "Jain", "Abhinandan", "", "Bob" ], [ "Bayard", "David", "", "Bob" ], [ "Chmielewski", "Art", "", "Bob" ], [ "Toupet", "Olivier", "", "Bob" ], [ "Burdick", "Joel", "", "Bob" ], [ "Gharib", "Morteza", "", "Bob" ], [ "J.", "", "", "Bob" ], [ "Balaram", "", "" ] ]
Mid-Air Helicopter Delivery (MAHD) is a new Entry, Descent and Landing (EDL) architecture to enable in situ mobility for Mars science at lower cost than previous missions. It uses a jetpack to slow down a Mars Science Helicopter (MSH) after separation from the backshell, and reach aerodynamic conditions suitable for helicopter take-off in mid air. For given aeroshell dimensions, only MAHD's lander-free approach leaves enough room in the aeroshell to accommodate the largest rotor option for MSH. This drastically improves flight performance, notably allowing +150\% increased science payload mass. Compared to heritage EDL approaches, the simpler MAHD architecture is also likely to reduce cost, and enables access to more hazardous and higher-elevation terrains on Mars. This paper introduces a design for the MAHD system architecture and operations. We present a mechanical configuration that fits both MSH and the jetpack within the 2.65-m Mars heritage aeroshell, and a jetpack control architecture which fully leverages the available helicopter avionics. We discuss preliminary numerical models of the flow dynamics resulting from the interaction between the jets, the rotors and the side winds. We define a force-torque sensing architecture capable of handling the wind and trimming the rotors to prepare for safe take-off. Finally, we analyze the dynamic environment and closed-loop control simulation results to demonstrate the preliminary feasibility of MAHD.
2211.11344
Jakub Tetek
Shyam Narayanan, Jakub T\v{e}tek
Estimating the Effective Support Size in Constant Query Complexity
null
null
null
null
cs.DS math.ST stat.TH
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Estimating the support size of a distribution is a well-studied problem in statistics. Motivated by the fact that this problem is highly non-robust (as small perturbations in the distributions can drastically affect the support size) and thus hard to estimate, Goldreich [ECCC 2019] studied the query complexity of estimating the $\epsilon$-\emph{effective support size} $\text{Ess}_\epsilon$ of a distribution ${P}$, which is equal to the smallest support size of a distribution that is $\epsilon$-far in total variation distance from ${P}$. In his paper, he shows an algorithm in the dual access setting (where we may both receive random samples and query the sampling probability $p(x)$ for any $x$) for a bicriteria approximation, giving an answer in $[\text{Ess}_{(1+\beta)\epsilon},(1+\gamma) \text{Ess}_{\epsilon}]$ for some values $\beta, \gamma > 0$. However, his algorithm has either super-constant query complexity in the support size or super-constant approximation ratio $1+\gamma = \omega(1)$. He then asked if this is necessary, or if it is possible to get a constant-factor approximation in a number of queries independent of the support size. We answer his question by showing that not only is complexity independent of $n$ possible for $\gamma>0$, but also for $\gamma=0$, that is, that the bicriteria relaxation is not necessary. Specifically, we show an algorithm with query complexity $O(\frac{1}{\beta^3 \epsilon^3})$. That is, for any $0 < \epsilon, \beta < 1$, we output in this complexity a number $\tilde{n} \in [\text{Ess}_{(1+\beta)\epsilon},\text{Ess}_\epsilon]$. We also show that it is possible to solve the approximate version with approximation ratio $1+\gamma$ in complexity $O\left(\frac{1}{\beta^2 \epsilon} + \frac{1}{\beta \epsilon \gamma^2}\right)$. Our algorithm is very simple, and has $4$ short lines of pseudocode.
[ { "created": "Mon, 21 Nov 2022 10:49:32 GMT", "version": "v1" } ]
2022-11-22
[ [ "Narayanan", "Shyam", "" ], [ "Tětek", "Jakub", "" ] ]
Estimating the support size of a distribution is a well-studied problem in statistics. Motivated by the fact that this problem is highly non-robust (as small perturbations in the distributions can drastically affect the support size) and thus hard to estimate, Goldreich [ECCC 2019] studied the query complexity of estimating the $\epsilon$-\emph{effective support size} $\text{Ess}_\epsilon$ of a distribution ${P}$, which is equal to the smallest support size of a distribution that is $\epsilon$-far in total variation distance from ${P}$. In his paper, he shows an algorithm in the dual access setting (where we may both receive random samples and query the sampling probability $p(x)$ for any $x$) for a bicriteria approximation, giving an answer in $[\text{Ess}_{(1+\beta)\epsilon},(1+\gamma) \text{Ess}_{\epsilon}]$ for some values $\beta, \gamma > 0$. However, his algorithm has either super-constant query complexity in the support size or super-constant approximation ratio $1+\gamma = \omega(1)$. He then asked if this is necessary, or if it is possible to get a constant-factor approximation in a number of queries independent of the support size. We answer his question by showing that not only is complexity independent of $n$ possible for $\gamma>0$, but also for $\gamma=0$, that is, that the bicriteria relaxation is not necessary. Specifically, we show an algorithm with query complexity $O(\frac{1}{\beta^3 \epsilon^3})$. That is, for any $0 < \epsilon, \beta < 1$, we output in this complexity a number $\tilde{n} \in [\text{Ess}_{(1+\beta)\epsilon},\text{Ess}_\epsilon]$. We also show that it is possible to solve the approximate version with approximation ratio $1+\gamma$ in complexity $O\left(\frac{1}{\beta^2 \epsilon} + \frac{1}{\beta \epsilon \gamma^2}\right)$. Our algorithm is very simple, and has $4$ short lines of pseudocode.
1901.10237
Hai Duong Nguyen
Hai-Duong Nguyen, Soo-Hyung Kim
Automatic Whole-body Bone Age Assessment Using Deep Hierarchical Features
null
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Bone age assessment gives us evidence to analyze the children growth status and the rejuvenation involved chronological and biological ages. All the previous works consider left-hand X-ray image of a child in their works. In this paper, we carry out a study on estimating human age using whole-body bone CT images and a novel convolutional neural network. Our model with additional connections shows an effective way to generate a massive number of vital features while reducing overfitting influence on small training data in the medical image analysis research area. A dataset and a comparison with common deep architectures will be provided for future research in this field.
[ { "created": "Tue, 29 Jan 2019 11:53:30 GMT", "version": "v1" } ]
2019-01-30
[ [ "Nguyen", "Hai-Duong", "" ], [ "Kim", "Soo-Hyung", "" ] ]
Bone age assessment gives us evidence to analyze the children growth status and the rejuvenation involved chronological and biological ages. All the previous works consider left-hand X-ray image of a child in their works. In this paper, we carry out a study on estimating human age using whole-body bone CT images and a novel convolutional neural network. Our model with additional connections shows an effective way to generate a massive number of vital features while reducing overfitting influence on small training data in the medical image analysis research area. A dataset and a comparison with common deep architectures will be provided for future research in this field.
2405.10098
Sen Huang
Sen Huang, Kaixiang Yang, Sheng Qi, Rui Wang
When Large Language Model Meets Optimization
null
null
null
null
cs.NE
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Optimization algorithms and large language models (LLMs) enhance decision-making in dynamic environments by integrating artificial intelligence with traditional techniques. LLMs, with extensive domain knowledge, facilitate intelligent modeling and strategic decision-making in optimization, while optimization algorithms refine LLM architectures and output quality. This synergy offers novel approaches for advancing general AI, addressing both the computational challenges of complex problems and the application of LLMs in practical scenarios. This review outlines the progress and potential of combining LLMs with optimization algorithms, providing insights for future research directions.
[ { "created": "Thu, 16 May 2024 13:54:37 GMT", "version": "v1" } ]
2024-05-17
[ [ "Huang", "Sen", "" ], [ "Yang", "Kaixiang", "" ], [ "Qi", "Sheng", "" ], [ "Wang", "Rui", "" ] ]
Optimization algorithms and large language models (LLMs) enhance decision-making in dynamic environments by integrating artificial intelligence with traditional techniques. LLMs, with extensive domain knowledge, facilitate intelligent modeling and strategic decision-making in optimization, while optimization algorithms refine LLM architectures and output quality. This synergy offers novel approaches for advancing general AI, addressing both the computational challenges of complex problems and the application of LLMs in practical scenarios. This review outlines the progress and potential of combining LLMs with optimization algorithms, providing insights for future research directions.
1504.01802
Morteza Hashemi
Morteza Hashemi, Yuval Cassuto, Ari Trachtenberg
Fountain Codes with Nonuniform Selection Distributions through Feedback
Submitted to the IEEE Transactions on Information Theory
null
10.1109/TIT.2016.2570232
null
cs.IT math.IT
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
One key requirement for fountain (rateless) coding schemes is to achieve a high intermediate symbol recovery rate. Recent coding schemes have incorporated the use of a feedback channel to improve intermediate performance of traditional rateless codes; however, these codes with feedback are designed based on uniformly at random selection of input symbols. In this paper, on the other hand, we develop feedback-based fountain codes with dynamically-adjusted nonuniform symbol selection distributions, and show that this characteristic can enhance the intermediate decoding rate. We provide an analysis of our codes, including bounds on computational complexity and failure probability for a maximum likelihood decoder; the latter are tighter than bounds known for classical rateless codes. Through numerical simulations, we also show that feedback information paired with a nonuniform selection distribution can highly improve the symbol recovery rate, and that the amount of feedback sent can be tuned to the specific transmission properties of a given feedback channel.
[ { "created": "Wed, 8 Apr 2015 02:11:09 GMT", "version": "v1" } ]
2016-11-17
[ [ "Hashemi", "Morteza", "" ], [ "Cassuto", "Yuval", "" ], [ "Trachtenberg", "Ari", "" ] ]
One key requirement for fountain (rateless) coding schemes is to achieve a high intermediate symbol recovery rate. Recent coding schemes have incorporated the use of a feedback channel to improve intermediate performance of traditional rateless codes; however, these codes with feedback are designed based on uniformly at random selection of input symbols. In this paper, on the other hand, we develop feedback-based fountain codes with dynamically-adjusted nonuniform symbol selection distributions, and show that this characteristic can enhance the intermediate decoding rate. We provide an analysis of our codes, including bounds on computational complexity and failure probability for a maximum likelihood decoder; the latter are tighter than bounds known for classical rateless codes. Through numerical simulations, we also show that feedback information paired with a nonuniform selection distribution can highly improve the symbol recovery rate, and that the amount of feedback sent can be tuned to the specific transmission properties of a given feedback channel.
1708.08844
Jan Czarnowski
Jan Czarnowski, Stefan Leutenegger, Andrew Davison
Semantic Texture for Robust Dense Tracking
null
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
We argue that robust dense SLAM systems can make valuable use of the layers of features coming from a standard CNN as a pyramid of `semantic texture' which is suitable for dense alignment while being much more robust to nuisance factors such as lighting than raw RGB values. We use a straightforward Lucas-Kanade formulation of image alignment, with a schedule of iterations over the coarse-to-fine levels of a pyramid, and simply replace the usual image pyramid by the hierarchy of convolutional feature maps from a pre-trained CNN. The resulting dense alignment performance is much more robust to lighting and other variations, as we show by camera rotation tracking experiments on time-lapse sequences captured over many hours. Looking towards the future of scene representation for real-time visual SLAM, we further demonstrate that a selection using simple criteria of a small number of the total set of features output by a CNN gives just as accurate but much more efficient tracking performance.
[ { "created": "Tue, 29 Aug 2017 15:58:18 GMT", "version": "v1" } ]
2017-08-30
[ [ "Czarnowski", "Jan", "" ], [ "Leutenegger", "Stefan", "" ], [ "Davison", "Andrew", "" ] ]
We argue that robust dense SLAM systems can make valuable use of the layers of features coming from a standard CNN as a pyramid of `semantic texture' which is suitable for dense alignment while being much more robust to nuisance factors such as lighting than raw RGB values. We use a straightforward Lucas-Kanade formulation of image alignment, with a schedule of iterations over the coarse-to-fine levels of a pyramid, and simply replace the usual image pyramid by the hierarchy of convolutional feature maps from a pre-trained CNN. The resulting dense alignment performance is much more robust to lighting and other variations, as we show by camera rotation tracking experiments on time-lapse sequences captured over many hours. Looking towards the future of scene representation for real-time visual SLAM, we further demonstrate that a selection using simple criteria of a small number of the total set of features output by a CNN gives just as accurate but much more efficient tracking performance.
2312.10029
Zachary Kenton
Sebastian Farquhar, Vikrant Varma, Zachary Kenton, Johannes Gasteiger, Vladimir Mikulik, Rohin Shah
Challenges with unsupervised LLM knowledge discovery
12 pages (38 including references and appendices). First three authors equal contribution, randomised order
null
null
null
cs.LG cs.AI
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
We show that existing unsupervised methods on large language model (LLM) activations do not discover knowledge -- instead they seem to discover whatever feature of the activations is most prominent. The idea behind unsupervised knowledge elicitation is that knowledge satisfies a consistency structure, which can be used to discover knowledge. We first prove theoretically that arbitrary features (not just knowledge) satisfy the consistency structure of a particular leading unsupervised knowledge-elicitation method, contrast-consistent search (Burns et al. - arXiv:2212.03827). We then present a series of experiments showing settings in which unsupervised methods result in classifiers that do not predict knowledge, but instead predict a different prominent feature. We conclude that existing unsupervised methods for discovering latent knowledge are insufficient, and we contribute sanity checks to apply to evaluating future knowledge elicitation methods. Conceptually, we hypothesise that the identification issues explored here, e.g. distinguishing a model's knowledge from that of a simulated character's, will persist for future unsupervised methods.
[ { "created": "Fri, 15 Dec 2023 18:49:43 GMT", "version": "v1" }, { "created": "Mon, 18 Dec 2023 16:43:35 GMT", "version": "v2" } ]
2023-12-19
[ [ "Farquhar", "Sebastian", "" ], [ "Varma", "Vikrant", "" ], [ "Kenton", "Zachary", "" ], [ "Gasteiger", "Johannes", "" ], [ "Mikulik", "Vladimir", "" ], [ "Shah", "Rohin", "" ] ]
We show that existing unsupervised methods on large language model (LLM) activations do not discover knowledge -- instead they seem to discover whatever feature of the activations is most prominent. The idea behind unsupervised knowledge elicitation is that knowledge satisfies a consistency structure, which can be used to discover knowledge. We first prove theoretically that arbitrary features (not just knowledge) satisfy the consistency structure of a particular leading unsupervised knowledge-elicitation method, contrast-consistent search (Burns et al. - arXiv:2212.03827). We then present a series of experiments showing settings in which unsupervised methods result in classifiers that do not predict knowledge, but instead predict a different prominent feature. We conclude that existing unsupervised methods for discovering latent knowledge are insufficient, and we contribute sanity checks to apply to evaluating future knowledge elicitation methods. Conceptually, we hypothesise that the identification issues explored here, e.g. distinguishing a model's knowledge from that of a simulated character's, will persist for future unsupervised methods.
1604.07153
Kim-Manuel Klein
Klaus Jansen, Kim-Manuel Klein, Jos\'e Verschae
Closing the Gap for Makespan Scheduling via Sparsification Techniques
20 pages, ICALP 2016
null
null
null
cs.DS
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Makespan scheduling on identical machines is one of the most basic and fundamental packing problems studied in the discrete optimization literature. It asks for an assignment of $n$ jobs to a set of $m$ identical machines that minimizes the makespan. The problem is strongly NP-hard, and thus we do not expect a $(1+\epsilon)$-approximation algorithm with a running time that depends polynomially on $1/\epsilon$. Furthermore, Chen et al. [3] recently showed that a running time of $2^{(1/\epsilon)^{1-\delta}}+\text{poly}(n)$ for any $\delta>0$ would imply that the Exponential Time Hypothesis (ETH) fails. A long sequence of algorithms have been developed that try to obtain low dependencies on $1/\epsilon$, the better of which achieves a running time of $2^{\tilde{O}(1/\epsilon^2)}+O(n\log n)$ [11]. In this paper we obtain an algorithm with a running time of $2^{\tilde{O}(1/\epsilon)}+O(n\log n)$, which is tight under ETH up to logarithmic factors on the exponent. Our main technical contribution is a new structural result on the configuration-IP. More precisely, we show the existence of a highly symmetric and sparse optimal solution, in which all but a constant number of machines are assigned a configuration with small support. This structure can then be exploited by integer programming techniques and enumeration. We believe that our structural result is of independent interest and should find applications to other settings. In particular, we show how the structure can be applied to the minimum makespan problem on related machines and to a larger class of objective functions on parallel machines. For all these cases we obtain an efficient PTAS with running time $2^{\tilde{O}(1/\epsilon)} + \text{poly}(n)$.
[ { "created": "Mon, 25 Apr 2016 07:47:34 GMT", "version": "v1" } ]
2016-04-26
[ [ "Jansen", "Klaus", "" ], [ "Klein", "Kim-Manuel", "" ], [ "Verschae", "José", "" ] ]
Makespan scheduling on identical machines is one of the most basic and fundamental packing problems studied in the discrete optimization literature. It asks for an assignment of $n$ jobs to a set of $m$ identical machines that minimizes the makespan. The problem is strongly NP-hard, and thus we do not expect a $(1+\epsilon)$-approximation algorithm with a running time that depends polynomially on $1/\epsilon$. Furthermore, Chen et al. [3] recently showed that a running time of $2^{(1/\epsilon)^{1-\delta}}+\text{poly}(n)$ for any $\delta>0$ would imply that the Exponential Time Hypothesis (ETH) fails. A long sequence of algorithms have been developed that try to obtain low dependencies on $1/\epsilon$, the better of which achieves a running time of $2^{\tilde{O}(1/\epsilon^2)}+O(n\log n)$ [11]. In this paper we obtain an algorithm with a running time of $2^{\tilde{O}(1/\epsilon)}+O(n\log n)$, which is tight under ETH up to logarithmic factors on the exponent. Our main technical contribution is a new structural result on the configuration-IP. More precisely, we show the existence of a highly symmetric and sparse optimal solution, in which all but a constant number of machines are assigned a configuration with small support. This structure can then be exploited by integer programming techniques and enumeration. We believe that our structural result is of independent interest and should find applications to other settings. In particular, we show how the structure can be applied to the minimum makespan problem on related machines and to a larger class of objective functions on parallel machines. For all these cases we obtain an efficient PTAS with running time $2^{\tilde{O}(1/\epsilon)} + \text{poly}(n)$.
1812.03953
Hadi Abdi Khojasteh
Hadi Abdi Khojasteh, Alireza Abbas Alipour, Ebrahim Ansari and Parvin Razzaghi
An Intelligent Safety System for Human-Centered Semi-Autonomous Vehicles
15 pages and 5 figures, Submitted to the international conference on Contemporary issues in Data Science (CiDaS 2019), Learn more about this project at https://iasbs.ac.ir/~ansari/faraz
Nature Switzerland AG - Springer LNDECT 45(2020) 322-336
10.1007/978-3-030-37309-2_26
null
cs.CV cs.LG cs.RO
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Nowadays, automobile manufacturers make efforts to develop ways to make cars fully safe. Monitoring driver's actions by computer vision techniques to detect driving mistakes in real-time and then planning for autonomous driving to avoid vehicle collisions is one of the most important issues that has been investigated in the machine vision and Intelligent Transportation Systems (ITS). The main goal of this study is to prevent accidents caused by fatigue, drowsiness, and driver distraction. To avoid these incidents, this paper proposes an integrated safety system that continuously monitors the driver's attention and vehicle surroundings, and finally decides whether the actual steering control status is safe or not. For this purpose, we equipped an ordinary car called FARAZ with a vision system consisting of four mounted cameras along with a universal car tool for communicating with surrounding factory-installed sensors and other car systems, and sending commands to actuators. The proposed system leverages a scene understanding pipeline using deep convolutional encoder-decoder networks and a driver state detection pipeline. We have been identifying and assessing domestic capabilities for the development of technologies specifically of the ordinary vehicles in order to manufacture smart cars and eke providing an intelligent system to increase safety and to assist the driver in various conditions/situations.
[ { "created": "Mon, 10 Dec 2018 18:08:18 GMT", "version": "v1" }, { "created": "Wed, 20 Feb 2019 22:57:02 GMT", "version": "v2" } ]
2020-02-28
[ [ "Khojasteh", "Hadi Abdi", "" ], [ "Alipour", "Alireza Abbas", "" ], [ "Ansari", "Ebrahim", "" ], [ "Razzaghi", "Parvin", "" ] ]
Nowadays, automobile manufacturers make efforts to develop ways to make cars fully safe. Monitoring driver's actions by computer vision techniques to detect driving mistakes in real-time and then planning for autonomous driving to avoid vehicle collisions is one of the most important issues that has been investigated in the machine vision and Intelligent Transportation Systems (ITS). The main goal of this study is to prevent accidents caused by fatigue, drowsiness, and driver distraction. To avoid these incidents, this paper proposes an integrated safety system that continuously monitors the driver's attention and vehicle surroundings, and finally decides whether the actual steering control status is safe or not. For this purpose, we equipped an ordinary car called FARAZ with a vision system consisting of four mounted cameras along with a universal car tool for communicating with surrounding factory-installed sensors and other car systems, and sending commands to actuators. The proposed system leverages a scene understanding pipeline using deep convolutional encoder-decoder networks and a driver state detection pipeline. We have been identifying and assessing domestic capabilities for the development of technologies specifically of the ordinary vehicles in order to manufacture smart cars and eke providing an intelligent system to increase safety and to assist the driver in various conditions/situations.
2011.10278
Junho Koh
Junho Koh, Jaekyum Kim, Younji Shin, Byeongwon Lee, Seungji Yang and Jun Won Choi
Joint Representation of Temporal Image Sequences and Object Motion for Video Object Detection
null
null
null
null
cs.CV
http://creativecommons.org/licenses/by/4.0/
In this paper, we propose a new video object detector (VoD) method referred to as temporal feature aggregation and motion-aware VoD (TM-VoD), which produces a joint representation of temporal image sequences and object motion. The proposed TM-VoD aggregates visual feature maps extracted by convolutional neural networks applying the temporal attention gating and spatial feature alignment. This temporal feature aggregation is performed in two stages in a hierarchical fashion. In the first stage, the visual feature maps are fused at a pixel level via gated attention model. In the second stage, the proposed method aggregates the features after aligning the object features using temporal box offset calibration and weights them according to the cosine similarity measure. The proposed TM-VoD also finds the representation of the motion of objects in two successive steps. The pixel-level motion features are first computed based on the incremental changes between the adjacent visual feature maps. Then, box-level motion features are obtained from both the region of interest (RoI)-aligned pixel-level motion features and the sequential changes of the box coordinates. Finally, all these features are concatenated to produce a joint representation of the objects for VoD. The experiments conducted on the ImageNet VID dataset demonstrate that the proposed method outperforms existing VoD methods and achieves a performance comparable to that of state-of-the-art VoDs.
[ { "created": "Fri, 20 Nov 2020 08:46:12 GMT", "version": "v1" } ]
2020-11-23
[ [ "Koh", "Junho", "" ], [ "Kim", "Jaekyum", "" ], [ "Shin", "Younji", "" ], [ "Lee", "Byeongwon", "" ], [ "Yang", "Seungji", "" ], [ "Choi", "Jun Won", "" ] ]
In this paper, we propose a new video object detector (VoD) method referred to as temporal feature aggregation and motion-aware VoD (TM-VoD), which produces a joint representation of temporal image sequences and object motion. The proposed TM-VoD aggregates visual feature maps extracted by convolutional neural networks applying the temporal attention gating and spatial feature alignment. This temporal feature aggregation is performed in two stages in a hierarchical fashion. In the first stage, the visual feature maps are fused at a pixel level via gated attention model. In the second stage, the proposed method aggregates the features after aligning the object features using temporal box offset calibration and weights them according to the cosine similarity measure. The proposed TM-VoD also finds the representation of the motion of objects in two successive steps. The pixel-level motion features are first computed based on the incremental changes between the adjacent visual feature maps. Then, box-level motion features are obtained from both the region of interest (RoI)-aligned pixel-level motion features and the sequential changes of the box coordinates. Finally, all these features are concatenated to produce a joint representation of the objects for VoD. The experiments conducted on the ImageNet VID dataset demonstrate that the proposed method outperforms existing VoD methods and achieves a performance comparable to that of state-of-the-art VoDs.
2102.07344
Ana Stanescu
Ana Stanescu and Gaurav Pandey
Developing parsimonious ensembles using predictor diversity within a reinforcement learning framework
This work was intended as a replacement of arXiv:1805.02103 and any subsequent updates will appear there
null
null
null
cs.LG
http://creativecommons.org/licenses/by/4.0/
Heterogeneous ensembles that can aggregate an unrestricted number and variety of base predictors can effectively address challenging prediction problems. In particular, accurate ensembles that are also parsimonious, i.e., consist of as few base predictors as possible, can help reveal potentially useful knowledge about the target problem domain. Although ensemble selection offers a potential approach to achieving these goals, the currently available algorithms are limited in their abilities. In this paper, we present several algorithms that incorporate ensemble diversity into a reinforcement learning (RL)-based ensemble selection framework to build accurate and parsimonious ensembles. These algorithms, as well as several baselines, are rigorously evaluated on datasets from diverse domains in terms of the predictive performance and parsimony of their ensembles. This evaluation demonstrates that our diversity-incorporated RL-based algorithms perform better than the others for constructing simultaneously accurate and parsimonious ensembles. These algorithms can eventually aid the interpretation or reverse engineering of predictive models assimilated into effective ensembles. To enable such a translation, an implementation of these algorithms, as well the experimental setup they are evaluated in, has been made available at https://github.com/GauravPandeyLab/lens-learning-ensembles-using-reinforcement-learning.
[ { "created": "Mon, 15 Feb 2021 05:00:19 GMT", "version": "v1" }, { "created": "Thu, 25 Feb 2021 23:43:42 GMT", "version": "v2" } ]
2021-03-01
[ [ "Stanescu", "Ana", "" ], [ "Pandey", "Gaurav", "" ] ]
Heterogeneous ensembles that can aggregate an unrestricted number and variety of base predictors can effectively address challenging prediction problems. In particular, accurate ensembles that are also parsimonious, i.e., consist of as few base predictors as possible, can help reveal potentially useful knowledge about the target problem domain. Although ensemble selection offers a potential approach to achieving these goals, the currently available algorithms are limited in their abilities. In this paper, we present several algorithms that incorporate ensemble diversity into a reinforcement learning (RL)-based ensemble selection framework to build accurate and parsimonious ensembles. These algorithms, as well as several baselines, are rigorously evaluated on datasets from diverse domains in terms of the predictive performance and parsimony of their ensembles. This evaluation demonstrates that our diversity-incorporated RL-based algorithms perform better than the others for constructing simultaneously accurate and parsimonious ensembles. These algorithms can eventually aid the interpretation or reverse engineering of predictive models assimilated into effective ensembles. To enable such a translation, an implementation of these algorithms, as well the experimental setup they are evaluated in, has been made available at https://github.com/GauravPandeyLab/lens-learning-ensembles-using-reinforcement-learning.
1910.06511
Ling-Xiao Zhang
Lin Gao, Ling-Xiao Zhang, Hsien-Yu Meng, Yi-Hui Ren, Yu-Kun Lai, Leif Kobbelt
PRS-Net: Planar Reflective Symmetry Detection Net for 3D Models
null
IEEE Transactions on Visualization and Computer Graphics, Volume: 27, Issue: 6, 2021
10.1109/TVCG.2020.3003823
null
cs.GR cs.LG
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
In geometry processing, symmetry is a universal type of high-level structural information of 3D models and benefits many geometry processing tasks including shape segmentation, alignment, matching, and completion. Thus it is an important problem to analyze various symmetry forms of 3D shapes. Planar reflective symmetry is the most fundamental one. Traditional methods based on spatial sampling can be time-consuming and may not be able to identify all the symmetry planes. In this paper, we present a novel learning framework to automatically discover global planar reflective symmetry of a 3D shape. Our framework trains an unsupervised 3D convolutional neural network to extract global model features and then outputs possible global symmetry parameters, where input shapes are represented using voxels. We introduce a dedicated symmetry distance loss along with a regularization loss to avoid generating duplicated symmetry planes. Our network can also identify generalized cylinders by predicting their rotation axes. We further provide a method to remove invalid and duplicated planes and axes. We demonstrate that our method is able to produce reliable and accurate results. Our neural network based method is hundreds of times faster than the state-of-the-art methods, which are based on sampling. Our method is also robust even with noisy or incomplete input surfaces.
[ { "created": "Tue, 15 Oct 2019 03:46:58 GMT", "version": "v1" }, { "created": "Fri, 18 Oct 2019 04:05:15 GMT", "version": "v2" }, { "created": "Thu, 24 Oct 2019 06:55:05 GMT", "version": "v3" }, { "created": "Sat, 16 May 2020 02:21:17 GMT", "version": "v4" }, { "created": "Mon, 1 Jun 2020 11:54:32 GMT", "version": "v5" }, { "created": "Tue, 14 Sep 2021 11:49:12 GMT", "version": "v6" } ]
2021-09-15
[ [ "Gao", "Lin", "" ], [ "Zhang", "Ling-Xiao", "" ], [ "Meng", "Hsien-Yu", "" ], [ "Ren", "Yi-Hui", "" ], [ "Lai", "Yu-Kun", "" ], [ "Kobbelt", "Leif", "" ] ]
In geometry processing, symmetry is a universal type of high-level structural information of 3D models and benefits many geometry processing tasks including shape segmentation, alignment, matching, and completion. Thus it is an important problem to analyze various symmetry forms of 3D shapes. Planar reflective symmetry is the most fundamental one. Traditional methods based on spatial sampling can be time-consuming and may not be able to identify all the symmetry planes. In this paper, we present a novel learning framework to automatically discover global planar reflective symmetry of a 3D shape. Our framework trains an unsupervised 3D convolutional neural network to extract global model features and then outputs possible global symmetry parameters, where input shapes are represented using voxels. We introduce a dedicated symmetry distance loss along with a regularization loss to avoid generating duplicated symmetry planes. Our network can also identify generalized cylinders by predicting their rotation axes. We further provide a method to remove invalid and duplicated planes and axes. We demonstrate that our method is able to produce reliable and accurate results. Our neural network based method is hundreds of times faster than the state-of-the-art methods, which are based on sampling. Our method is also robust even with noisy or incomplete input surfaces.
2103.03666
Benedikt Kleppmann
Benedikt T. Kleppmann
Tree of Knowledge: an Online Platform for Learning the Behaviour of Complex Systems
10 pages, 5 figures
null
null
null
cs.AI
http://creativecommons.org/licenses/by/4.0/
Many social sciences such as psychology and economics try to learn the behaviour of complex agents such as humans, organisations and countries. The current statistical methods used for learning this behaviour try to infer generally valid behaviour, but can only learn from one type of study at a time. Furthermore, only data from carefully designed studies can be used, as the phenomenon of interest has to be isolated and confounding factors accounted for. These restrictions limit the robustness and accuracy of insights that can be gained from social/economic systems. Here we present the online platform TreeOfKnowledge which implements a new methodology specifically designed for learning complex behaviours from complex systems: agent-based behaviour learning. With agent-based behaviour learning it is possible to gain more accurate and robust insights as it does not have the restriction of conventional statistics. It learns agent behaviour from many heterogenous datasets and can learn from these datasets even if the phenomenon of interest is not directly observed, but appears deep within complex systems. This new methodology shows how the internet and advances in computational power allow for more accurate and powerful mathematical models.
[ { "created": "Sat, 27 Feb 2021 19:39:14 GMT", "version": "v1" } ]
2021-03-08
[ [ "Kleppmann", "Benedikt T.", "" ] ]
Many social sciences such as psychology and economics try to learn the behaviour of complex agents such as humans, organisations and countries. The current statistical methods used for learning this behaviour try to infer generally valid behaviour, but can only learn from one type of study at a time. Furthermore, only data from carefully designed studies can be used, as the phenomenon of interest has to be isolated and confounding factors accounted for. These restrictions limit the robustness and accuracy of insights that can be gained from social/economic systems. Here we present the online platform TreeOfKnowledge which implements a new methodology specifically designed for learning complex behaviours from complex systems: agent-based behaviour learning. With agent-based behaviour learning it is possible to gain more accurate and robust insights as it does not have the restriction of conventional statistics. It learns agent behaviour from many heterogenous datasets and can learn from these datasets even if the phenomenon of interest is not directly observed, but appears deep within complex systems. This new methodology shows how the internet and advances in computational power allow for more accurate and powerful mathematical models.
1902.10439
Su Yang
Su Yang, Yuqing Zhang, Chensi Wu
Attack-Defense Quantification Based On Game-Theory
null
null
null
null
cs.CR
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
With the developing of the attack and defense technology, the cyber environment has been more and more sophisticated. We failed to give an accurate evaluation of network security situation, as we lack a more accurate quantitative evaluation of attack-defense behaviors. In response to this situation, we proposed an attack-defense stochastic game model (ADSGM), analyzed the different security property of distinct defense mechanism, and put forward a corresponding utility calculation coping with the distinct defense mechanism. Through a case study, we showed the impact of active defense and the risk of attack exposure, demonstrated the effectiveness of our methods on attack-defense behavior quantification. This paper filled with the gap in the quantitative assessment of defensive measures, to make the quantitative evaluation of attack-defense more comprehensive and accurate.
[ { "created": "Wed, 27 Feb 2019 10:28:34 GMT", "version": "v1" } ]
2019-02-28
[ [ "Yang", "Su", "" ], [ "Zhang", "Yuqing", "" ], [ "Wu", "Chensi", "" ] ]
With the developing of the attack and defense technology, the cyber environment has been more and more sophisticated. We failed to give an accurate evaluation of network security situation, as we lack a more accurate quantitative evaluation of attack-defense behaviors. In response to this situation, we proposed an attack-defense stochastic game model (ADSGM), analyzed the different security property of distinct defense mechanism, and put forward a corresponding utility calculation coping with the distinct defense mechanism. Through a case study, we showed the impact of active defense and the risk of attack exposure, demonstrated the effectiveness of our methods on attack-defense behavior quantification. This paper filled with the gap in the quantitative assessment of defensive measures, to make the quantitative evaluation of attack-defense more comprehensive and accurate.
2405.11320
Emmanouil Maragkoudakis
Emmanouil Maragkoudakis, Symeon Papadopoulos, Iraklis Varlamis and Christos Diou
Sampling Strategies for Mitigating Bias in Face Synthesis Methods
Accepted to the BIAS 2023 ECML-PKDD Workshop
null
null
null
cs.LG cs.CV
http://creativecommons.org/licenses/by/4.0/
Synthetically generated images can be used to create media content or to complement datasets for training image analysis models. Several methods have recently been proposed for the synthesis of high-fidelity face images; however, the potential biases introduced by such methods have not been sufficiently addressed. This paper examines the bias introduced by the widely popular StyleGAN2 generative model trained on the Flickr Faces HQ dataset and proposes two sampling strategies to balance the representation of selected attributes in the generated face images. We focus on two protected attributes, gender and age, and reveal that biases arise in the distribution of randomly sampled images against very young and very old age groups, as well as against female faces. These biases are also assessed for different image quality levels based on the GIQA score. To mitigate bias, we propose two alternative methods for sampling on selected lines or spheres of the latent space to increase the number of generated samples from the under-represented classes. The experimental results show a decrease in bias against underrepresented groups and a more uniform distribution of the protected features at different levels of image quality.
[ { "created": "Sat, 18 May 2024 15:30:14 GMT", "version": "v1" } ]
2024-05-21
[ [ "Maragkoudakis", "Emmanouil", "" ], [ "Papadopoulos", "Symeon", "" ], [ "Varlamis", "Iraklis", "" ], [ "Diou", "Christos", "" ] ]
Synthetically generated images can be used to create media content or to complement datasets for training image analysis models. Several methods have recently been proposed for the synthesis of high-fidelity face images; however, the potential biases introduced by such methods have not been sufficiently addressed. This paper examines the bias introduced by the widely popular StyleGAN2 generative model trained on the Flickr Faces HQ dataset and proposes two sampling strategies to balance the representation of selected attributes in the generated face images. We focus on two protected attributes, gender and age, and reveal that biases arise in the distribution of randomly sampled images against very young and very old age groups, as well as against female faces. These biases are also assessed for different image quality levels based on the GIQA score. To mitigate bias, we propose two alternative methods for sampling on selected lines or spheres of the latent space to increase the number of generated samples from the under-represented classes. The experimental results show a decrease in bias against underrepresented groups and a more uniform distribution of the protected features at different levels of image quality.
2205.09351
Arnab Dey
Arnab Dey, Yassine Ahmine, Andrew I. Comport
Mip-NeRF RGB-D: Depth Assisted Fast Neural Radiance Fields
null
Journal of WSCG 2022
10.24132/JWSCG.2022.5
Vol.30., No.1-2, ISSN 1213-6972
cs.CV
http://creativecommons.org/licenses/by-nc-sa/4.0/
Neural scene representations, such as Neural Radiance Fields (NeRF), are based on training a multilayer perceptron (MLP) using a set of color images with known poses. An increasing number of devices now produce RGB-D(color + depth) information, which has been shown to be very important for a wide range of tasks. Therefore, the aim of this paper is to investigate what improvements can be made to these promising implicit representations by incorporating depth information with the color images. In particular, the recently proposed Mip-NeRF approach, which uses conical frustums instead of rays for volume rendering, allows one to account for the varying area of a pixel with distance from the camera center. The proposed method additionally models depth uncertainty. This allows to address major limitations of NeRF-based approaches including improving the accuracy of geometry, reduced artifacts, faster training time, and shortened prediction time. Experiments are performed on well-known benchmark scenes, and comparisons show improved accuracy in scene geometry and photometric reconstruction, while reducing the training time by 3 - 5 times.
[ { "created": "Thu, 19 May 2022 07:11:42 GMT", "version": "v1" }, { "created": "Thu, 9 Jun 2022 11:35:53 GMT", "version": "v2" }, { "created": "Mon, 7 Nov 2022 13:57:58 GMT", "version": "v3" } ]
2022-11-08
[ [ "Dey", "Arnab", "" ], [ "Ahmine", "Yassine", "" ], [ "Comport", "Andrew I.", "" ] ]
Neural scene representations, such as Neural Radiance Fields (NeRF), are based on training a multilayer perceptron (MLP) using a set of color images with known poses. An increasing number of devices now produce RGB-D(color + depth) information, which has been shown to be very important for a wide range of tasks. Therefore, the aim of this paper is to investigate what improvements can be made to these promising implicit representations by incorporating depth information with the color images. In particular, the recently proposed Mip-NeRF approach, which uses conical frustums instead of rays for volume rendering, allows one to account for the varying area of a pixel with distance from the camera center. The proposed method additionally models depth uncertainty. This allows to address major limitations of NeRF-based approaches including improving the accuracy of geometry, reduced artifacts, faster training time, and shortened prediction time. Experiments are performed on well-known benchmark scenes, and comparisons show improved accuracy in scene geometry and photometric reconstruction, while reducing the training time by 3 - 5 times.