id stringlengths 9 14 | submitter stringlengths 1 64 ⌀ | authors stringlengths 4 9.62k | title stringlengths 4 343 | comments stringlengths 1 609 ⌀ | journal-ref stringlengths 4 404 ⌀ | doi stringlengths 12 153 ⌀ | report-no stringlengths 2 254 ⌀ | categories stringlengths 5 112 | license stringclasses 9 values | orig_abstract stringlengths 14 3.76k | versions listlengths 1 60 | update_date stringlengths 10 10 | authors_parsed listlengths 1 535 | abstract stringlengths 11 3.75k |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2011.09533 | Christian Schroeder de Witt | Christian Schroeder de Witt, Tarun Gupta, Denys Makoviichuk, Viktor
Makoviychuk, Philip H.S. Torr, Mingfei Sun, Shimon Whiteson | Is Independent Learning All You Need in the StarCraft Multi-Agent
Challenge? | null | null | null | null | cs.AI | http://creativecommons.org/licenses/by/4.0/ | Most recently developed approaches to cooperative multi-agent reinforcement
learning in the \emph{centralized training with decentralized execution}
setting involve estimating a centralized, joint value function. In this paper,
we demonstrate that, despite its various theoretical shortcomings, Independent
PPO (IPPO), a form of independent learning in which each agent simply estimates
its local value function, can perform just as well as or better than
state-of-the-art joint learning approaches on popular multi-agent benchmark
suite SMAC with little hyperparameter tuning. We also compare IPPO to several
variants; the results suggest that IPPO's strong performance may be due to its
robustness to some forms of environment non-stationarity.
| [
{
"created": "Wed, 18 Nov 2020 20:29:59 GMT",
"version": "v1"
}
] | 2020-11-20 | [
[
"de Witt",
"Christian Schroeder",
""
],
[
"Gupta",
"Tarun",
""
],
[
"Makoviichuk",
"Denys",
""
],
[
"Makoviychuk",
"Viktor",
""
],
[
"Torr",
"Philip H. S.",
""
],
[
"Sun",
"Mingfei",
""
],
[
"Whiteson",
"Shimon... | Most recently developed approaches to cooperative multi-agent reinforcement learning in the \emph{centralized training with decentralized execution} setting involve estimating a centralized, joint value function. In this paper, we demonstrate that, despite its various theoretical shortcomings, Independent PPO (IPPO), a form of independent learning in which each agent simply estimates its local value function, can perform just as well as or better than state-of-the-art joint learning approaches on popular multi-agent benchmark suite SMAC with little hyperparameter tuning. We also compare IPPO to several variants; the results suggest that IPPO's strong performance may be due to its robustness to some forms of environment non-stationarity. |
1807.09464 | Colas Le Guernic | Julien Duchene (CALID, LAAS-TSF), Eric Alata (LAAS-TSF), Vincent
Nicomette (LAAS-TSF), Mohamed Ka\^aniche (LAAS-TSF), Colas Le Guernic
(DGA.MI, TAMIS) | Specification-Based Protocol Obfuscation | null | 2018 48th Annual IEEE/IFIP International Conference on Dependable
Systems and Networks (DSN), Jun 2018, Luxembourg City, France. IEEE, 2018 | 10.1109/DSN.2018.00056 | null | cs.CR cs.NI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper proposes a new obfuscation technique of a communication protocol
that is aimed at making the reverse engineering of the protocol more complex.
The obfuscation is based on the transformation of protocol message format
specification. The obfuscating transformations are applied to the Abstract
Syntax Tree (AST) representation of the messages and mainly concern the
ordering or aggregation of the AST nodes. The paper also presents the design of
a framework that implements the proposed obfuscation technique by automatically
generating, from the specification of the message format, a library performing
the corresponding transformations. Finally, our framework is applied to two
real application protocols (Modbus and HTTP) to illustrate the relevance and
efficiency of the proposed approach. Various metrics recorded from the
experiments show the significant increase of the complexity of the obfuscated
protocol binary compared to the non-obfuscated code. It is also shown that the
execution time and memory overheads remain acceptable for a practical
deployment of the approach in operation.
| [
{
"created": "Wed, 25 Jul 2018 07:49:25 GMT",
"version": "v1"
}
] | 2018-07-26 | [
[
"Duchene",
"Julien",
"",
"CALID, LAAS-TSF"
],
[
"Alata",
"Eric",
"",
"LAAS-TSF"
],
[
"Nicomette",
"Vincent",
"",
"LAAS-TSF"
],
[
"Kaâniche",
"Mohamed",
"",
"LAAS-TSF"
],
[
"Guernic",
"Colas Le",
"",
"DGA.MI, TAMIS"
]... | This paper proposes a new obfuscation technique of a communication protocol that is aimed at making the reverse engineering of the protocol more complex. The obfuscation is based on the transformation of protocol message format specification. The obfuscating transformations are applied to the Abstract Syntax Tree (AST) representation of the messages and mainly concern the ordering or aggregation of the AST nodes. The paper also presents the design of a framework that implements the proposed obfuscation technique by automatically generating, from the specification of the message format, a library performing the corresponding transformations. Finally, our framework is applied to two real application protocols (Modbus and HTTP) to illustrate the relevance and efficiency of the proposed approach. Various metrics recorded from the experiments show the significant increase of the complexity of the obfuscated protocol binary compared to the non-obfuscated code. It is also shown that the execution time and memory overheads remain acceptable for a practical deployment of the approach in operation. |
1706.01269 | Alex Jourjine | Alex Jourjine | Extended Gauge Theory, Bi-Spinors, and Scalar Supersymmetry | Typos cleanup. 9 pages | null | null | null | hep-th hep-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Within the context of the extended bi-spinor gauge theory we describe a new
off-shell realization of scalar supersymmetry (s-susy) of massless interacting
fields with U(1), U(1) x SU(N) and U(1) x SU(N_1) x SU(N_2) gauge groups.
S-susy acts in the space of graded differential forms. The realization is
non-linear in the non-abelian case. S-susy would not require the doubling of
the SM particle spectrum. Instead, essentially only the forth generation of
quarks and leptons would be needed as extra field content. The theory is by
construction globally U(2,2) invariant and is an example of a supersymmetric
CFT.
| [
{
"created": "Mon, 5 Jun 2017 10:46:20 GMT",
"version": "v1"
},
{
"created": "Tue, 26 Sep 2017 16:49:19 GMT",
"version": "v2"
},
{
"created": "Mon, 16 Oct 2017 18:11:58 GMT",
"version": "v3"
}
] | 2017-10-18 | [
[
"Jourjine",
"Alex",
""
]
] | Within the context of the extended bi-spinor gauge theory we describe a new off-shell realization of scalar supersymmetry (s-susy) of massless interacting fields with U(1), U(1) x SU(N) and U(1) x SU(N_1) x SU(N_2) gauge groups. S-susy acts in the space of graded differential forms. The realization is non-linear in the non-abelian case. S-susy would not require the doubling of the SM particle spectrum. Instead, essentially only the forth generation of quarks and leptons would be needed as extra field content. The theory is by construction globally U(2,2) invariant and is an example of a supersymmetric CFT. |
2401.13213 | Miao Zhang | Miao Zhang, Zee fryer, Ben Colman, Ali Shahriyari, Gaurav Bharaj | Common-Sense Bias Discovery and Mitigation for Classification Tasks | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Machine learning model bias can arise from dataset composition: sensitive
features correlated to the learning target disturb the model decision rule and
lead to performance differences along the features. Existing de-biasing work
captures prominent and delicate image features which are traceable in model
latent space, like colors of digits or background of animals. However, using
the latent space is not sufficient to understand all dataset feature
correlations. In this work, we propose a framework to extract feature clusters
in a dataset based on image descriptions, allowing us to capture both subtle
and coarse features of the images. The feature co-occurrence pattern is
formulated and correlation is measured, utilizing a human-in-the-loop for
examination. The analyzed features and correlations are human-interpretable, so
we name the method Common-Sense Bias Discovery (CSBD). Having exposed sensitive
correlations in a dataset, we demonstrate that downstream model bias can be
mitigated by adjusting image sampling weights, without requiring a sensitive
group label supervision. Experiments show that our method discovers novel
biases on multiple classification tasks for two benchmark image datasets, and
the intervention outperforms state-of-the-art unsupervised bias mitigation
methods.
| [
{
"created": "Wed, 24 Jan 2024 03:56:07 GMT",
"version": "v1"
},
{
"created": "Thu, 8 Feb 2024 05:38:54 GMT",
"version": "v2"
}
] | 2024-02-09 | [
[
"Zhang",
"Miao",
""
],
[
"fryer",
"Zee",
""
],
[
"Colman",
"Ben",
""
],
[
"Shahriyari",
"Ali",
""
],
[
"Bharaj",
"Gaurav",
""
]
] | Machine learning model bias can arise from dataset composition: sensitive features correlated to the learning target disturb the model decision rule and lead to performance differences along the features. Existing de-biasing work captures prominent and delicate image features which are traceable in model latent space, like colors of digits or background of animals. However, using the latent space is not sufficient to understand all dataset feature correlations. In this work, we propose a framework to extract feature clusters in a dataset based on image descriptions, allowing us to capture both subtle and coarse features of the images. The feature co-occurrence pattern is formulated and correlation is measured, utilizing a human-in-the-loop for examination. The analyzed features and correlations are human-interpretable, so we name the method Common-Sense Bias Discovery (CSBD). Having exposed sensitive correlations in a dataset, we demonstrate that downstream model bias can be mitigated by adjusting image sampling weights, without requiring a sensitive group label supervision. Experiments show that our method discovers novel biases on multiple classification tasks for two benchmark image datasets, and the intervention outperforms state-of-the-art unsupervised bias mitigation methods. |
hep-th/0703257 | Masakazu Sano | Masakazu Sano, Hisao Suzuki | Integrable Cosmological Models From Higher Dimensional Einstein
Equations | 10 pages, 2 figures, added reference, corrected typos(v2),
explanation improved and references and acknowledgments added, accepted for
publication in PRD(v3) | Phys.Rev.D76:064006,2007 | 10.1103/PhysRevD.76.064006 | EPHOU 07-003 | hep-th gr-qc | null | We consider the cosmological models for the higher dimensional spacetime
which includes the curvatures of our space as well as the curvatures of the
internal space. We find that the condition for the integrability of the
cosmological equations is that the total space-time dimensions are D=10 or D=11
which is exactly the conditions for superstrings or M-theory. We obtain
analytic solutions with generic initial conditions in the four dimensional
Einstein frame and study the accelerating universe when both our space and the
internal space have negative curvatures.
| [
{
"created": "Wed, 28 Mar 2007 12:01:45 GMT",
"version": "v1"
},
{
"created": "Thu, 5 Apr 2007 05:04:55 GMT",
"version": "v2"
},
{
"created": "Thu, 19 Jul 2007 02:34:21 GMT",
"version": "v3"
}
] | 2008-11-26 | [
[
"Sano",
"Masakazu",
""
],
[
"Suzuki",
"Hisao",
""
]
] | We consider the cosmological models for the higher dimensional spacetime which includes the curvatures of our space as well as the curvatures of the internal space. We find that the condition for the integrability of the cosmological equations is that the total space-time dimensions are D=10 or D=11 which is exactly the conditions for superstrings or M-theory. We obtain analytic solutions with generic initial conditions in the four dimensional Einstein frame and study the accelerating universe when both our space and the internal space have negative curvatures. |
2310.12124 | Luca Smaldone Ph.D | Giuseppe Gaetano Luciano and Luca Smaldone | Time-energy uncertainty relation for neutrino oscillations: historical
development, applications and future prospects | 20 pages, published version | Symmetry 15(11), 2032 (2023) | 10.3390/sym15112032 | null | hep-th gr-qc quant-ph | http://creativecommons.org/licenses/by/4.0/ | Time-energy uncertainty relation (TEUR) plays a fundamental role in quantum
mechanics, as it allows to grasp peculiar aspects of a variety of phenomena
based on very general principles and symmetries of the theory. Using the
Mandelstam-Tamm method, TEUR has been recently derived for neutrino
oscillations by connecting the uncertainty on neutrino energy with the
characteristic time-scale of oscillations. Interestingly enough, the suggestive
interpretation of neutrinos as unstable-like particles has proved to naturally
emerge in this context. Further aspects have been later discussed in
semiclassical gravity by computing corrections to the neutrino energy
uncertainty in a generic stationary curved spacetime, and in quantum field
theory, where the clock observable turns out to be identified with the
non-conserved flavor charge operator. In the present work, we give an overview
on the above achievements. In particular, we analyze the implications of TEUR
and explore the impact of gravitational and non-relativistic effects on the
standard condition for neutrino oscillations.
| [
{
"created": "Wed, 18 Oct 2023 17:31:14 GMT",
"version": "v1"
},
{
"created": "Thu, 9 Nov 2023 10:53:41 GMT",
"version": "v2"
}
] | 2023-11-10 | [
[
"Luciano",
"Giuseppe Gaetano",
""
],
[
"Smaldone",
"Luca",
""
]
] | Time-energy uncertainty relation (TEUR) plays a fundamental role in quantum mechanics, as it allows to grasp peculiar aspects of a variety of phenomena based on very general principles and symmetries of the theory. Using the Mandelstam-Tamm method, TEUR has been recently derived for neutrino oscillations by connecting the uncertainty on neutrino energy with the characteristic time-scale of oscillations. Interestingly enough, the suggestive interpretation of neutrinos as unstable-like particles has proved to naturally emerge in this context. Further aspects have been later discussed in semiclassical gravity by computing corrections to the neutrino energy uncertainty in a generic stationary curved spacetime, and in quantum field theory, where the clock observable turns out to be identified with the non-conserved flavor charge operator. In the present work, we give an overview on the above achievements. In particular, we analyze the implications of TEUR and explore the impact of gravitational and non-relativistic effects on the standard condition for neutrino oscillations. |
2405.09155 | Manoj Gulati | Lim Chang Quan Thaddeus, C. Rajashekar Reddy, Yuvraj Singh Bhadauria,
Dhairya Shah, Manoj Gulati, Ambuj Varshney | TunnelSense: Low-power, Non-Contact Sensing using Tunnel Diodes | This work is accepted at IEEE RFID 2024 | null | null | null | cs.ET | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Sensing the motion of physical objects in an environment enables numerous
applications, from tracking occupancy in buildings and monitoring vital signs
to diagnosing faults in machines. Typically, these application scenarios
involve attaching a sensor, such as an accelerometer, to the object of
interest, like a wearable device that tracks our steps. However, many of these
scenarios require tracking motion in a noncontact manner where the sensor is
not in touch with the object. A sensor in such a scenario observes variations
in radio, light, acoustic, and infrared fields disturbed by the object's
motion. Current noncontact sensing mechanisms often require substantial energy
and involve complex processing on sophisticated hardware. We present
TunnelSense, a novel mechanism that rethinks noncontact sensing using tunnel
diode oscillators. They are highly sensitive to changes in their
electromagnetic environments. The motion of an object near a tunnel diode
oscillator induces corresponding changes in its resonant frequency and thus in
the generated radio waves. Additionally, the low-power characteristics of the
tunnel diode allow tags designed using them to operate on less than
100microwatt of power consumption and with a biasing voltage starting at 70
millivolt. This enables prolonged tag operation on a small battery or energy
harvested from the environment. Among numerous applications enabled by the
TunnelSense system, this work demonstrates its ability to detect breathing at
distances up to 30 centimeter between the subject and the TunnelSense tag.
| [
{
"created": "Wed, 15 May 2024 07:39:13 GMT",
"version": "v1"
}
] | 2024-05-16 | [
[
"Thaddeus",
"Lim Chang Quan",
""
],
[
"Reddy",
"C. Rajashekar",
""
],
[
"Bhadauria",
"Yuvraj Singh",
""
],
[
"Shah",
"Dhairya",
""
],
[
"Gulati",
"Manoj",
""
],
[
"Varshney",
"Ambuj",
""
]
] | Sensing the motion of physical objects in an environment enables numerous applications, from tracking occupancy in buildings and monitoring vital signs to diagnosing faults in machines. Typically, these application scenarios involve attaching a sensor, such as an accelerometer, to the object of interest, like a wearable device that tracks our steps. However, many of these scenarios require tracking motion in a noncontact manner where the sensor is not in touch with the object. A sensor in such a scenario observes variations in radio, light, acoustic, and infrared fields disturbed by the object's motion. Current noncontact sensing mechanisms often require substantial energy and involve complex processing on sophisticated hardware. We present TunnelSense, a novel mechanism that rethinks noncontact sensing using tunnel diode oscillators. They are highly sensitive to changes in their electromagnetic environments. The motion of an object near a tunnel diode oscillator induces corresponding changes in its resonant frequency and thus in the generated radio waves. Additionally, the low-power characteristics of the tunnel diode allow tags designed using them to operate on less than 100microwatt of power consumption and with a biasing voltage starting at 70 millivolt. This enables prolonged tag operation on a small battery or energy harvested from the environment. Among numerous applications enabled by the TunnelSense system, this work demonstrates its ability to detect breathing at distances up to 30 centimeter between the subject and the TunnelSense tag. |
hep-th/0104060 | Li Yu Qi | Han-Ying Guo, Xiao-mei Ji, Yu-Qi Li, and Ke Wu | A Note on Symplectic, Multisymplectic Scheme in Finite Element Method | 7 pages, 3 figures | null | 10.1088/0253-6102/36/3/259 | null | hep-th | null | We find that with uniform mesh, the numerical schemes derived from finite
element method can keep a preserved symplectic structure in one-dimensional
case and a preserved multisymplectic structure in two-dimentional case in
certain discrete version respectively. These results are in fact the intrinsic
reason that the numerical experiments indicate that such finite element
algorithms are accurate in practice.
| [
{
"created": "Fri, 6 Apr 2001 09:03:42 GMT",
"version": "v1"
}
] | 2018-01-17 | [
[
"Guo",
"Han-Ying",
""
],
[
"Ji",
"Xiao-mei",
""
],
[
"Li",
"Yu-Qi",
""
],
[
"Wu",
"Ke",
""
]
] | We find that with uniform mesh, the numerical schemes derived from finite element method can keep a preserved symplectic structure in one-dimensional case and a preserved multisymplectic structure in two-dimentional case in certain discrete version respectively. These results are in fact the intrinsic reason that the numerical experiments indicate that such finite element algorithms are accurate in practice. |
2207.07597 | Minsang Kim | Minsang Kim, Sang-hyun Je, Eunjoo Park | OASYS: Domain-Agnostic Automated System for Constructing Knowledge Base
from Unstructured Text | ACM SIGKDD Workshop on Mining and Learning with Graphs 2022, Accepted | null | null | null | cs.CL cs.AI cs.LG | http://creativecommons.org/licenses/by-nc-nd/4.0/ | In recent years, creating and managing knowledge bases have become crucial to
the retail product and enterprise domains. We present an automatic knowledge
base construction system that mines data from documents. This system can
generate training data during the training process without human intervention.
Therefore, it is domain-agnostic trainable using only the target domain text
corpus and a pre-defined knowledge base. This system is called OASYS and is the
first system built with the Korean language in mind. In addition, we also have
constructed a new human-annotated benchmark dataset of the Korean Wikipedia
corpus paired with a Korean DBpedia to aid system evaluation. The system
performance results on human-annotated benchmark test dataset are meaningful
and show that the generated knowledge base from OASYS trained on only
auto-generated data is useful. We provide both a human-annotated test dataset
and an auto-generated dataset.
| [
{
"created": "Wed, 29 Jun 2022 22:03:38 GMT",
"version": "v1"
}
] | 2022-07-18 | [
[
"Kim",
"Minsang",
""
],
[
"Je",
"Sang-hyun",
""
],
[
"Park",
"Eunjoo",
""
]
] | In recent years, creating and managing knowledge bases have become crucial to the retail product and enterprise domains. We present an automatic knowledge base construction system that mines data from documents. This system can generate training data during the training process without human intervention. Therefore, it is domain-agnostic trainable using only the target domain text corpus and a pre-defined knowledge base. This system is called OASYS and is the first system built with the Korean language in mind. In addition, we also have constructed a new human-annotated benchmark dataset of the Korean Wikipedia corpus paired with a Korean DBpedia to aid system evaluation. The system performance results on human-annotated benchmark test dataset are meaningful and show that the generated knowledge base from OASYS trained on only auto-generated data is useful. We provide both a human-annotated test dataset and an auto-generated dataset. |
hep-th/0202142 | Robert Brandenberger | Robert H. Brandenberger (CERN & Brown Univ.), Jerome Martin (IAP) | On Signatures of Short Distance Physics in the Cosmic Microwave
Background | 11 pages, 4 figures | Int.J.Mod.Phys. A17 (2002) 3663-3680 | 10.1142/S0217751X02010765 | BRWON-HET-1302 | hep-th astro-ph gr-qc hep-ph | null | Following a self-contained review of the basics of the theory of cosmological
perturbations, we discuss why the conclusions reached in the recent paper by
Kaloper et al are too pessimistic estimates of the amplitude of possible
imprints of trans-Planckian (string) physics on the spectrum of cosmic
microwave anisotropies in an inflationary Universe. It is shown that the likely
origin of large trans-Planckian effects on late time cosmological fluctuations
comes from nonadiabatic evolution of the state of fluctuations while the
wavelength is smaller than the Planck (string) scale, resulting in an excited
state at the time that the wavelength crosses the Hubble radius during
inflation.
| [
{
"created": "Thu, 21 Feb 2002 11:35:00 GMT",
"version": "v1"
}
] | 2016-09-06 | [
[
"Brandenberger",
"Robert H.",
"",
"CERN & Brown Univ."
],
[
"Martin",
"Jerome",
"",
"IAP"
]
] | Following a self-contained review of the basics of the theory of cosmological perturbations, we discuss why the conclusions reached in the recent paper by Kaloper et al are too pessimistic estimates of the amplitude of possible imprints of trans-Planckian (string) physics on the spectrum of cosmic microwave anisotropies in an inflationary Universe. It is shown that the likely origin of large trans-Planckian effects on late time cosmological fluctuations comes from nonadiabatic evolution of the state of fluctuations while the wavelength is smaller than the Planck (string) scale, resulting in an excited state at the time that the wavelength crosses the Hubble radius during inflation. |
hep-th/0212226 | Andrei Ivanov | M. Faber, A. N. Ivanov | On the ground state of a free massless (pseudo)scalar field in two
dimensions | 20 pages, Latex, no figures | null | null | null | hep-th | null | We investigate the ground state of a free massless (pseudo)scalar field in
1+1-dimensional space-time. We argue that in the quantum field theory of a free
massless (pseudo)scalar field without infrared divergences (Eur. Phys. J. C24,
653 (2002)) the ground state can be represented by a tensor product of wave
functions of the fiducial vacuum and of the collective zero-mode, describing
the motion of the ``center of mass'' of a free massless (pseudo)scalar field.
We show that the bosonized version of the BCS wave function of the ground state
of the massless Thirring model obtained in (Phys.Lett. B563, 231 (2003))
describes the ground state of the free massless (pseudo)scalar field.
| [
{
"created": "Wed, 18 Dec 2002 22:07:01 GMT",
"version": "v1"
},
{
"created": "Sun, 29 Jun 2003 09:08:29 GMT",
"version": "v2"
}
] | 2007-05-23 | [
[
"Faber",
"M.",
""
],
[
"Ivanov",
"A. N.",
""
]
] | We investigate the ground state of a free massless (pseudo)scalar field in 1+1-dimensional space-time. We argue that in the quantum field theory of a free massless (pseudo)scalar field without infrared divergences (Eur. Phys. J. C24, 653 (2002)) the ground state can be represented by a tensor product of wave functions of the fiducial vacuum and of the collective zero-mode, describing the motion of the ``center of mass'' of a free massless (pseudo)scalar field. We show that the bosonized version of the BCS wave function of the ground state of the massless Thirring model obtained in (Phys.Lett. B563, 231 (2003)) describes the ground state of the free massless (pseudo)scalar field. |
2111.07492 | Chen Ma | Chen Ma, Xiangyu Guo, Li Chen, Jun-Hai Yong, Yisen Wang | Finding Optimal Tangent Points for Reducing Distortions of Hard-label
Attacks | Accepted at NeurIPS 2021. The missing square term in Eqn.(13), as
well as many other mistakes of the previous version, have been fixed in the
current version | null | null | null | cs.CV cs.CR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | One major problem in black-box adversarial attacks is the high query
complexity in the hard-label attack setting, where only the top-1 predicted
label is available. In this paper, we propose a novel geometric-based approach
called Tangent Attack (TA), which identifies an optimal tangent point of a
virtual hemisphere located on the decision boundary to reduce the distortion of
the attack. Assuming the decision boundary is locally flat, we theoretically
prove that the minimum $\ell_2$ distortion can be obtained by reaching the
decision boundary along the tangent line passing through such tangent point in
each iteration. To improve the robustness of our method, we further propose a
generalized method which replaces the hemisphere with a semi-ellipsoid to adapt
to curved decision boundaries. Our approach is free of pre-training. Extensive
experiments conducted on the ImageNet and CIFAR-10 datasets demonstrate that
our approach can consume only a small number of queries to achieve the
low-magnitude distortion. The implementation source code is released online at
https://github.com/machanic/TangentAttack.
| [
{
"created": "Mon, 15 Nov 2021 01:51:37 GMT",
"version": "v1"
},
{
"created": "Thu, 18 Nov 2021 05:21:57 GMT",
"version": "v2"
},
{
"created": "Thu, 16 Dec 2021 13:20:41 GMT",
"version": "v3"
},
{
"created": "Sun, 16 Jan 2022 09:41:09 GMT",
"version": "v4"
},
{
"c... | 2022-03-01 | [
[
"Ma",
"Chen",
""
],
[
"Guo",
"Xiangyu",
""
],
[
"Chen",
"Li",
""
],
[
"Yong",
"Jun-Hai",
""
],
[
"Wang",
"Yisen",
""
]
] | One major problem in black-box adversarial attacks is the high query complexity in the hard-label attack setting, where only the top-1 predicted label is available. In this paper, we propose a novel geometric-based approach called Tangent Attack (TA), which identifies an optimal tangent point of a virtual hemisphere located on the decision boundary to reduce the distortion of the attack. Assuming the decision boundary is locally flat, we theoretically prove that the minimum $\ell_2$ distortion can be obtained by reaching the decision boundary along the tangent line passing through such tangent point in each iteration. To improve the robustness of our method, we further propose a generalized method which replaces the hemisphere with a semi-ellipsoid to adapt to curved decision boundaries. Our approach is free of pre-training. Extensive experiments conducted on the ImageNet and CIFAR-10 datasets demonstrate that our approach can consume only a small number of queries to achieve the low-magnitude distortion. The implementation source code is released online at https://github.com/machanic/TangentAttack. |
2403.20312 | Jaisidh Singh | Jaisidh Singh, Ishaan Shrivastava, Mayank Vatsa, Richa Singh, Aparna
Bharati | Learn "No" to Say "Yes" Better: Improving Vision-Language Models via
Negations | 14 pages + 6 figures in main manuscript (excluding references) | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Existing vision-language models (VLMs) treat text descriptions as a unit,
confusing individual concepts in a prompt and impairing visual semantic
matching and reasoning. An important aspect of reasoning in logic and language
is negations. This paper highlights the limitations of popular VLMs such as
CLIP, at understanding the implications of negations, i.e., the effect of the
word "not" in a given prompt. To enable evaluation of VLMs on fluent prompts
with negations, we present CC-Neg, a dataset containing 228,246 images, true
captions and their corresponding negated captions. Using CC-Neg along with
modifications to the contrastive loss of CLIP, our proposed CoN-CLIP framework,
has an improved understanding of negations. This training paradigm improves
CoN-CLIP's ability to encode semantics reliably, resulting in 3.85% average
gain in top-1 accuracy for zero-shot image classification across 8 datasets.
Further, CoN-CLIP outperforms CLIP on challenging compositionality benchmarks
such as SugarCREPE by 4.4%, showcasing emergent compositional understanding of
objects, relations, and attributes in text. Overall, our work addresses a
crucial limitation of VLMs by introducing a dataset and framework that
strengthens semantic associations between images and text, demonstrating
improved large-scale foundation models with significantly reduced computational
cost, promoting efficiency and accessibility.
| [
{
"created": "Fri, 29 Mar 2024 17:33:42 GMT",
"version": "v1"
}
] | 2024-04-01 | [
[
"Singh",
"Jaisidh",
""
],
[
"Shrivastava",
"Ishaan",
""
],
[
"Vatsa",
"Mayank",
""
],
[
"Singh",
"Richa",
""
],
[
"Bharati",
"Aparna",
""
]
] | Existing vision-language models (VLMs) treat text descriptions as a unit, confusing individual concepts in a prompt and impairing visual semantic matching and reasoning. An important aspect of reasoning in logic and language is negations. This paper highlights the limitations of popular VLMs such as CLIP, at understanding the implications of negations, i.e., the effect of the word "not" in a given prompt. To enable evaluation of VLMs on fluent prompts with negations, we present CC-Neg, a dataset containing 228,246 images, true captions and their corresponding negated captions. Using CC-Neg along with modifications to the contrastive loss of CLIP, our proposed CoN-CLIP framework, has an improved understanding of negations. This training paradigm improves CoN-CLIP's ability to encode semantics reliably, resulting in 3.85% average gain in top-1 accuracy for zero-shot image classification across 8 datasets. Further, CoN-CLIP outperforms CLIP on challenging compositionality benchmarks such as SugarCREPE by 4.4%, showcasing emergent compositional understanding of objects, relations, and attributes in text. Overall, our work addresses a crucial limitation of VLMs by introducing a dataset and framework that strengthens semantic associations between images and text, demonstrating improved large-scale foundation models with significantly reduced computational cost, promoting efficiency and accessibility. |
1804.03242 | Junyu Liu | Ning Bao, Junyu Liu | Quantum complexity and the virial theorem | v2: add references and a footnote. v3: published version, with
concrete examples | JHEP 1808 (2018) 144 | 10.1007/JHEP08(2018)144 | CALT-TH-2018-016 | hep-th gr-qc quant-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | It is conjectured that in the geometric formulation of quantum computing, one
can study quantum complexity through classical entropy of statistical ensembles
established non-relativistically in the group manifold of unitary operators.
The kinetic and positional decompositions of statistical entropy are
conjectured to correspond to the Kolmogorov complexity and computational
complexity, respectively, of corresponding quantum circuits. In this paper, we
claim that by applying the virial theorem to the group manifold, one can derive
a generic relation between Kolmogorov complexity and computational complexity
in the thermal equilibrium.
| [
{
"created": "Mon, 9 Apr 2018 21:20:38 GMT",
"version": "v1"
},
{
"created": "Tue, 15 May 2018 22:50:20 GMT",
"version": "v2"
},
{
"created": "Tue, 21 Aug 2018 19:33:10 GMT",
"version": "v3"
}
] | 2018-08-27 | [
[
"Bao",
"Ning",
""
],
[
"Liu",
"Junyu",
""
]
] | It is conjectured that in the geometric formulation of quantum computing, one can study quantum complexity through classical entropy of statistical ensembles established non-relativistically in the group manifold of unitary operators. The kinetic and positional decompositions of statistical entropy are conjectured to correspond to the Kolmogorov complexity and computational complexity, respectively, of corresponding quantum circuits. In this paper, we claim that by applying the virial theorem to the group manifold, one can derive a generic relation between Kolmogorov complexity and computational complexity in the thermal equilibrium. |
hep-th/0607017 | Han-Ying Guo | Han-Ying Guo | The Beltrami Model of De Sitter Space: From Snyder's quantized
space-time to de Sitter invariant relativity | 15 pages. Invited talk given at `International workshop on
noncommutative geometry and physics', Beijing, Nov. 7-10, 2005. To appear in
the proceedings | null | null | null | hep-th | null | In terms of the Beltrami model of de Sitter space we show that there is an
interchangeable relation between Snyder's quantized space-time model in
dS-space of momenta at the Planck length $\ell_P=(G\hbar c^{-3})^{1/2}$ and the
dS-invariant special relativity in dS-spacetime of radius
$R\simeq(3\Lambda^{-1})^{1/2}$, which is another fundamental length related to
the cosmological constant. Here, the cosmological constant $\Lambda$ is
regarded as a fundamental constant together with the speed of light $c$, Newton
constant $G$ and Planck constant $\hbar$. Furthermore, the physics at two
fundamental scales of length, the \dS-radius $R$ and the Planck length
$\ell_P$, should be dual to each other and linked via the gravity with local
dS-invariance characterized by a dimensionless coupling constant $g= \sqrt{3}
\ell_P/R\simeq(G\hbar c^{-3}\Lambda)^{1/2}\sim 10^{-61}$.
| [
{
"created": "Tue, 4 Jul 2006 04:09:19 GMT",
"version": "v1"
}
] | 2007-05-23 | [
[
"Guo",
"Han-Ying",
""
]
] | In terms of the Beltrami model of de Sitter space we show that there is an interchangeable relation between Snyder's quantized space-time model in dS-space of momenta at the Planck length $\ell_P=(G\hbar c^{-3})^{1/2}$ and the dS-invariant special relativity in dS-spacetime of radius $R\simeq(3\Lambda^{-1})^{1/2}$, which is another fundamental length related to the cosmological constant. Here, the cosmological constant $\Lambda$ is regarded as a fundamental constant together with the speed of light $c$, Newton constant $G$ and Planck constant $\hbar$. Furthermore, the physics at two fundamental scales of length, the \dS-radius $R$ and the Planck length $\ell_P$, should be dual to each other and linked via the gravity with local dS-invariance characterized by a dimensionless coupling constant $g= \sqrt{3} \ell_P/R\simeq(G\hbar c^{-3}\Lambda)^{1/2}\sim 10^{-61}$. |
2403.17369 | ZiYang Gong | Ziyang Gong, Fuhao Li, Yupeng Deng, Deblina Bhattacharjee, Xianzheng
Ma, Xiangwei Zhu, Zhenming Ji | CoDA: Instructive Chain-of-Domain Adaptation with Severity-Aware Visual
Prompt Tuning | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Unsupervised Domain Adaptation (UDA) aims to adapt models from labeled source
domains to unlabeled target domains. When adapting to adverse scenes, existing
UDA methods fail to perform well due to the lack of instructions, leading their
models to overlook discrepancies within all adverse scenes. To tackle this, we
propose CoDA which instructs models to distinguish, focus, and learn from these
discrepancies at scene and image levels. Specifically, CoDA consists of a
Chain-of-Domain (CoD) strategy and a Severity-Aware Visual Prompt Tuning
(SAVPT) mechanism. CoD focuses on scene-level instructions to divide all
adverse scenes into easy and hard scenes, guiding models to adapt from source
to easy domains with easy scene images, and then to hard domains with hard
scene images, thereby laying a solid foundation for whole adaptations. Building
upon this foundation, we employ SAVPT to dive into more detailed image-level
instructions to boost performance. SAVPT features a novel metric Severity that
divides all adverse scene images into low-severity and high-severity images.
Then Severity directs visual prompts and adapters, instructing models to
concentrate on unified severity features instead of scene-specific features,
without adding complexity to the model architecture. CoDA achieves SOTA
performances on widely-used benchmarks under all adverse scenes. Notably, CoDA
outperforms the existing ones by 4.6%, and 10.3% mIoU on the Foggy Driving, and
Foggy Zurich benchmarks, respectively. Our code is available at
https://github.com/Cuzyoung/CoDA
| [
{
"created": "Tue, 26 Mar 2024 04:09:08 GMT",
"version": "v1"
},
{
"created": "Thu, 4 Apr 2024 08:05:06 GMT",
"version": "v2"
},
{
"created": "Mon, 15 Jul 2024 06:34:03 GMT",
"version": "v3"
}
] | 2024-07-16 | [
[
"Gong",
"Ziyang",
""
],
[
"Li",
"Fuhao",
""
],
[
"Deng",
"Yupeng",
""
],
[
"Bhattacharjee",
"Deblina",
""
],
[
"Ma",
"Xianzheng",
""
],
[
"Zhu",
"Xiangwei",
""
],
[
"Ji",
"Zhenming",
""
]
] | Unsupervised Domain Adaptation (UDA) aims to adapt models from labeled source domains to unlabeled target domains. When adapting to adverse scenes, existing UDA methods fail to perform well due to the lack of instructions, leading their models to overlook discrepancies within all adverse scenes. To tackle this, we propose CoDA which instructs models to distinguish, focus, and learn from these discrepancies at scene and image levels. Specifically, CoDA consists of a Chain-of-Domain (CoD) strategy and a Severity-Aware Visual Prompt Tuning (SAVPT) mechanism. CoD focuses on scene-level instructions to divide all adverse scenes into easy and hard scenes, guiding models to adapt from source to easy domains with easy scene images, and then to hard domains with hard scene images, thereby laying a solid foundation for whole adaptations. Building upon this foundation, we employ SAVPT to dive into more detailed image-level instructions to boost performance. SAVPT features a novel metric Severity that divides all adverse scene images into low-severity and high-severity images. Then Severity directs visual prompts and adapters, instructing models to concentrate on unified severity features instead of scene-specific features, without adding complexity to the model architecture. CoDA achieves SOTA performances on widely-used benchmarks under all adverse scenes. Notably, CoDA outperforms the existing ones by 4.6%, and 10.3% mIoU on the Foggy Driving, and Foggy Zurich benchmarks, respectively. Our code is available at https://github.com/Cuzyoung/CoDA |
1704.04268 | John Rhodes | Elizabeth S. Allman and James H. Degnan and John A. Rhodes | Split probabilities and species tree inference under the multispecies
coalescent model | 43 pages | null | null | null | q-bio.PE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Using topological summaries of gene trees as a basis for species tree
inference is a promising approach to obtain acceptable speed on genomic-scale
datasets, and to avoid some undesirable modeling assumptions. Here we study the
probabilities of splits on gene trees under the multispecies coalescent model,
and how their features might inform species tree inference. After investigating
the behavior of split consensus methods, we investigate split invariants ---
that is, polynomial relationships between split probabilities. These invariants
are then used to show that, even though a split is an unrooted notion, split
probabilities retain enough information to identify the rooted species tree
topology for trees of more than 5 taxa, with one possible 6-taxon exception.
| [
{
"created": "Thu, 13 Apr 2017 19:45:39 GMT",
"version": "v1"
}
] | 2017-04-17 | [
[
"Allman",
"Elizabeth S.",
""
],
[
"Degnan",
"James H.",
""
],
[
"Rhodes",
"John A.",
""
]
] | Using topological summaries of gene trees as a basis for species tree inference is a promising approach to obtain acceptable speed on genomic-scale datasets, and to avoid some undesirable modeling assumptions. Here we study the probabilities of splits on gene trees under the multispecies coalescent model, and how their features might inform species tree inference. After investigating the behavior of split consensus methods, we investigate split invariants --- that is, polynomial relationships between split probabilities. These invariants are then used to show that, even though a split is an unrooted notion, split probabilities retain enough information to identify the rooted species tree topology for trees of more than 5 taxa, with one possible 6-taxon exception. |
hep-th/0104194 | Leonardo Castellani | L. Castellani and L. Sommovigo | Supersymmetric domain wall x G/H solutions of IIB supergravity | 8 pages, latex | null | null | DFTT-11/2001 | hep-th | null | 1-brane nonmaximally supersymmetric solutions of D=10 chiral supergravity are
discussed. In the dual frame, their near brane geometry is the product of a
3-dimensional domain wall spacetime and a 7-dimensional homogeneous Einstein
space G/H.
| [
{
"created": "Mon, 23 Apr 2001 16:00:44 GMT",
"version": "v1"
}
] | 2007-05-23 | [
[
"Castellani",
"L.",
""
],
[
"Sommovigo",
"L.",
""
]
] | 1-brane nonmaximally supersymmetric solutions of D=10 chiral supergravity are discussed. In the dual frame, their near brane geometry is the product of a 3-dimensional domain wall spacetime and a 7-dimensional homogeneous Einstein space G/H. |
hep-th/9506197 | Rainer Dick | Rainer Dick | Remarks on chiral symmetry breaking with massless fermions | LaTex, 7 pages, one misleading remark corrected and a comment added | null | null | null | hep-th hep-ph | null | In this talk I present recent results on Lorentz covariant correlation
functions $\langle q(p_1)\overline{q}(p_2)\rangle$ on the cone $p^2=0$. In
particular, chiral symmetry breaking terms are constructed which resemble
fermionic 2--point functions of 2--D CFT up to a scalar factor.
| [
{
"created": "Thu, 29 Jun 1995 18:45:35 GMT",
"version": "v1"
},
{
"created": "Thu, 6 Jul 1995 02:49:58 GMT",
"version": "v2"
}
] | 2008-02-03 | [
[
"Dick",
"Rainer",
""
]
] | In this talk I present recent results on Lorentz covariant correlation functions $\langle q(p_1)\overline{q}(p_2)\rangle$ on the cone $p^2=0$. In particular, chiral symmetry breaking terms are constructed which resemble fermionic 2--point functions of 2--D CFT up to a scalar factor. |
hep-th/0506001 | Valeri Frolov | Valeri P. Frolov, Werner Israel, and Andrei Zelnikov | Gravitational field of relativistic gyratons | 11 pages | Phys.Rev. D72 (2005) 084031 | 10.1103/PhysRevD.72.084031 | Alberta-Thy-08-05 | hep-th gr-qc | null | The metric ansatz is used to describe the gravitational field of a beam-pulse
of spinning radiation (gyraton) in an arbitrary number of spacetime dimensions
D. First we demonstrate that this metric belongs to the class of metrics for
which all scalar invariants constructed from the curvature and its covariant
derivatives vanish. Next, it is shown that the vacuum Einstein equations reduce
to two linear problems in (D-2)-dimensional Euclidean space. The first is to
find the static magnetic potential created by a point-like source. The second
requires finding the electric potential created by a point-like source
surrounded by given distribution of the electric charge. To obtain a generic
gyraton-type solution of the vacuum Einstein equations it is sufficient to
allow the coefficients in the corresponding harmonic decompositions of
solutions of the linear problems to depend arbitrarily on retarded time and
substitute the obtained expressions in the metric ansatz. We discuss properties
of the solutions for relativistic gyratons and consider special examples.
| [
{
"created": "Wed, 1 Jun 2005 00:06:46 GMT",
"version": "v1"
},
{
"created": "Thu, 1 Sep 2005 21:11:45 GMT",
"version": "v2"
},
{
"created": "Tue, 8 Nov 2005 20:57:24 GMT",
"version": "v3"
}
] | 2009-11-11 | [
[
"Frolov",
"Valeri P.",
""
],
[
"Israel",
"Werner",
""
],
[
"Zelnikov",
"Andrei",
""
]
] | The metric ansatz is used to describe the gravitational field of a beam-pulse of spinning radiation (gyraton) in an arbitrary number of spacetime dimensions D. First we demonstrate that this metric belongs to the class of metrics for which all scalar invariants constructed from the curvature and its covariant derivatives vanish. Next, it is shown that the vacuum Einstein equations reduce to two linear problems in (D-2)-dimensional Euclidean space. The first is to find the static magnetic potential created by a point-like source. The second requires finding the electric potential created by a point-like source surrounded by given distribution of the electric charge. To obtain a generic gyraton-type solution of the vacuum Einstein equations it is sufficient to allow the coefficients in the corresponding harmonic decompositions of solutions of the linear problems to depend arbitrarily on retarded time and substitute the obtained expressions in the metric ansatz. We discuss properties of the solutions for relativistic gyratons and consider special examples. |
1609.02318 | Alex Alvarado | Nikita A. Shevchenko, Stanislav A. Derevyanko, Jaroslaw E. Prilepsky,
Alex Alvarado, Polina Bayvel, and Sergei K. Turitsyn | Capacity Lower Bounds of the Noncentral Chi-Channel with Applications to
Soliton Amplitude Modulation | null | null | 10.1109/TCOMM.2018.2808286 | null | cs.IT math.IT physics.optics | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The channel law for amplitude-modulated solitons transmitted through a
nonlinear optical fibre with ideal distributed amplification and a receiver
based on the nonlinear Fourier transform is a noncentral chi-distribution with
$2n$ degrees of freedom, where $n=2$ and $n=3$ correspond to the single- and
dual-polarisation cases, respectively. In this paper, we study capacity lower
bounds of this channel under an average power constraint in bits per channel
use. We develop an asymptotic semi-analytic approximation for a capacity lower
bound for arbitrary $n$ and a Rayleigh input distribution. It is shown that
this lower bound grows logarithmically with signal-to-noise ratio (SNR),
independently of the value of $n$. Numerical results for other continuous input
distributions are also provided. A half-Gaussian input distribution is shown to
give larger rates than a Rayleigh input distribution for $n=1,2,3$. At an SNR
of $25$ dB, the best lower bounds we developed are approximately $3.68$ bit per
channel use. The practically relevant case of amplitude shift-keying (ASK)
constellations is also numerically analysed. For the same SNR of $25$ dB, a
$16$-ASK constellation yields a rate of approximately $3.45$ bit per channel
use.
| [
{
"created": "Thu, 8 Sep 2016 08:20:23 GMT",
"version": "v1"
},
{
"created": "Sat, 1 Apr 2017 16:46:21 GMT",
"version": "v2"
},
{
"created": "Sun, 10 Sep 2017 08:21:15 GMT",
"version": "v3"
},
{
"created": "Fri, 16 Feb 2018 16:04:17 GMT",
"version": "v4"
}
] | 2020-06-05 | [
[
"Shevchenko",
"Nikita A.",
""
],
[
"Derevyanko",
"Stanislav A.",
""
],
[
"Prilepsky",
"Jaroslaw E.",
""
],
[
"Alvarado",
"Alex",
""
],
[
"Bayvel",
"Polina",
""
],
[
"Turitsyn",
"Sergei K.",
""
]
] | The channel law for amplitude-modulated solitons transmitted through a nonlinear optical fibre with ideal distributed amplification and a receiver based on the nonlinear Fourier transform is a noncentral chi-distribution with $2n$ degrees of freedom, where $n=2$ and $n=3$ correspond to the single- and dual-polarisation cases, respectively. In this paper, we study capacity lower bounds of this channel under an average power constraint in bits per channel use. We develop an asymptotic semi-analytic approximation for a capacity lower bound for arbitrary $n$ and a Rayleigh input distribution. It is shown that this lower bound grows logarithmically with signal-to-noise ratio (SNR), independently of the value of $n$. Numerical results for other continuous input distributions are also provided. A half-Gaussian input distribution is shown to give larger rates than a Rayleigh input distribution for $n=1,2,3$. At an SNR of $25$ dB, the best lower bounds we developed are approximately $3.68$ bit per channel use. The practically relevant case of amplitude shift-keying (ASK) constellations is also numerically analysed. For the same SNR of $25$ dB, a $16$-ASK constellation yields a rate of approximately $3.45$ bit per channel use. |
1301.3106 | Syed Jafar | Syed A. Jafar | Topological Interference Management through Index Coding | Revised for the IEEE Transactions on Information Theory | null | 10.1109/TIT.2013.2285151 | null | cs.IT math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This work studies linear interference networks, both wired and wireless, with
no channel state information at the transmitters (CSIT) except a coarse
knowledge of the end-to-end one-hop topology of the network that only allows a
distinction between weak (zero) and significant (non-zero) channels and no
further knowledge of the channel coefficients' realizations. The network
capacity (wired) and DoF (wireless) are found to be bounded above by the
capacity of an index coding problem for which the antidote graph is the
complement of the given interference graph. The problems are shown to be
equivalent under linear solutions. An interference alignment perspective is
then used to translate the existing index coding solutions into the wired
network capacity and wireless network DoF solutions, as well as to find new and
unified solutions to different classes of all three problems.
| [
{
"created": "Mon, 14 Jan 2013 19:55:43 GMT",
"version": "v1"
},
{
"created": "Mon, 30 Sep 2013 01:12:25 GMT",
"version": "v2"
}
] | 2016-11-17 | [
[
"Jafar",
"Syed A.",
""
]
] | This work studies linear interference networks, both wired and wireless, with no channel state information at the transmitters (CSIT) except a coarse knowledge of the end-to-end one-hop topology of the network that only allows a distinction between weak (zero) and significant (non-zero) channels and no further knowledge of the channel coefficients' realizations. The network capacity (wired) and DoF (wireless) are found to be bounded above by the capacity of an index coding problem for which the antidote graph is the complement of the given interference graph. The problems are shown to be equivalent under linear solutions. An interference alignment perspective is then used to translate the existing index coding solutions into the wired network capacity and wireless network DoF solutions, as well as to find new and unified solutions to different classes of all three problems. |
2208.14827 | Niloofar Vardian | Niloofar Vardian | Entanglement Renormalization of the class of Continuous Matrix Product
States | 6 pages, 1 figure | null | 10.1103/PhysRevD.108.094029 | null | hep-th cond-mat.stat-mech cond-mat.str-el | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Continuous tensor network gives a variational ansatz for the ground state of
the quantum field theories (QFTs). The notable examples are the continuous
matrix product state (cMPS) and the continuous multiscale entanglement
renormalization ansatz (cMERA). While cMPS is just adapted to the
non-relativistic QFTs, only the Gaussian cMERA is well-understood which we can
not use to approximate the ground state of the interacting relativistic QFTs.
But instead, cMERA also corresponds to a real-space renormalization group flow
in the context of the wave functions. In this letter, we investigate the
backward Gaussian cMERA renormalization group flow of the class of cMPS by
putting the standard cMPS at the IR scale. At the UV scale, for the bosonic
systems in the thermodynamic limit, we achieve the variational class of states
that has been proposed recently as the relativistic cMPS (RCMPS) is adapted to
the relativistic QFTs without requiring to introduce of any additional IR or UV
cut-off. We also extend the RCMPS to fermionic systems and theories on a finite
circle.
| [
{
"created": "Wed, 31 Aug 2022 12:56:15 GMT",
"version": "v1"
}
] | 2023-11-23 | [
[
"Vardian",
"Niloofar",
""
]
] | Continuous tensor network gives a variational ansatz for the ground state of the quantum field theories (QFTs). The notable examples are the continuous matrix product state (cMPS) and the continuous multiscale entanglement renormalization ansatz (cMERA). While cMPS is just adapted to the non-relativistic QFTs, only the Gaussian cMERA is well-understood which we can not use to approximate the ground state of the interacting relativistic QFTs. But instead, cMERA also corresponds to a real-space renormalization group flow in the context of the wave functions. In this letter, we investigate the backward Gaussian cMERA renormalization group flow of the class of cMPS by putting the standard cMPS at the IR scale. At the UV scale, for the bosonic systems in the thermodynamic limit, we achieve the variational class of states that has been proposed recently as the relativistic cMPS (RCMPS) is adapted to the relativistic QFTs without requiring to introduce of any additional IR or UV cut-off. We also extend the RCMPS to fermionic systems and theories on a finite circle. |
1410.2090 | Amirpasha Shirazinia Dr. | Amirpasha Shirazinia, Subhrakanti Dey | Power-Constrained Sparse Gaussian Linear Dimensionality Reduction over
Noisy Channels | Accepted for publication in IEEE Transactions on Signal Processing
(16 pages) | null | 10.1109/TSP.2015.2455521 | null | cs.IT math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we investigate power-constrained sensing matrix design in a
sparse Gaussian linear dimensionality reduction framework. Our study is carried
out in a single--terminal setup as well as in a multi--terminal setup
consisting of orthogonal or coherent multiple access channels (MAC). We adopt
the mean square error (MSE) performance criterion for sparse source
reconstruction in a system where source-to-sensor channel(s) and
sensor-to-decoder communication channel(s) are noisy. Our proposed sensing
matrix design procedure relies upon minimizing a lower-bound on the MSE in
single-- and multiple--terminal setups. We propose a three-stage sensing matrix
optimization scheme that combines semi-definite relaxation (SDR) programming, a
low-rank approximation problem and power-rescaling. Under certain conditions,
we derive closed-form solutions to the proposed optimization procedure. Through
numerical experiments, by applying practical sparse reconstruction algorithms,
we show the superiority of the proposed scheme by comparing it with other
relevant methods. This performance improvement is achieved at the price of
higher computational complexity. Hence, in order to address the complexity
burden, we present an equivalent stochastic optimization method to the problem
of interest that can be solved approximately, while still providing a superior
performance over the popular methods.
| [
{
"created": "Wed, 8 Oct 2014 13:01:25 GMT",
"version": "v1"
},
{
"created": "Thu, 9 Oct 2014 13:10:37 GMT",
"version": "v2"
},
{
"created": "Mon, 27 Jul 2015 10:25:35 GMT",
"version": "v3"
}
] | 2015-10-28 | [
[
"Shirazinia",
"Amirpasha",
""
],
[
"Dey",
"Subhrakanti",
""
]
] | In this paper, we investigate power-constrained sensing matrix design in a sparse Gaussian linear dimensionality reduction framework. Our study is carried out in a single--terminal setup as well as in a multi--terminal setup consisting of orthogonal or coherent multiple access channels (MAC). We adopt the mean square error (MSE) performance criterion for sparse source reconstruction in a system where source-to-sensor channel(s) and sensor-to-decoder communication channel(s) are noisy. Our proposed sensing matrix design procedure relies upon minimizing a lower-bound on the MSE in single-- and multiple--terminal setups. We propose a three-stage sensing matrix optimization scheme that combines semi-definite relaxation (SDR) programming, a low-rank approximation problem and power-rescaling. Under certain conditions, we derive closed-form solutions to the proposed optimization procedure. Through numerical experiments, by applying practical sparse reconstruction algorithms, we show the superiority of the proposed scheme by comparing it with other relevant methods. This performance improvement is achieved at the price of higher computational complexity. Hence, in order to address the complexity burden, we present an equivalent stochastic optimization method to the problem of interest that can be solved approximately, while still providing a superior performance over the popular methods. |
0906.3368 | Skenderis Kostas | Joost Hoogeveen and Kostas Skenderis | Decoupling of unphysical states in the minimal pure spinor formalism I | 77 pages (51 pages + appendices), added hyperrefs | JHEP01(2010)041 | 10.1007/JHEP01(2010)041 | NSF-KITP-09-102, ITF-2009-15 | hep-th | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This is the first of a series of two papers where decoupling of unphysical
states in the minimal pure spinor formalism is investigated. The multi-loop
amplitude prescription for the minimal pure spinor superstring formulated in
hep-th/0406055 involves the insertion of picture changing operators in the path
integral. These operators are BRST closed in a distributional sense and depend
on a number of constant tensors. One can trace the origin of these insertions
to gauge fixing, so the amplitudes are formally independent of the constant
tensors. We show however by explicit tree-level and one-loop computations that
the picture changing operators are not BRST closed inside correlators and the
amplitudes do depend on these constant tensors. This is due to the fact that
the gauge fixing condition implicit in the existing minimal amplitude
prescription is singular and this can lead to Lorentz violation and
non-decoupling of BRST exact states. As discussed in hep-th/0406055, a
manifestly Lorentz invariant prescription can be obtained by integrating over
the constant tensors and in the sequel to this paper, it is shown that when one
includes these integrations unphysical states do decouple to all orders despite
the fact that the PCO's are not BRST closed inside correlators.
| [
{
"created": "Thu, 18 Jun 2009 08:36:32 GMT",
"version": "v1"
},
{
"created": "Fri, 19 Jun 2009 11:42:15 GMT",
"version": "v2"
}
] | 2010-01-16 | [
[
"Hoogeveen",
"Joost",
""
],
[
"Skenderis",
"Kostas",
""
]
] | This is the first of a series of two papers where decoupling of unphysical states in the minimal pure spinor formalism is investigated. The multi-loop amplitude prescription for the minimal pure spinor superstring formulated in hep-th/0406055 involves the insertion of picture changing operators in the path integral. These operators are BRST closed in a distributional sense and depend on a number of constant tensors. One can trace the origin of these insertions to gauge fixing, so the amplitudes are formally independent of the constant tensors. We show however by explicit tree-level and one-loop computations that the picture changing operators are not BRST closed inside correlators and the amplitudes do depend on these constant tensors. This is due to the fact that the gauge fixing condition implicit in the existing minimal amplitude prescription is singular and this can lead to Lorentz violation and non-decoupling of BRST exact states. As discussed in hep-th/0406055, a manifestly Lorentz invariant prescription can be obtained by integrating over the constant tensors and in the sequel to this paper, it is shown that when one includes these integrations unphysical states do decouple to all orders despite the fact that the PCO's are not BRST closed inside correlators. |
1210.4211 | Wei Lu | Wei Lu, Laks V.S. Lakshmanan | Profit Maximization over Social Networks | 19 pages, 8 figures. An abbreviated version appears in 2012 IEEE
International Conference on Data Mining (ICDM'12). The second version
includes some minor fixes | null | 10.1109/ICDM.2012.145 | null | cs.SI cs.GT physics.soc-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Influence maximization is the problem of finding a set of influential users
in a social network such that the expected spread of influence under a certain
propagation model is maximized. Much of the previous work has neglected the
important distinction between social influence and actual product adoption.
However, as recognized in the management science literature, an individual who
gets influenced by social acquaintances may not necessarily adopt a product (or
technology), due, e.g., to monetary concerns. In this work, we distinguish
between influence and adoption by explicitly modeling the states of being
influenced and of adopting a product. We extend the classical Linear Threshold
(LT) model to incorporate prices and valuations, and factor them into users'
decision-making process of adopting a product. We show that the expected profit
function under our proposed model maintains submodularity under certain
conditions, but no longer exhibits monotonicity, unlike the expected influence
spread function. To maximize the expected profit under our extended LT model,
we employ an unbudgeted greedy framework to propose three profit maximization
algorithms. The results of our detailed experimental study on three real-world
datasets demonstrate that of the three algorithms, \textsf{PAGE}, which assigns
prices dynamically based on the profit potential of each candidate seed, has
the best performance both in the expected profit achieved and in running time.
| [
{
"created": "Mon, 15 Oct 2012 22:32:37 GMT",
"version": "v1"
},
{
"created": "Wed, 5 Jun 2013 16:41:04 GMT",
"version": "v2"
}
] | 2016-11-18 | [
[
"Lu",
"Wei",
""
],
[
"Lakshmanan",
"Laks V. S.",
""
]
] | Influence maximization is the problem of finding a set of influential users in a social network such that the expected spread of influence under a certain propagation model is maximized. Much of the previous work has neglected the important distinction between social influence and actual product adoption. However, as recognized in the management science literature, an individual who gets influenced by social acquaintances may not necessarily adopt a product (or technology), due, e.g., to monetary concerns. In this work, we distinguish between influence and adoption by explicitly modeling the states of being influenced and of adopting a product. We extend the classical Linear Threshold (LT) model to incorporate prices and valuations, and factor them into users' decision-making process of adopting a product. We show that the expected profit function under our proposed model maintains submodularity under certain conditions, but no longer exhibits monotonicity, unlike the expected influence spread function. To maximize the expected profit under our extended LT model, we employ an unbudgeted greedy framework to propose three profit maximization algorithms. The results of our detailed experimental study on three real-world datasets demonstrate that of the three algorithms, \textsf{PAGE}, which assigns prices dynamically based on the profit potential of each candidate seed, has the best performance both in the expected profit achieved and in running time. |
0710.5051 | Nikolaos Tetradis | G. Dvali, H. B. Nielsen, N. Tetradis | Localization of Gauge Fields and Monopole Tunnelling | 11 pages, 3 figures, improvements in the presentation, version to
appear in Physical Review D | Phys.Rev.D77:085005,2008 | 10.1103/PhysRevD.77.085005 | null | hep-th hep-ph | null | We study the dynamical localization of a massless gauge field on a
lower-dimensional surface (2-brane). In flat space, the necessary and
sufficient condition for this phenomenon is the existence of confinement in the
bulk. The resulting configuration is equivalent to a dual Josephson junction.
This duality leads to an interesting puzzle, as it implies that a localized
massless theory, even in the Abelian case, must become confining at
exponentially large distances. Through the use of topological arguments we
clarify the physics behind this large-distance confinement and identify the
instantons of the brane world-volume theory that are responsible for its
appearance. We show that they correspond to the (condensed) bulk magnetic
charges (monopoles), that occasionally tunnel through the brane and induce weak
confinement of the brane theory. We consider the possible generalization of
this effect to higher dimensions and discuss phenomenological bounds on the
confinement of electric charges at exponentially large distances within our
Universe.
| [
{
"created": "Fri, 26 Oct 2007 10:51:34 GMT",
"version": "v1"
},
{
"created": "Thu, 28 Feb 2008 11:29:54 GMT",
"version": "v2"
}
] | 2008-11-26 | [
[
"Dvali",
"G.",
""
],
[
"Nielsen",
"H. B.",
""
],
[
"Tetradis",
"N.",
""
]
] | We study the dynamical localization of a massless gauge field on a lower-dimensional surface (2-brane). In flat space, the necessary and sufficient condition for this phenomenon is the existence of confinement in the bulk. The resulting configuration is equivalent to a dual Josephson junction. This duality leads to an interesting puzzle, as it implies that a localized massless theory, even in the Abelian case, must become confining at exponentially large distances. Through the use of topological arguments we clarify the physics behind this large-distance confinement and identify the instantons of the brane world-volume theory that are responsible for its appearance. We show that they correspond to the (condensed) bulk magnetic charges (monopoles), that occasionally tunnel through the brane and induce weak confinement of the brane theory. We consider the possible generalization of this effect to higher dimensions and discuss phenomenological bounds on the confinement of electric charges at exponentially large distances within our Universe. |
1012.1909 | Manar Mohaisen | Manar Mohaisen, KyungHi Chang | On Transmit Antenna Selection for Multiuser MIMO Systems with Dirty
Paper Coding | 5 pages, 6 figures, 1 table, [The 20th Personal, Indoor and Mobile
Radio Communications Symposium 2009 (PIMRC-09)] | The 20th Personal, Indoor and Mobile Radio Communications
Symposium 2009 (PIMRC-09) | null | null | cs.IT math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we address the transmit antenna selection in multi-user MIMO
systems with precoding. The optimum and reduced complexity sub-optimum antenna
selection algorithms are introduced. QR-decomposition (QRD) based antenna
selection is investigated and the reason behind its sub-optimality is
analytically derived. We introduce the conventional QRD-based algorithm and
propose an efficient QRD-based transmit antenna scheme (maxR) that is both
implementation and performance efficient. Moreover, we derive explicit formulae
for the computational complexities of the aforementioned algorithms. Simulation
results and analysis demonstrate that the proposed maxR algorithm requires only
1% of the computational efforts required by the optimal algorithm for a
degradation of 1dB and 0.1dB in the case of linear zero-forcing and
Tomlinson-Harashima precoding schemes, respectively.
| [
{
"created": "Thu, 9 Dec 2010 02:03:17 GMT",
"version": "v1"
}
] | 2010-12-10 | [
[
"Mohaisen",
"Manar",
""
],
[
"Chang",
"KyungHi",
""
]
] | In this paper, we address the transmit antenna selection in multi-user MIMO systems with precoding. The optimum and reduced complexity sub-optimum antenna selection algorithms are introduced. QR-decomposition (QRD) based antenna selection is investigated and the reason behind its sub-optimality is analytically derived. We introduce the conventional QRD-based algorithm and propose an efficient QRD-based transmit antenna scheme (maxR) that is both implementation and performance efficient. Moreover, we derive explicit formulae for the computational complexities of the aforementioned algorithms. Simulation results and analysis demonstrate that the proposed maxR algorithm requires only 1% of the computational efforts required by the optimal algorithm for a degradation of 1dB and 0.1dB in the case of linear zero-forcing and Tomlinson-Harashima precoding schemes, respectively. |
1902.06066 | Varshaneya V | Varshaneya V, Balasubramanian S and Darshan Gera | RES-SE-NET: Boosting Performance of Resnets by Enhancing
Bridge-connections | null | null | null | null | cs.LG cs.CV stat.ML | http://creativecommons.org/licenses/by-sa/4.0/ | One of the ways to train deep neural networks effectively is to use residual
connections. Residual connections can be classified as being either identity
connections or bridge-connections with a reshaping convolution. Empirical
observations on CIFAR-10 and CIFAR-100 datasets using a baseline Resnet model,
with bridge-connections removed, have shown a significant reduction in
accuracy. This reduction is due to lack of contribution, in the form of feature
maps, by the bridge-connections. Hence bridge-connections are vital for Resnet.
However, all feature maps in the bridge-connections are considered to be
equally important. In this work, an upgraded architecture "Res-SE-Net" is
proposed to further strengthen the contribution from the bridge-connections by
quantifying the importance of each feature map and weighting them accordingly
using Squeeze-and-Excitation (SE) block. It is demonstrated that Res-SE-Net
generalizes much better than Resnet and SE-Resnet on the benchmark CIFAR-10 and
CIFAR-100 datasets.
| [
{
"created": "Sat, 16 Feb 2019 08:25:16 GMT",
"version": "v1"
}
] | 2019-02-19 | [
[
"V",
"Varshaneya",
""
],
[
"S",
"Balasubramanian",
""
],
[
"Gera",
"Darshan",
""
]
] | One of the ways to train deep neural networks effectively is to use residual connections. Residual connections can be classified as being either identity connections or bridge-connections with a reshaping convolution. Empirical observations on CIFAR-10 and CIFAR-100 datasets using a baseline Resnet model, with bridge-connections removed, have shown a significant reduction in accuracy. This reduction is due to lack of contribution, in the form of feature maps, by the bridge-connections. Hence bridge-connections are vital for Resnet. However, all feature maps in the bridge-connections are considered to be equally important. In this work, an upgraded architecture "Res-SE-Net" is proposed to further strengthen the contribution from the bridge-connections by quantifying the importance of each feature map and weighting them accordingly using Squeeze-and-Excitation (SE) block. It is demonstrated that Res-SE-Net generalizes much better than Resnet and SE-Resnet on the benchmark CIFAR-10 and CIFAR-100 datasets. |
1501.02855 | Luis Sentis | Donghyun Kim, Ye Zhao, Gray Thomas, and Luis Sentis | Assessing Whole-Body Operational Space Control in a Point-Foot Series
Elastic Biped: Balance on Split Terrain and Undirected Walking | 17 pages, 9 figures, 4 tables | null | null | null | cs.RO cs.SY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper we present advancements in control and trajectory generation
for agile behavior in bipedal robots. We demonstrate that Whole-Body
Operational Space Control (WBOSC), developed a few years ago, is well suited
for achieving two types of agile behaviors, namely, balancing on a high pitch
split terrain and achieving undirected walking on flat terrain. The work
presented here is the first implementation of WBOSC on a biped robot, and more
specifically a biped robot with series elastic actuators. We present and
analyze a new algorithm that dynamically balances point foot robots by choosing
footstep placements. Dealing with the naturally unstable dynamics of these type
of systems is a difficult problem that requires both the controller and the
trajectory generation algorithm to operate quickly and efficiently. We put
forth a comprehensive development and integration effort: the design and
construction of the biped system and experimental infrastructure, a
customization of WBOSC for the agile behaviors, and new trajectory generation
algorithms. Using this custom built controller, we conduct, for first time, an
experiment in which a biped robot balances in a high pitch split terrain,
demonstrating our ability to precisely regulate internal forces using force
sensing feedback techniques. Finally, we demonstrate the stabilizing
capabilities of our online trajectory generation algorithm in the physics-based
simulator and through physical experiments with a planarized locomotion setup.
| [
{
"created": "Tue, 13 Jan 2015 00:17:39 GMT",
"version": "v1"
}
] | 2015-01-14 | [
[
"Kim",
"Donghyun",
""
],
[
"Zhao",
"Ye",
""
],
[
"Thomas",
"Gray",
""
],
[
"Sentis",
"Luis",
""
]
] | In this paper we present advancements in control and trajectory generation for agile behavior in bipedal robots. We demonstrate that Whole-Body Operational Space Control (WBOSC), developed a few years ago, is well suited for achieving two types of agile behaviors, namely, balancing on a high pitch split terrain and achieving undirected walking on flat terrain. The work presented here is the first implementation of WBOSC on a biped robot, and more specifically a biped robot with series elastic actuators. We present and analyze a new algorithm that dynamically balances point foot robots by choosing footstep placements. Dealing with the naturally unstable dynamics of these type of systems is a difficult problem that requires both the controller and the trajectory generation algorithm to operate quickly and efficiently. We put forth a comprehensive development and integration effort: the design and construction of the biped system and experimental infrastructure, a customization of WBOSC for the agile behaviors, and new trajectory generation algorithms. Using this custom built controller, we conduct, for first time, an experiment in which a biped robot balances in a high pitch split terrain, demonstrating our ability to precisely regulate internal forces using force sensing feedback techniques. Finally, we demonstrate the stabilizing capabilities of our online trajectory generation algorithm in the physics-based simulator and through physical experiments with a planarized locomotion setup. |
1009.5705 | Grenville Croll | Angus Dunn | Spreadsheets - the Good, the Bad and the Downright Ugly | 8 Pages | Proc. European Spreadsheet Risks Int. Grp. (EuSpRIG) 2010 157-164
ISBN 978-1-905404-50-6 | null | null | cs.SE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Spreadsheets are ubiquitous, heavily relied on throughout vast swathes of
finance, commerce, industry, academia and Government. They are also
acknowledged to be extraordinarily and unacceptably prone to error. If these
two points are accepted, it has to follow that their uncontrolled use has the
potential to inflict considerable damage. One approach to controlling such
error should be to define as "good practice" a set of characteristics that a
spreadsheet must possess and as "bad practice" another set that it must avoid.
Defining such characteristics should, in principle, perfectly do-able. However,
being able to say with authority at a definite moment that any particular
spreadsheet complies with these characteristics is very much more difficult.
The author asserts that the use of automated spreadsheet development could
markedly help in ensuring and demonstrating such compliance.
| [
{
"created": "Tue, 28 Sep 2010 21:46:50 GMT",
"version": "v1"
}
] | 2010-09-30 | [
[
"Dunn",
"Angus",
""
]
] | Spreadsheets are ubiquitous, heavily relied on throughout vast swathes of finance, commerce, industry, academia and Government. They are also acknowledged to be extraordinarily and unacceptably prone to error. If these two points are accepted, it has to follow that their uncontrolled use has the potential to inflict considerable damage. One approach to controlling such error should be to define as "good practice" a set of characteristics that a spreadsheet must possess and as "bad practice" another set that it must avoid. Defining such characteristics should, in principle, perfectly do-able. However, being able to say with authority at a definite moment that any particular spreadsheet complies with these characteristics is very much more difficult. The author asserts that the use of automated spreadsheet development could markedly help in ensuring and demonstrating such compliance. |
2207.14417 | Alexandros Evangelidis | Muqsit Azeem, Alexandros Evangelidis, Jan K\v{r}et\'insk\'y, Alexander
Slivinskiy, and Maximilian Weininger | Optimistic and Topological Value Iteration for Simple Stochastic Games | null | null | null | null | cs.GT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | While value iteration (VI) is a standard solution approach to simple
stochastic games (SSGs), it suffered from the lack of a stopping criterion.
Recently, several solutions have appeared, among them also "optimistic" VI
(OVI). However, OVI is applicable only to one-player SSGs with no end
components. We lift these two assumptions, making it available to general SSGs.
Further, we utilize the idea in the context of topological VI, where we provide
an efficient precise solution. In order to compare the new algorithms with the
state of the art, we use not only the standard benchmarks, but we also design a
random generator of SSGs, which can be biased towards various types of models,
aiding in understanding the advantages of different algorithms on SSGs.
| [
{
"created": "Fri, 29 Jul 2022 00:34:47 GMT",
"version": "v1"
}
] | 2022-08-01 | [
[
"Azeem",
"Muqsit",
""
],
[
"Evangelidis",
"Alexandros",
""
],
[
"Křetínský",
"Jan",
""
],
[
"Slivinskiy",
"Alexander",
""
],
[
"Weininger",
"Maximilian",
""
]
] | While value iteration (VI) is a standard solution approach to simple stochastic games (SSGs), it suffered from the lack of a stopping criterion. Recently, several solutions have appeared, among them also "optimistic" VI (OVI). However, OVI is applicable only to one-player SSGs with no end components. We lift these two assumptions, making it available to general SSGs. Further, we utilize the idea in the context of topological VI, where we provide an efficient precise solution. In order to compare the new algorithms with the state of the art, we use not only the standard benchmarks, but we also design a random generator of SSGs, which can be biased towards various types of models, aiding in understanding the advantages of different algorithms on SSGs. |
1208.5350 | Man Yi Yim | Man Yi Yim, Ad Aertsen, Stefan Rotter | Impact of intrinsic biophysical diversity on the activity of spiking
neurons | 4 pages, 5 figures | Phys. Rev. E 87, 032710 (2013) | 10.1103/PhysRevE.87.032710 | null | q-bio.NC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We study the effect of intrinsic heterogeneity on the activity of a
population of leaky integrate-and-fire neurons. By rescaling the dynamical
equation, we derive mathematical relations between multiple neuronal parameters
and a fluctuating input noise. To this end, common input to heterogeneous
neurons is conceived as an identical noise with neuron-specific mean and
variance. As a consequence, the neuronal output rates can differ considerably,
and their relative spike timing becomes desynchronized. This theory can
quantitatively explain some recent experimental findings.
| [
{
"created": "Mon, 27 Aug 2012 10:06:12 GMT",
"version": "v1"
},
{
"created": "Wed, 16 Jan 2013 10:27:55 GMT",
"version": "v2"
},
{
"created": "Thu, 31 Jan 2013 14:42:54 GMT",
"version": "v3"
},
{
"created": "Tue, 19 Feb 2013 17:08:00 GMT",
"version": "v4"
}
] | 2013-08-21 | [
[
"Yim",
"Man Yi",
""
],
[
"Aertsen",
"Ad",
""
],
[
"Rotter",
"Stefan",
""
]
] | We study the effect of intrinsic heterogeneity on the activity of a population of leaky integrate-and-fire neurons. By rescaling the dynamical equation, we derive mathematical relations between multiple neuronal parameters and a fluctuating input noise. To this end, common input to heterogeneous neurons is conceived as an identical noise with neuron-specific mean and variance. As a consequence, the neuronal output rates can differ considerably, and their relative spike timing becomes desynchronized. This theory can quantitatively explain some recent experimental findings. |
cs/0703053 | Nicolas Lomenie | Guray Erus (CRIP5), Nicolas Lom\'enie (CRIP5) | Extraction of cartographic objects in high resolution satellite images
for object model generation | null | 4th Workshop on pattern Recognition in Remote Sensing in
conjunction with ICPR2006 (08/2006) 00-00 | null | null | cs.CV | null | The aim of this study is to detect man-made cartographic objects in
high-resolution satellite images. New generation satellites offer a sub-metric
spatial resolution, in which it is possible (and necessary) to develop methods
at object level rather than at pixel level, and to exploit structural features
of objects. With this aim, a method to generate structural object models from
manually segmented images has been developed. To generate the model from
non-segmented images, extraction of the objects from the sample images is
required. A hybrid method of extraction (both in terms of input sources and
segmentation algorithms) is proposed: A region based segmentation is applied on
a 10 meter resolution multi-spectral image. The result is used as marker in a
"marker-controlled watershed method using edges" on a 2.5 meter resolution
panchromatic image. Very promising results have been obtained even on images
where the limits of the target objects are not apparent.
| [
{
"created": "Mon, 12 Mar 2007 15:57:23 GMT",
"version": "v1"
}
] | 2016-08-14 | [
[
"Erus",
"Guray",
"",
"CRIP5"
],
[
"Loménie",
"Nicolas",
"",
"CRIP5"
]
] | The aim of this study is to detect man-made cartographic objects in high-resolution satellite images. New generation satellites offer a sub-metric spatial resolution, in which it is possible (and necessary) to develop methods at object level rather than at pixel level, and to exploit structural features of objects. With this aim, a method to generate structural object models from manually segmented images has been developed. To generate the model from non-segmented images, extraction of the objects from the sample images is required. A hybrid method of extraction (both in terms of input sources and segmentation algorithms) is proposed: A region based segmentation is applied on a 10 meter resolution multi-spectral image. The result is used as marker in a "marker-controlled watershed method using edges" on a 2.5 meter resolution panchromatic image. Very promising results have been obtained even on images where the limits of the target objects are not apparent. |
2208.14153 | Yuhang Liu | Yuhang Liu, Zhen Zhang, Dong Gong, Mingming Gong, Biwei Huang, Anton
van den Hengel, Kun Zhang, Javen Qinfeng Shi | Identifying Weight-Variant Latent Causal Models | null | null | null | null | cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The task of causal representation learning aims to uncover latent
higher-level causal representations that affect lower-level observations.
Identifying true latent causal representations from observed data, while
allowing instantaneous causal relations among latent variables, remains a
challenge, however. To this end, we start from the analysis of three intrinsic
properties in identifying latent space from observations: transitivity,
permutation indeterminacy, and scaling indeterminacy. We find that transitivity
acts as a key role in impeding the identifiability of latent causal
representations. To address the unidentifiable issue due to transitivity, we
introduce a novel identifiability condition where the underlying latent causal
model satisfies a linear-Gaussian model, in which the causal coefficients and
the distribution of Gaussian noise are modulated by an additional observed
variable. Under some mild assumptions, we can show that the latent causal
representations can be identified up to trivial permutation and scaling.
Furthermore, based on this theoretical result, we propose a novel method,
termed Structural caUsAl Variational autoEncoder, which directly learns latent
causal representations and causal relationships among them, together with the
mapping from the latent causal variables to the observed ones. We show that the
proposed method learns the true parameters asymptotically. Experimental results
on synthetic and real data demonstrate the identifiability and consistency
results and the efficacy of the proposed method in learning latent causal
representations.
| [
{
"created": "Tue, 30 Aug 2022 11:12:59 GMT",
"version": "v1"
},
{
"created": "Fri, 30 Sep 2022 07:14:54 GMT",
"version": "v2"
},
{
"created": "Wed, 16 Nov 2022 10:36:08 GMT",
"version": "v3"
},
{
"created": "Tue, 6 Dec 2022 06:15:03 GMT",
"version": "v4"
},
{
"cr... | 2023-02-21 | [
[
"Liu",
"Yuhang",
""
],
[
"Zhang",
"Zhen",
""
],
[
"Gong",
"Dong",
""
],
[
"Gong",
"Mingming",
""
],
[
"Huang",
"Biwei",
""
],
[
"Hengel",
"Anton van den",
""
],
[
"Zhang",
"Kun",
""
],
[
"Shi",
... | The task of causal representation learning aims to uncover latent higher-level causal representations that affect lower-level observations. Identifying true latent causal representations from observed data, while allowing instantaneous causal relations among latent variables, remains a challenge, however. To this end, we start from the analysis of three intrinsic properties in identifying latent space from observations: transitivity, permutation indeterminacy, and scaling indeterminacy. We find that transitivity acts as a key role in impeding the identifiability of latent causal representations. To address the unidentifiable issue due to transitivity, we introduce a novel identifiability condition where the underlying latent causal model satisfies a linear-Gaussian model, in which the causal coefficients and the distribution of Gaussian noise are modulated by an additional observed variable. Under some mild assumptions, we can show that the latent causal representations can be identified up to trivial permutation and scaling. Furthermore, based on this theoretical result, we propose a novel method, termed Structural caUsAl Variational autoEncoder, which directly learns latent causal representations and causal relationships among them, together with the mapping from the latent causal variables to the observed ones. We show that the proposed method learns the true parameters asymptotically. Experimental results on synthetic and real data demonstrate the identifiability and consistency results and the efficacy of the proposed method in learning latent causal representations. |
0806.4959 | Troels Harmark | Gianluca Grignani, Troels Harmark and Marta Orselli | The SU(2) x SU(2) sector in the string dual of N=6 superconformal
Chern-Simons theory | 19 pages.: Typos fixed, Sec. 6 improved | Nucl.Phys.B810:115-134,2009 | 10.1016/j.nuclphysb.2008.10.019 | null | hep-th | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We examine the string dual of the recently constructed $\mathcal{N}=6$
superconformal Chern-Simons theory of Aharony, Bergman, Jafferis and Maldacena
(ABJM theory). We focus in particular on the $SU(2)\times SU(2)$ sector. We
find a sigma-model limit in which the resulting sigma-model is two
Landau-Lifshitz models added together. We consider a Penrose limit for which we
can approach the $SU(2)\times SU(2)$ sector. Finally, we find a new Giant
Magnon solution in the $SU(2)\times SU(2)$ sector corresponding to one magnon
in each $SU(2)$. We put these results together to find the full magnon
dispersion relation and we compare this to recently found results for ABJM
theory at weak coupling.
| [
{
"created": "Mon, 30 Jun 2008 17:45:13 GMT",
"version": "v1"
},
{
"created": "Tue, 1 Jul 2008 17:26:39 GMT",
"version": "v2"
},
{
"created": "Mon, 14 Jul 2008 14:18:33 GMT",
"version": "v3"
},
{
"created": "Thu, 17 Jul 2008 16:54:23 GMT",
"version": "v4"
}
] | 2017-09-07 | [
[
"Grignani",
"Gianluca",
""
],
[
"Harmark",
"Troels",
""
],
[
"Orselli",
"Marta",
""
]
] | We examine the string dual of the recently constructed $\mathcal{N}=6$ superconformal Chern-Simons theory of Aharony, Bergman, Jafferis and Maldacena (ABJM theory). We focus in particular on the $SU(2)\times SU(2)$ sector. We find a sigma-model limit in which the resulting sigma-model is two Landau-Lifshitz models added together. We consider a Penrose limit for which we can approach the $SU(2)\times SU(2)$ sector. Finally, we find a new Giant Magnon solution in the $SU(2)\times SU(2)$ sector corresponding to one magnon in each $SU(2)$. We put these results together to find the full magnon dispersion relation and we compare this to recently found results for ABJM theory at weak coupling. |
1908.10004 | Jia Tian | Jia Tian, Jue Hou, Bin Chen | Asymmetric $\lambda$-deformed cosets | 23 pages ver2 references added | null | 10.1016/j.nuclphysb.2020.114944 | null | hep-th | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We study the integrable asymmetric $\lambda$-deformations of the
$SO(n+1)/SO(n)$ coset models, following the prescription proposed in
\cite{AsyLambda}. We construct all corresponding deformed geometries in an
inductive way. Remarkably we find a $Z_2$ transformation which maps the
asymmetric $\lambda$--deformed models to the symmetric $\lambda$--deformed
models.
| [
{
"created": "Tue, 27 Aug 2019 03:17:17 GMT",
"version": "v1"
},
{
"created": "Tue, 3 Sep 2019 05:37:36 GMT",
"version": "v2"
}
] | 2020-03-18 | [
[
"Tian",
"Jia",
""
],
[
"Hou",
"Jue",
""
],
[
"Chen",
"Bin",
""
]
] | We study the integrable asymmetric $\lambda$-deformations of the $SO(n+1)/SO(n)$ coset models, following the prescription proposed in \cite{AsyLambda}. We construct all corresponding deformed geometries in an inductive way. Remarkably we find a $Z_2$ transformation which maps the asymmetric $\lambda$--deformed models to the symmetric $\lambda$--deformed models. |
1202.5255 | Avihay Kadosh | Avihay Kadosh, Aharon Davidson and Elisabetta Pallante | Slinky evolution of domain wall brane cosmology | 24 pages, 4 figures, extended discussion of slinky evolution, minor
revisions, conclusions unchanged | Phys. Rev. D 86, 124015 (2012) | 10.1103/PhysRevD.86.124015 | null | hep-th gr-qc hep-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Invoking an initial symmetry between the time $ t $ and some extra spatial
dimension $ y $, we discuss a novel scenario where the dynamical formation of
the 4-dim brane and its cosmological evolution are induced simultaneously by a
common $ t<->y $ symmetry breaking mechanism. The local maximum of the
underlying scalar potential is mapped onto a 'watershed' curve in the $ (t,y) $
plane; the direction tangent to this curve is identified as the cosmic time,
whereas the perpendicular direction serves locally as the extra spatial
dimension. Special attention is devoted to the so-called slinky configurations,
whose brane cosmology is characterized by a decaying cosmological constant
along the watershed curve. Such a slinky solution is first constructed within a
simplified case where the watershed is constrained by $ y = 0 $. The physical
requirements for a slinky configuration to generate a realistic model of
cosmological evolution are then discussed in a more elaborated framework.
| [
{
"created": "Thu, 23 Feb 2012 18:19:06 GMT",
"version": "v1"
},
{
"created": "Thu, 9 Aug 2012 10:18:03 GMT",
"version": "v2"
}
] | 2012-12-06 | [
[
"Kadosh",
"Avihay",
""
],
[
"Davidson",
"Aharon",
""
],
[
"Pallante",
"Elisabetta",
""
]
] | Invoking an initial symmetry between the time $ t $ and some extra spatial dimension $ y $, we discuss a novel scenario where the dynamical formation of the 4-dim brane and its cosmological evolution are induced simultaneously by a common $ t<->y $ symmetry breaking mechanism. The local maximum of the underlying scalar potential is mapped onto a 'watershed' curve in the $ (t,y) $ plane; the direction tangent to this curve is identified as the cosmic time, whereas the perpendicular direction serves locally as the extra spatial dimension. Special attention is devoted to the so-called slinky configurations, whose brane cosmology is characterized by a decaying cosmological constant along the watershed curve. Such a slinky solution is first constructed within a simplified case where the watershed is constrained by $ y = 0 $. The physical requirements for a slinky configuration to generate a realistic model of cosmological evolution are then discussed in a more elaborated framework. |
2007.01520 | Alexander Mitchell Mr | Alexander L. Mitchell, Martin Engelcke, Oiwi Parker Jones, David
Surovik, Siddhant Gangapurwala, Oliwier Melon, Ioannis Havoutis, and Ingmar
Posner | First Steps: Latent-Space Control with Semantic Constraints for
Quadruped Locomotion | 8 pages, 7 figures, accepted at IROS 2020 | null | null | null | cs.RO cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Traditional approaches to quadruped control frequently employ simplified,
hand-derived models. This significantly reduces the capability of the robot
since its effective kinematic range is curtailed. In addition, kinodynamic
constraints are often non-differentiable and difficult to implement in an
optimisation approach. In this work, these challenges are addressed by framing
quadruped control as optimisation in a structured latent space. A deep
generative model captures a statistical representation of feasible joint
configurations, whilst complex dynamic and terminal constraints are expressed
via high-level, semantic indicators and represented by learned classifiers
operating upon the latent space. As a consequence, complex constraints are
rendered differentiable and evaluated an order of magnitude faster than
analytical approaches. We validate the feasibility of locomotion trajectories
optimised using our approach both in simulation and on a real-world ANYmal
quadruped. Our results demonstrate that this approach is capable of generating
smooth and realisable trajectories. To the best of our knowledge, this is the
first time latent space control has been successfully applied to a complex,
real robot platform.
| [
{
"created": "Fri, 3 Jul 2020 07:04:18 GMT",
"version": "v1"
},
{
"created": "Fri, 20 Nov 2020 16:31:46 GMT",
"version": "v2"
}
] | 2020-11-23 | [
[
"Mitchell",
"Alexander L.",
""
],
[
"Engelcke",
"Martin",
""
],
[
"Jones",
"Oiwi Parker",
""
],
[
"Surovik",
"David",
""
],
[
"Gangapurwala",
"Siddhant",
""
],
[
"Melon",
"Oliwier",
""
],
[
"Havoutis",
"Ioannis... | Traditional approaches to quadruped control frequently employ simplified, hand-derived models. This significantly reduces the capability of the robot since its effective kinematic range is curtailed. In addition, kinodynamic constraints are often non-differentiable and difficult to implement in an optimisation approach. In this work, these challenges are addressed by framing quadruped control as optimisation in a structured latent space. A deep generative model captures a statistical representation of feasible joint configurations, whilst complex dynamic and terminal constraints are expressed via high-level, semantic indicators and represented by learned classifiers operating upon the latent space. As a consequence, complex constraints are rendered differentiable and evaluated an order of magnitude faster than analytical approaches. We validate the feasibility of locomotion trajectories optimised using our approach both in simulation and on a real-world ANYmal quadruped. Our results demonstrate that this approach is capable of generating smooth and realisable trajectories. To the best of our knowledge, this is the first time latent space control has been successfully applied to a complex, real robot platform. |
2203.00770 | Kan Yu | Ming Zhan (1), Zhibo Pang (2 and 3), Dacfey Dzung (2), Kan Yu (4),
Ming Xiao (3) ((1) Southwest University, (2) ABB Corporate Research, (3) KTH
Royal Institute of Technology, (4) La Trobe University) | Short-Packet Interleaver against Impulse Interference in Practical
Industrial Environments | 14 pages, 12 figures, submitted to IEEE Transactions on Wireless
Communications | null | null | null | cs.IT eess.SP math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The most common cause of transmission failure in Wireless High Performance
(WirelessHP) target industry environments is impulse interference. As
interleavers are commonly used to improve the reliability on the Orthogonal
Frequency Division Multiplexing (OFDM) symbol level for long packet
transmission, this paper considers the feasibility of applying short-packet bit
interleaving to enhance the impulse/burst interference resisting capability on
both OFDM symbol and frame level. Using the Universal Software Radio
Peripherals (USRP) and PC hardware platform, the Packet Error Rate (PER)
performance of interleaved coded short-packet transmission with Convolutional
Codes (CC), Reed-Solomon codes (RS) and RS+CC concatenated codes are tested and
analyzed. Applying the IEEE 1613 standard for impulse interference generation,
extensive PER tests of CC(1=2) and RS(31; 21)+CC(1=2) concatenated codes are
performed. With practical experiments, we prove the effectiveness of bit in
terleaved coded short-packet transmission in real factory environments. We also
investigate how PER performance depends on the interleavers, codes and impulse
interference power and frequency.
| [
{
"created": "Tue, 1 Mar 2022 22:24:37 GMT",
"version": "v1"
}
] | 2022-03-03 | [
[
"Zhan",
"Ming",
"",
"2 and 3"
],
[
"Pang",
"Zhibo",
"",
"2 and 3"
],
[
"Dzung",
"Dacfey",
""
],
[
"Yu",
"Kan",
""
],
[
"Xiao",
"Ming",
""
]
] | The most common cause of transmission failure in Wireless High Performance (WirelessHP) target industry environments is impulse interference. As interleavers are commonly used to improve the reliability on the Orthogonal Frequency Division Multiplexing (OFDM) symbol level for long packet transmission, this paper considers the feasibility of applying short-packet bit interleaving to enhance the impulse/burst interference resisting capability on both OFDM symbol and frame level. Using the Universal Software Radio Peripherals (USRP) and PC hardware platform, the Packet Error Rate (PER) performance of interleaved coded short-packet transmission with Convolutional Codes (CC), Reed-Solomon codes (RS) and RS+CC concatenated codes are tested and analyzed. Applying the IEEE 1613 standard for impulse interference generation, extensive PER tests of CC(1=2) and RS(31; 21)+CC(1=2) concatenated codes are performed. With practical experiments, we prove the effectiveness of bit in terleaved coded short-packet transmission in real factory environments. We also investigate how PER performance depends on the interleavers, codes and impulse interference power and frequency. |
0802.1520 | Ophir Flomenbom | O. Flomenbom, and R. J. Silbey | Toolbox for analyzing finite two-state trajectories | null | Phys. Rev. E 78, 066105 (2008) | 10.1103/PhysRevE.78.066105 | null | q-bio.QM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In many experiments, the aim is to deduce an underlying multi-substate on-off
kinetic scheme (KS) from the statistical properties of a two-state trajectory.
However, the mapping of a KS into a two-state trajectory leads to the loss of
information about the KS, and so, in many cases, more than one KS can be
associated with the data. We recently showed that the optimal way to solve this
problem is to use canonical forms of reduced dimensions (RD). RD forms are
on-off networks with connections only between substates of different states,
where the connections can have non-exponential waiting time probability density
functions (WT-PDFs). In theory, only a single RD form can be associated with
the data. To utilize RD forms in the analysis of the data, a RD form should be
associated with the data. Here, we give a toolbox for building a RD form from a
finite two-state trajectory. The methods in the toolbox are based on known
statistical methods in data analysis, combined with statistical methods and
numerical algorithms designed specifically for the current problem. Our toolbox
is self-contained - it builds a mechanism based only on the information it
extracts from the data, and its implementation on the data is fast (analyzing a
10^6 cycle trajectory from a thirty-parameter mechanism takes a couple of hours
on a PC with a 2.66 GHz processor). The toolbox is automated and is freely
available for academic research upon electronic request.
| [
{
"created": "Mon, 11 Feb 2008 20:07:26 GMT",
"version": "v1"
},
{
"created": "Wed, 8 Oct 2008 23:31:58 GMT",
"version": "v2"
},
{
"created": "Thu, 25 Dec 2008 03:07:42 GMT",
"version": "v3"
}
] | 2010-08-16 | [
[
"Flomenbom",
"O.",
""
],
[
"Silbey",
"R. J.",
""
]
] | In many experiments, the aim is to deduce an underlying multi-substate on-off kinetic scheme (KS) from the statistical properties of a two-state trajectory. However, the mapping of a KS into a two-state trajectory leads to the loss of information about the KS, and so, in many cases, more than one KS can be associated with the data. We recently showed that the optimal way to solve this problem is to use canonical forms of reduced dimensions (RD). RD forms are on-off networks with connections only between substates of different states, where the connections can have non-exponential waiting time probability density functions (WT-PDFs). In theory, only a single RD form can be associated with the data. To utilize RD forms in the analysis of the data, a RD form should be associated with the data. Here, we give a toolbox for building a RD form from a finite two-state trajectory. The methods in the toolbox are based on known statistical methods in data analysis, combined with statistical methods and numerical algorithms designed specifically for the current problem. Our toolbox is self-contained - it builds a mechanism based only on the information it extracts from the data, and its implementation on the data is fast (analyzing a 10^6 cycle trajectory from a thirty-parameter mechanism takes a couple of hours on a PC with a 2.66 GHz processor). The toolbox is automated and is freely available for academic research upon electronic request. |
2102.11448 | DiJia Su | DiJia Su, Jason D. Lee, John M. Mulvey, H. Vincent Poor | MUSBO: Model-based Uncertainty Regularized and Sample Efficient Batch
Optimization for Deployment Constrained Reinforcement Learning | null | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In many contemporary applications such as healthcare, finance, robotics, and
recommendation systems, continuous deployment of new policies for data
collection and online learning is either cost ineffective or impractical. We
consider a setting that lies between pure offline reinforcement learning (RL)
and pure online RL called deployment constrained RL in which the number of
policy deployments for data sampling is limited. To solve this challenging
task, we propose a new algorithmic learning framework called Model-based
Uncertainty regularized and Sample Efficient Batch Optimization (MUSBO). Our
framework discovers novel and high quality samples for each deployment to
enable efficient data collection. During each offline training session, we
bootstrap the policy update by quantifying the amount of uncertainty within our
collected data. In the high support region (low uncertainty), we encourage our
policy by taking an aggressive update. In the low support region (high
uncertainty) when the policy bootstraps into the out-of-distribution region, we
downweight it by our estimated uncertainty quantification. Experimental results
show that MUSBO achieves state-of-the-art performance in the deployment
constrained RL setting.
| [
{
"created": "Tue, 23 Feb 2021 01:30:55 GMT",
"version": "v1"
},
{
"created": "Thu, 3 Jun 2021 23:59:52 GMT",
"version": "v2"
}
] | 2021-06-07 | [
[
"Su",
"DiJia",
""
],
[
"Lee",
"Jason D.",
""
],
[
"Mulvey",
"John M.",
""
],
[
"Poor",
"H. Vincent",
""
]
] | In many contemporary applications such as healthcare, finance, robotics, and recommendation systems, continuous deployment of new policies for data collection and online learning is either cost ineffective or impractical. We consider a setting that lies between pure offline reinforcement learning (RL) and pure online RL called deployment constrained RL in which the number of policy deployments for data sampling is limited. To solve this challenging task, we propose a new algorithmic learning framework called Model-based Uncertainty regularized and Sample Efficient Batch Optimization (MUSBO). Our framework discovers novel and high quality samples for each deployment to enable efficient data collection. During each offline training session, we bootstrap the policy update by quantifying the amount of uncertainty within our collected data. In the high support region (low uncertainty), we encourage our policy by taking an aggressive update. In the low support region (high uncertainty) when the policy bootstraps into the out-of-distribution region, we downweight it by our estimated uncertainty quantification. Experimental results show that MUSBO achieves state-of-the-art performance in the deployment constrained RL setting. |
2005.04864 | Xingyu Chen | Xingyu Chen and Zijie Liu | The Fairness of Leximin in Allocation of Indivisible Chores | null | null | null | null | cs.GT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The leximin solution -- which selects an allocation that maximizes the
minimum utility, then the second minimum utility, and so forth -- is known to
provide EFX (envy-free up to any good) fairness guarantee in some contexts when
allocating indivisible goods. However, it remains unknown how fair the leximin
solution is when used to allocate indivisible chores. In this paper, we
demonstrate that the leximin solution can be modified to also provide
compelling fairness guarantees for the allocation of indivisible chores. First,
we generalize the definition of the leximin solution. Then, we show that the
leximin solution finds a PROP1 (proportional up to one good) and PO
(Pareto-optimal) allocation for 3 or 4 agents in the context of chores
allocation with additive distinct valuations. Additionally, we prove that the
leximin solution is EFX for combinations of goods and chores for agents with
general but identical valuations.
| [
{
"created": "Mon, 11 May 2020 05:15:43 GMT",
"version": "v1"
}
] | 2020-05-12 | [
[
"Chen",
"Xingyu",
""
],
[
"Liu",
"Zijie",
""
]
] | The leximin solution -- which selects an allocation that maximizes the minimum utility, then the second minimum utility, and so forth -- is known to provide EFX (envy-free up to any good) fairness guarantee in some contexts when allocating indivisible goods. However, it remains unknown how fair the leximin solution is when used to allocate indivisible chores. In this paper, we demonstrate that the leximin solution can be modified to also provide compelling fairness guarantees for the allocation of indivisible chores. First, we generalize the definition of the leximin solution. Then, we show that the leximin solution finds a PROP1 (proportional up to one good) and PO (Pareto-optimal) allocation for 3 or 4 agents in the context of chores allocation with additive distinct valuations. Additionally, we prove that the leximin solution is EFX for combinations of goods and chores for agents with general but identical valuations. |
1611.06065 | Maude Pupin | Qassim Esmaeel, Maude Pupin (CRIStAL, BONSAI), Nam Phuong Kieu,
Gabrielle Chataign\'e, Max B\'echet, Jovana Deravel, Fran\c{c}ois Krier,
Monica H\"ofte, Philippe Jacques, Val\'erie Lecl\`ere (CRIStAL, BONSAI) | Burkholderia genome mining for nonribosomal peptide synthetases reveals
a great potential for novel siderophores and lipopeptides synthesis | null | MicrobiologyOpen, 2016, 5 (3), pp.512 - 526 | 10.1002/mbo3.347 | null | q-bio.QM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Burkholderia is an important genus encompassing a variety of species,
including pathogenic strains as well as strains that promote plant growth. We
have carried out a global strategy, which combined two complementary
approaches. The first one is genome guided with deep analysis of genome
sequences and the second one is assay guided with experiments to support the
predictions obtained in silico. This efficient screening for new secondary
metabolites, performed on 48 gapless genomes of Burkholderia species, revealed
a total of 161 clusters containing nonribosomal peptide synthetases (NRPSs),
with the potential to synthesize at least 11 novel products. Most of them are
siderophores or lipopeptides, two classes of products with potential
application in biocontrol. The strategy led to the identification, for the
first time, of the cluster for cepaciachelin biosynthesis in the genome of
Burkholderia ambifaria AMMD and a cluster corresponding to a new
malleobactin-like siderophore, called phymabactin, was identified in
Burkholderia phymatum STM815 genome. In both cases, the siderophore was
produced when the strain was grown in iron-limited conditions. Elsewhere, the
cluster for the antifungal burkholdin was detected in the genome of B.
ambifaria AMMD and also Burkholderia sp. KJ006. Burkholderia pseudomallei
strains harbor the genetic potential to produce a novel lipopeptide called
burkhomycin, containing a peptidyl moiety of 12 monomers. A mixture of
lipopeptides produced by Burkholderia rhizoxinica lowered the surface tension
of the supernatant from 70 to 27 mN/m. The production of nonribosomal secondary
metabolites seems related to the three phylogenetic groups obtained from 16S
rRNA sequences. Moreover, the genome-mining approach gave new insights into the
nonribosomal synthesis exemplified by the identification of dual C/E domains in
lipopeptide NRPSs, up to now essentially found in Pseudomonas strains.
| [
{
"created": "Fri, 18 Nov 2016 13:28:43 GMT",
"version": "v1"
}
] | 2016-11-21 | [
[
"Esmaeel",
"Qassim",
"",
"CRIStAL, BONSAI"
],
[
"Pupin",
"Maude",
"",
"CRIStAL, BONSAI"
],
[
"Kieu",
"Nam Phuong",
"",
"CRIStAL, BONSAI"
],
[
"Chataigné",
"Gabrielle",
"",
"CRIStAL, BONSAI"
],
[
"Béchet",
"Max",
"",
"C... | Burkholderia is an important genus encompassing a variety of species, including pathogenic strains as well as strains that promote plant growth. We have carried out a global strategy, which combined two complementary approaches. The first one is genome guided with deep analysis of genome sequences and the second one is assay guided with experiments to support the predictions obtained in silico. This efficient screening for new secondary metabolites, performed on 48 gapless genomes of Burkholderia species, revealed a total of 161 clusters containing nonribosomal peptide synthetases (NRPSs), with the potential to synthesize at least 11 novel products. Most of them are siderophores or lipopeptides, two classes of products with potential application in biocontrol. The strategy led to the identification, for the first time, of the cluster for cepaciachelin biosynthesis in the genome of Burkholderia ambifaria AMMD and a cluster corresponding to a new malleobactin-like siderophore, called phymabactin, was identified in Burkholderia phymatum STM815 genome. In both cases, the siderophore was produced when the strain was grown in iron-limited conditions. Elsewhere, the cluster for the antifungal burkholdin was detected in the genome of B. ambifaria AMMD and also Burkholderia sp. KJ006. Burkholderia pseudomallei strains harbor the genetic potential to produce a novel lipopeptide called burkhomycin, containing a peptidyl moiety of 12 monomers. A mixture of lipopeptides produced by Burkholderia rhizoxinica lowered the surface tension of the supernatant from 70 to 27 mN/m. The production of nonribosomal secondary metabolites seems related to the three phylogenetic groups obtained from 16S rRNA sequences. Moreover, the genome-mining approach gave new insights into the nonribosomal synthesis exemplified by the identification of dual C/E domains in lipopeptide NRPSs, up to now essentially found in Pseudomonas strains. |
1108.5184 | Hong Lu | Yi-Xin Chen, H. Lu and Kai-Nan Shao | Linearized Modes in Extended and Critical Gravities | 24 pages, 2 figures | null | 10.1088/0264-9381/29/8/085017 | CAS-KITPC/ITP-277 | hep-th gr-qc | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We construct explicit solutions for the linearized massive and massless
spin-2, vector and scalar modes around the AdS spacetimes in diverse
dimensions. These modes may arise in extended (super)gravities with higher
curvature terms in general dimensions. Log modes in critical gravities can also
be straightforwardly deduced. We analyze the properties of these modes and
obtain the tachyon-free condition, which allows negative mass square for these
modes. However, such modes may not satisfy the standard AdS boundary condition
and can be truncated out from the spectrum.
| [
{
"created": "Thu, 25 Aug 2011 20:05:42 GMT",
"version": "v1"
}
] | 2015-05-30 | [
[
"Chen",
"Yi-Xin",
""
],
[
"Lu",
"H.",
""
],
[
"Shao",
"Kai-Nan",
""
]
] | We construct explicit solutions for the linearized massive and massless spin-2, vector and scalar modes around the AdS spacetimes in diverse dimensions. These modes may arise in extended (super)gravities with higher curvature terms in general dimensions. Log modes in critical gravities can also be straightforwardly deduced. We analyze the properties of these modes and obtain the tachyon-free condition, which allows negative mass square for these modes. However, such modes may not satisfy the standard AdS boundary condition and can be truncated out from the spectrum. |
hep-th/9504095 | Paul Townsend | P.K. Townsend | String-Membrane Duality in Seven Dimensions | The original version of this paper dealt mostly with one side of
string-membrane duality: the solitonic interpretation of the heterotic string
as a $K_3$ compactified D=11 superfivebrane. The revised version includes a
discussion of the converse prediction: that the supermembrane has a solitonic
interpretation as a $T^3$ compactified heterotic fivebrane. It also includes
a discussion of D=8 membrane-membrane duality, and various changes to the
references | Phys.Lett.B354:247-255,1995 | 10.1016/0370-2693(95)00649-6 | DAMTP, R/95/15 | hep-th | null | The conjectured equivalence of the heterotic string to a $K_3$ compactified
type IIA superstring is combined with the conjectured equivalence of the latter
to a compactified 11-dimensional supermembrane to derive a string membrane
duality in seven dimensions; the membrane is a soliton of the string theory and
vice versa. A prediction of this duality is that the heterotic string is a
$K_3$ compactification of the solitonic 11-dimensional fivebrane. It is
verified that the worldsheet action of the D=10 heterotic string is indeed
obtainable by $K_3$ compactification of the worldvolume action of the
11-dimensional fivebrane, and it is suggested how the worldvolume action of the
D=11 supermebrane may be similarly obtained by $T^3$ compactification of the
worldvolume action of a D=10 heterotic fivebrane. Generalizations to $D=8$
string-threebrane and membrane-membrane duality are also discussed.
| [
{
"created": "Tue, 18 Apr 1995 12:22:24 GMT",
"version": "v1"
},
{
"created": "Wed, 10 May 1995 16:41:11 GMT",
"version": "v2"
}
] | 2010-11-01 | [
[
"Townsend",
"P. K.",
""
]
] | The conjectured equivalence of the heterotic string to a $K_3$ compactified type IIA superstring is combined with the conjectured equivalence of the latter to a compactified 11-dimensional supermembrane to derive a string membrane duality in seven dimensions; the membrane is a soliton of the string theory and vice versa. A prediction of this duality is that the heterotic string is a $K_3$ compactification of the solitonic 11-dimensional fivebrane. It is verified that the worldsheet action of the D=10 heterotic string is indeed obtainable by $K_3$ compactification of the worldvolume action of the 11-dimensional fivebrane, and it is suggested how the worldvolume action of the D=11 supermebrane may be similarly obtained by $T^3$ compactification of the worldvolume action of a D=10 heterotic fivebrane. Generalizations to $D=8$ string-threebrane and membrane-membrane duality are also discussed. |
2307.04815 | Raul De Palma Aristides | R. P. Aristides and A. J. Pons and H. A. Cerdeira and C. Masoller and
G. Tirabass | Parameter and coupling estimation in small groups of Izhikevich neurons | null | Chaos, vol. 33, n. 4, 2023 | 10.1063/5.0144499 | null | q-bio.NC | http://creativecommons.org/licenses/by/4.0/ | Nowadays, experimental techniques allow scientists to have access to large
amounts of data. In order to obtain reliable information from the complex
systems which produce these data, appropriate analysis tools are needed}. The
Kalman filter is a {frequently used} technique to infer, assuming a model of
the system, the parameters of the model from uncertain observations. A
well-known implementation of the Kalman filter, the Unscented Kalman filter
(UKF), was recently shown to be able to infer the connectivity of a set of
coupled chaotic oscillators. {I}n this work, we test whether the UKF can also
reconstruct the connectivity of {small groups of} coupled neurons when their
links are either electrical or chemical {synapses}. {In particular, w}e
consider Izhikevich neurons, and aim to infer which neurons influence each
other, considering {simulated spike trains as the experimental observations
used by the UKF}. First, we {verify} that the UKF can recover the parameters of
a single neuron, even when the parameters vary in time. Second, we analyze
small neural ensembles and}} demonstrate that the UKF allows inferring the
connectivity between the neurons, even for heterogeneous, directed, and
{temporally evolving} networks. {Our results show that time-dependent parameter
and coupling estimation is possible in this nonlinearly coupled system.
| [
{
"created": "Fri, 16 Jun 2023 10:27:50 GMT",
"version": "v1"
}
] | 2023-07-12 | [
[
"Aristides",
"R. P.",
""
],
[
"Pons",
"A. J.",
""
],
[
"Cerdeira",
"H. A.",
""
],
[
"Masoller",
"C.",
""
],
[
"Tirabass",
"G.",
""
]
] | Nowadays, experimental techniques allow scientists to have access to large amounts of data. In order to obtain reliable information from the complex systems which produce these data, appropriate analysis tools are needed}. The Kalman filter is a {frequently used} technique to infer, assuming a model of the system, the parameters of the model from uncertain observations. A well-known implementation of the Kalman filter, the Unscented Kalman filter (UKF), was recently shown to be able to infer the connectivity of a set of coupled chaotic oscillators. {I}n this work, we test whether the UKF can also reconstruct the connectivity of {small groups of} coupled neurons when their links are either electrical or chemical {synapses}. {In particular, w}e consider Izhikevich neurons, and aim to infer which neurons influence each other, considering {simulated spike trains as the experimental observations used by the UKF}. First, we {verify} that the UKF can recover the parameters of a single neuron, even when the parameters vary in time. Second, we analyze small neural ensembles and}} demonstrate that the UKF allows inferring the connectivity between the neurons, even for heterogeneous, directed, and {temporally evolving} networks. {Our results show that time-dependent parameter and coupling estimation is possible in this nonlinearly coupled system. |
hep-th/9408139 | Hu Zhan-ning | Zhan-Ning Hu, Bo-Yu Hou | Remarks on the Star-Triangle Relation in the Baxter-Bazhanov Model | 6 pages, latex file, AS-ITP-94-39 | null | 10.1007/BF02184882 | null | hep-th | null | In this letter we show that the restricted star-triangle relation introduced
by Bazhanov and Baxter can be obtained either from the star-triangle relation
of chiral Potts model or from the star-square relation which is proposed by
Kashaev $et ~al$ and give a response of the guess which is suggested by
Bazhanov and Baxter in Ref. \cite{b2}.
| [
{
"created": "Thu, 25 Aug 1994 23:08:28 GMT",
"version": "v1"
}
] | 2009-10-28 | [
[
"Hu",
"Zhan-Ning",
""
],
[
"Hou",
"Bo-Yu",
""
]
] | In this letter we show that the restricted star-triangle relation introduced by Bazhanov and Baxter can be obtained either from the star-triangle relation of chiral Potts model or from the star-square relation which is proposed by Kashaev $et ~al$ and give a response of the guess which is suggested by Bazhanov and Baxter in Ref. \cite{b2}. |
1107.5827 | Girma Hailu | Girma Hailu | Linear Confinement of Quarks from Supergravity | 8 pages, PDFLaTeX | Phys.Rev. D84 (2011) 106008 | 10.1103/PhysRevD.84.106008 | null | hep-th hep-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A supergravity background that produces linear confinement of quarks in four
dimensions is presented.
| [
{
"created": "Thu, 28 Jul 2011 20:44:25 GMT",
"version": "v1"
}
] | 2015-05-30 | [
[
"Hailu",
"Girma",
""
]
] | A supergravity background that produces linear confinement of quarks in four dimensions is presented. |
hep-th/9306115 | null | Jan Sladkowski | Does noncommutative geometry predict nonlinear Higgs mechanism? | 12 pages, LaTeX file, BI-TP 93/26 | Int.J.Theor.Phys. 33 (1994) 2381-2388 | 10.1007/BF00673963 | null | hep-th hep-ph | null | It is argued that the noncommutative geometry construction of the standard
model predicts a nonlinear symmetry breaking mechanism rather than the orthodox
Higgs mechanism. Such models have experimentally verifiable consequences.
| [
{
"created": "Tue, 22 Jun 1993 15:11:14 GMT",
"version": "v1"
}
] | 2009-10-22 | [
[
"Sladkowski",
"Jan",
""
]
] | It is argued that the noncommutative geometry construction of the standard model predicts a nonlinear symmetry breaking mechanism rather than the orthodox Higgs mechanism. Such models have experimentally verifiable consequences. |
1208.0811 | Guanhong Pei | Guanhong Pei and Anil Kumar S. Vullikanti | Efficient Algorithms for Maximum Link Scheduling in Distributed
Computing Models with SINR Constraints | null | null | null | null | cs.DC cs.DS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A fundamental problem in wireless networks is the maximum link scheduling
problem: given a set $L$ of links, compute the largest possible subset
$L'\subseteq L$ of links that can be scheduled simultaneously without
interference. This problem is particularly challenging in the physical
interference model based on SINR constraints (referred to as the SINR model),
which has gained a lot of interest in recent years. Constant factor
approximation algorithms have been developed for this problem, but low
complexity distributed algorithms that give the same approximation guarantee in
the SINR model are not known. Distributed algorithms are especially challenging
in this model, because of its non-locality.
In this paper, we develop a set of fast distributed algorithms in the SINR
model, providing constant approximation for the maximum link scheduling problem
under uniform power assignment. We find that different aspects of available
technology, such as full/half-duplex communication, and non-adaptive/adaptive
power control, have a significant impact on the performance of the algorithm;
these issues have not been explored in the context of distributed algorithms in
the SINR model before. Our algorithms' running time is $O(g(L) \log^c m)$,
where $c=1,2,3$ for different problem instances, and $g(L)$ is the "link
diversity" determined by the logarithmic scale of a communication link length.
Since $g(L)$ is small and remains in a constant range in most cases, our
algorithms serve as the first set of "sublinear" time distributed solution. The
algorithms are randomized and crucially use physical carrier sensing in
distributed communication steps.
| [
{
"created": "Fri, 3 Aug 2012 18:26:06 GMT",
"version": "v1"
},
{
"created": "Fri, 16 Nov 2012 15:46:26 GMT",
"version": "v2"
}
] | 2012-11-19 | [
[
"Pei",
"Guanhong",
""
],
[
"Vullikanti",
"Anil Kumar S.",
""
]
] | A fundamental problem in wireless networks is the maximum link scheduling problem: given a set $L$ of links, compute the largest possible subset $L'\subseteq L$ of links that can be scheduled simultaneously without interference. This problem is particularly challenging in the physical interference model based on SINR constraints (referred to as the SINR model), which has gained a lot of interest in recent years. Constant factor approximation algorithms have been developed for this problem, but low complexity distributed algorithms that give the same approximation guarantee in the SINR model are not known. Distributed algorithms are especially challenging in this model, because of its non-locality. In this paper, we develop a set of fast distributed algorithms in the SINR model, providing constant approximation for the maximum link scheduling problem under uniform power assignment. We find that different aspects of available technology, such as full/half-duplex communication, and non-adaptive/adaptive power control, have a significant impact on the performance of the algorithm; these issues have not been explored in the context of distributed algorithms in the SINR model before. Our algorithms' running time is $O(g(L) \log^c m)$, where $c=1,2,3$ for different problem instances, and $g(L)$ is the "link diversity" determined by the logarithmic scale of a communication link length. Since $g(L)$ is small and remains in a constant range in most cases, our algorithms serve as the first set of "sublinear" time distributed solution. The algorithms are randomized and crucially use physical carrier sensing in distributed communication steps. |
2002.05712 | Zhuliang Yao | Zhuliang Yao, Yue Cao, Shuxin Zheng, Gao Huang, Stephen Lin | Cross-Iteration Batch Normalization | Accepted to CVPR 2021 | null | null | null | cs.LG cs.CV stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A well-known issue of Batch Normalization is its significantly reduced
effectiveness in the case of small mini-batch sizes. When a mini-batch contains
few examples, the statistics upon which the normalization is defined cannot be
reliably estimated from it during a training iteration. To address this
problem, we present Cross-Iteration Batch Normalization (CBN), in which
examples from multiple recent iterations are jointly utilized to enhance
estimation quality. A challenge of computing statistics over multiple
iterations is that the network activations from different iterations are not
comparable to each other due to changes in network weights. We thus compensate
for the network weight changes via a proposed technique based on Taylor
polynomials, so that the statistics can be accurately estimated and batch
normalization can be effectively applied. On object detection and image
classification with small mini-batch sizes, CBN is found to outperform the
original batch normalization and a direct calculation of statistics over
previous iterations without the proposed compensation technique. Code is
available at https://github.com/Howal/Cross-iterationBatchNorm .
| [
{
"created": "Thu, 13 Feb 2020 18:52:57 GMT",
"version": "v1"
},
{
"created": "Fri, 14 Feb 2020 11:10:04 GMT",
"version": "v2"
},
{
"created": "Thu, 25 Mar 2021 06:57:36 GMT",
"version": "v3"
}
] | 2021-03-26 | [
[
"Yao",
"Zhuliang",
""
],
[
"Cao",
"Yue",
""
],
[
"Zheng",
"Shuxin",
""
],
[
"Huang",
"Gao",
""
],
[
"Lin",
"Stephen",
""
]
] | A well-known issue of Batch Normalization is its significantly reduced effectiveness in the case of small mini-batch sizes. When a mini-batch contains few examples, the statistics upon which the normalization is defined cannot be reliably estimated from it during a training iteration. To address this problem, we present Cross-Iteration Batch Normalization (CBN), in which examples from multiple recent iterations are jointly utilized to enhance estimation quality. A challenge of computing statistics over multiple iterations is that the network activations from different iterations are not comparable to each other due to changes in network weights. We thus compensate for the network weight changes via a proposed technique based on Taylor polynomials, so that the statistics can be accurately estimated and batch normalization can be effectively applied. On object detection and image classification with small mini-batch sizes, CBN is found to outperform the original batch normalization and a direct calculation of statistics over previous iterations without the proposed compensation technique. Code is available at https://github.com/Howal/Cross-iterationBatchNorm . |
2405.13182 | Zachary Kilpatrick PhD | Heather L Cihak and Zachary P Kilpatrick | Robustly encoding certainty in a metastable neural circuit model | 15 pages, 10 figures | null | null | null | q-bio.NC nlin.PS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Localized persistent neural activity can encode delayed estimates of
continuous variables. Common experiments require that subjects store and report
the feature value (e.g., orientation) of a particular cue (e.g., oriented bar
on a screen) after a delay. Visualizing recorded activity of neurons along
their feature tuning reveals activity bumps whose centers wander
stochastically, degrading the estimate over time. Bump position therefore
represents the remembered estimate. Recent work suggests bump amplitude may
represent estimate certainty reflecting a probabilistic population code for a
Bayesian posterior. Idealized models of this type are fragile due to the fine
tuning common to constructed continuum attractors in dynamical systems. Here we
propose an alternative metastable model for robustly supporting multiple bump
amplitudes by extending neural circuit models to include quantized
nonlinearities. Asymptotic projections of circuit activity produce
low-dimensional evolution equations for the amplitude and position of bump
solutions in response to external stimuli and noise perturbations. Analysis of
reduced equations accurately characterizes phase variance and the dynamics of
amplitude transitions between stable discrete values. More salient cues
generate bumps of higher amplitude which wander less, consistent with the
experimental finding that greater certainty correlates with more accurate
memories.
| [
{
"created": "Tue, 21 May 2024 20:13:35 GMT",
"version": "v1"
},
{
"created": "Tue, 30 Jul 2024 19:15:50 GMT",
"version": "v2"
}
] | 2024-08-01 | [
[
"Cihak",
"Heather L",
""
],
[
"Kilpatrick",
"Zachary P",
""
]
] | Localized persistent neural activity can encode delayed estimates of continuous variables. Common experiments require that subjects store and report the feature value (e.g., orientation) of a particular cue (e.g., oriented bar on a screen) after a delay. Visualizing recorded activity of neurons along their feature tuning reveals activity bumps whose centers wander stochastically, degrading the estimate over time. Bump position therefore represents the remembered estimate. Recent work suggests bump amplitude may represent estimate certainty reflecting a probabilistic population code for a Bayesian posterior. Idealized models of this type are fragile due to the fine tuning common to constructed continuum attractors in dynamical systems. Here we propose an alternative metastable model for robustly supporting multiple bump amplitudes by extending neural circuit models to include quantized nonlinearities. Asymptotic projections of circuit activity produce low-dimensional evolution equations for the amplitude and position of bump solutions in response to external stimuli and noise perturbations. Analysis of reduced equations accurately characterizes phase variance and the dynamics of amplitude transitions between stable discrete values. More salient cues generate bumps of higher amplitude which wander less, consistent with the experimental finding that greater certainty correlates with more accurate memories. |
2011.02574 | Andrei Cramariuc | Le Chen, Yunke Ao, Florian Tschopp, Andrei Cramariuc, Michel Breyer,
Jen Jen Chung, Roland Siegwart, Cesar Cadena | Learning Trajectories for Visual-Inertial System Calibration via
Model-based Heuristic Deep Reinforcement Learning | null | Proceedings of the 4th Conference on Robot Learning (CoRL) 2020 | null | null | cs.RO cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Visual-inertial systems rely on precise calibrations of both camera
intrinsics and inter-sensor extrinsics, which typically require manually
performing complex motions in front of a calibration target. In this work we
present a novel approach to obtain favorable trajectories for visual-inertial
system calibration, using model-based deep reinforcement learning. Our key
contribution is to model the calibration process as a Markov decision process
and then use model-based deep reinforcement learning with particle swarm
optimization to establish a sequence of calibration trajectories to be
performed by a robot arm. Our experiments show that while maintaining similar
or shorter path lengths, the trajectories generated by our learned policy
result in lower calibration errors compared to random or handcrafted
trajectories.
| [
{
"created": "Wed, 4 Nov 2020 23:20:15 GMT",
"version": "v1"
}
] | 2021-02-17 | [
[
"Chen",
"Le",
""
],
[
"Ao",
"Yunke",
""
],
[
"Tschopp",
"Florian",
""
],
[
"Cramariuc",
"Andrei",
""
],
[
"Breyer",
"Michel",
""
],
[
"Chung",
"Jen Jen",
""
],
[
"Siegwart",
"Roland",
""
],
[
"Caden... | Visual-inertial systems rely on precise calibrations of both camera intrinsics and inter-sensor extrinsics, which typically require manually performing complex motions in front of a calibration target. In this work we present a novel approach to obtain favorable trajectories for visual-inertial system calibration, using model-based deep reinforcement learning. Our key contribution is to model the calibration process as a Markov decision process and then use model-based deep reinforcement learning with particle swarm optimization to establish a sequence of calibration trajectories to be performed by a robot arm. Our experiments show that while maintaining similar or shorter path lengths, the trajectories generated by our learned policy result in lower calibration errors compared to random or handcrafted trajectories. |
1805.04625 | Shun Watanabe | Himanshu Tyagi, Shun Watanabe | Strong Converse using Change of Measure Arguments | 35 pages, no figure; v2 updated references | null | null | null | cs.IT math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The strong converse for a coding theorem shows that the optimal asymptotic
rate possible with vanishing error cannot be improved by allowing a fixed
error. Building on a method introduced by Gu and Effros for centralized coding
problems, we develop a general and simple recipe for proving strong converse
that is applicable for distributed problems as well. Heuristically, our proof
of strong converse mimics the standard steps for proving a weak converse,
except that we apply those steps to a modified distribution obtained by
conditioning the original distribution on the event that no error occurs. A key
component of our recipe is the replacement of the hard Markov constraints
implied by the distributed nature of the problem with a soft information cost
using a variational formula introduced by Oohama. We illustrate our method by
providing a short proof of the strong converse for the Wyner-Ziv problem and
strong converse theorems for interactive function computation, common
randomness and secret key agreement, and the wiretap channel; the latter three
strong converse problems were open prior to this work.
| [
{
"created": "Sat, 12 May 2018 00:34:37 GMT",
"version": "v1"
},
{
"created": "Wed, 21 Aug 2019 14:13:36 GMT",
"version": "v2"
}
] | 2019-08-22 | [
[
"Tyagi",
"Himanshu",
""
],
[
"Watanabe",
"Shun",
""
]
] | The strong converse for a coding theorem shows that the optimal asymptotic rate possible with vanishing error cannot be improved by allowing a fixed error. Building on a method introduced by Gu and Effros for centralized coding problems, we develop a general and simple recipe for proving strong converse that is applicable for distributed problems as well. Heuristically, our proof of strong converse mimics the standard steps for proving a weak converse, except that we apply those steps to a modified distribution obtained by conditioning the original distribution on the event that no error occurs. A key component of our recipe is the replacement of the hard Markov constraints implied by the distributed nature of the problem with a soft information cost using a variational formula introduced by Oohama. We illustrate our method by providing a short proof of the strong converse for the Wyner-Ziv problem and strong converse theorems for interactive function computation, common randomness and secret key agreement, and the wiretap channel; the latter three strong converse problems were open prior to this work. |
1510.02840 | Mauricio Toro | Mauricio Toro | Concurrent Constraint Machine Improvisation: Models and Implementation | 8 pages | null | null | null | cs.LO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Machine improvisation creates music either by explicit coding of rules or by
applying machine learning methods. We deal with the latter case. An
improvisation system capable of real-time must execute two process
concurrently: one to apply machine learning methods to musical sequences in
order to capture prominent musical features, and one to produce musical
sequences stylistically consistent with the learned material. As an example,
the Concurrent Constraint Factor Oracle Model for Music Improvisation (ccfomi),
based upon Non-deterministic Timed Concurrent Constraint (ntcc) calculus, uses
the Factor Oracle to store the learned sequences.
| [
{
"created": "Fri, 9 Oct 2015 22:22:01 GMT",
"version": "v1"
}
] | 2015-10-13 | [
[
"Toro",
"Mauricio",
""
]
] | Machine improvisation creates music either by explicit coding of rules or by applying machine learning methods. We deal with the latter case. An improvisation system capable of real-time must execute two process concurrently: one to apply machine learning methods to musical sequences in order to capture prominent musical features, and one to produce musical sequences stylistically consistent with the learned material. As an example, the Concurrent Constraint Factor Oracle Model for Music Improvisation (ccfomi), based upon Non-deterministic Timed Concurrent Constraint (ntcc) calculus, uses the Factor Oracle to store the learned sequences. |
2108.08679 | Alexander Barg | Alexander Barg, Zitan Chen, and Itzhak Tamo | A construction of maximally recoverable codes | null | Designs, Codes and Cryptography, 2022, vol. 90, pp. 939-945 | 10.1007/s10623-022-01020-8 | null | cs.IT math.CO math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We construct a family of linear maximally recoverable codes with locality $r$
and dimension $r+1.$ For codes of length $n$ with $r\approx n^\alpha,
0\le\alpha\le 1$ the code alphabet is of the order $n^{1+3\alpha},$ which
improves upon the previously known constructions of maximally recoverable
codes.
| [
{
"created": "Thu, 19 Aug 2021 13:40:55 GMT",
"version": "v1"
}
] | 2023-03-07 | [
[
"Barg",
"Alexander",
""
],
[
"Chen",
"Zitan",
""
],
[
"Tamo",
"Itzhak",
""
]
] | We construct a family of linear maximally recoverable codes with locality $r$ and dimension $r+1.$ For codes of length $n$ with $r\approx n^\alpha, 0\le\alpha\le 1$ the code alphabet is of the order $n^{1+3\alpha},$ which improves upon the previously known constructions of maximally recoverable codes. |
2309.14950 | Atif Belal | Atif Belal, Akhil Meethal, Francisco Perdigon Romero, Marco Pedersoli,
Eric Granger | Multi-Source Domain Adaptation for Object Detection with Prototype-based
Mean-teacher | null | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by-sa/4.0/ | Adapting visual object detectors to operational target domains is a
challenging task, commonly achieved using unsupervised domain adaptation (UDA)
methods. Recent studies have shown that when the labeled dataset comes from
multiple source domains, treating them as separate domains and performing a
multi-source domain adaptation (MSDA) improves the accuracy and robustness over
blending these source domains and performing a UDA. For adaptation, existing
MSDA methods learn domain-invariant and domain-specific parameters (for each
source domain). However, unlike single-source UDA methods, learning
domain-specific parameters makes them grow significantly in proportion to the
number of source domains. This paper proposes a novel MSDA method called
Prototype-based Mean Teacher (PMT), which uses class prototypes instead of
domain-specific subnets to encode domain-specific information. These prototypes
are learned using a contrastive loss, aligning the same categories across
domains and separating different categories far apart. Given the use of
prototypes, the number of parameters required for our PMT method does not
increase significantly with the number of source domains, thus reducing memory
issues and possible overfitting. Empirical studies indicate that PMT
outperforms state-of-the-art MSDA methods on several challenging object
detection datasets. Our code is available at
https://github.com/imatif17/Prototype-Mean-Teacher.
| [
{
"created": "Tue, 26 Sep 2023 14:08:03 GMT",
"version": "v1"
},
{
"created": "Fri, 15 Dec 2023 21:00:50 GMT",
"version": "v2"
},
{
"created": "Wed, 31 Jul 2024 20:04:53 GMT",
"version": "v3"
}
] | 2024-08-02 | [
[
"Belal",
"Atif",
""
],
[
"Meethal",
"Akhil",
""
],
[
"Romero",
"Francisco Perdigon",
""
],
[
"Pedersoli",
"Marco",
""
],
[
"Granger",
"Eric",
""
]
] | Adapting visual object detectors to operational target domains is a challenging task, commonly achieved using unsupervised domain adaptation (UDA) methods. Recent studies have shown that when the labeled dataset comes from multiple source domains, treating them as separate domains and performing a multi-source domain adaptation (MSDA) improves the accuracy and robustness over blending these source domains and performing a UDA. For adaptation, existing MSDA methods learn domain-invariant and domain-specific parameters (for each source domain). However, unlike single-source UDA methods, learning domain-specific parameters makes them grow significantly in proportion to the number of source domains. This paper proposes a novel MSDA method called Prototype-based Mean Teacher (PMT), which uses class prototypes instead of domain-specific subnets to encode domain-specific information. These prototypes are learned using a contrastive loss, aligning the same categories across domains and separating different categories far apart. Given the use of prototypes, the number of parameters required for our PMT method does not increase significantly with the number of source domains, thus reducing memory issues and possible overfitting. Empirical studies indicate that PMT outperforms state-of-the-art MSDA methods on several challenging object detection datasets. Our code is available at https://github.com/imatif17/Prototype-Mean-Teacher. |
2303.17251 | Stefano Cresci | Stefano Cresci, Kai-Cheng Yang, Angelo Spognardi, Roberto Di Pietro,
Filippo Menczer, Marinella Petrocchi | Demystifying Misconceptions in Social Bots Research | null | null | null | null | cs.SI cs.AI cs.CY cs.LG | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Research on social bots aims at advancing knowledge and providing solutions
to one of the most debated forms of online manipulation. Yet, social bot
research is plagued by widespread biases, hyped results, and misconceptions
that set the stage for ambiguities, unrealistic expectations, and seemingly
irreconcilable findings. Overcoming such issues is instrumental towards
ensuring reliable solutions and reaffirming the validity of the scientific
method. In this contribution, we review some recent results in social bots
research, highlighting and revising factual errors as well as methodological
and conceptual biases. More importantly, we demystify common misconceptions,
addressing fundamental points on how social bots research is discussed. Our
analysis surfaces the need to discuss research about online disinformation and
manipulation in a rigorous, unbiased, and responsible way. This article
bolsters such effort by identifying and refuting common fallacious arguments
used by both proponents and opponents of social bots research, as well as
providing directions toward sound methodologies for future research in the
field.
| [
{
"created": "Thu, 30 Mar 2023 09:29:53 GMT",
"version": "v1"
},
{
"created": "Wed, 27 Mar 2024 14:48:48 GMT",
"version": "v2"
}
] | 2024-03-28 | [
[
"Cresci",
"Stefano",
""
],
[
"Yang",
"Kai-Cheng",
""
],
[
"Spognardi",
"Angelo",
""
],
[
"Di Pietro",
"Roberto",
""
],
[
"Menczer",
"Filippo",
""
],
[
"Petrocchi",
"Marinella",
""
]
] | Research on social bots aims at advancing knowledge and providing solutions to one of the most debated forms of online manipulation. Yet, social bot research is plagued by widespread biases, hyped results, and misconceptions that set the stage for ambiguities, unrealistic expectations, and seemingly irreconcilable findings. Overcoming such issues is instrumental towards ensuring reliable solutions and reaffirming the validity of the scientific method. In this contribution, we review some recent results in social bots research, highlighting and revising factual errors as well as methodological and conceptual biases. More importantly, we demystify common misconceptions, addressing fundamental points on how social bots research is discussed. Our analysis surfaces the need to discuss research about online disinformation and manipulation in a rigorous, unbiased, and responsible way. This article bolsters such effort by identifying and refuting common fallacious arguments used by both proponents and opponents of social bots research, as well as providing directions toward sound methodologies for future research in the field. |
1607.04063 | Carsten Witt | Dirk Sudholt and Carsten Witt | Update Strength in EDAs and ACO: How to Avoid Genetic Drift | 32 pages. An extended abstract of this work will appear in the
proceedings of the Genetic and Evolutionary Computation Conference (GECCO
2016). This revision fixes the abstract in the metadata | null | null | null | cs.NE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We provide a rigorous runtime analysis concerning the update strength, a
vital parameter in probabilistic model-building GAs such as the step size $1/K$
in the compact Genetic Algorithm (cGA) and the evaporation factor $\rho$ in
ACO. While a large update strength is desirable for exploitation, there is a
general trade-off: too strong updates can lead to genetic drift and poor
performance. We demonstrate this trade-off for the cGA and a simple MMAS ACO
algorithm on the OneMax function. More precisely, we obtain lower bounds on the
expected runtime of $\Omega(K\sqrt{n} + n \log n)$ and $\Omega(\sqrt{n}/\rho +
n \log n)$, respectively, showing that the update strength should be limited to
$1/K, \rho = O(1/(\sqrt{n} \log n))$. In fact, choosing $1/K, \rho \sim
1/(\sqrt{n}\log n)$ both algorithms efficiently optimize OneMax in expected
time $O(n \log n)$. Our analyses provide new insights into the stochastic
behavior of probabilistic model-building GAs and propose new guidelines for
setting the update strength in global optimization.
| [
{
"created": "Thu, 14 Jul 2016 10:11:59 GMT",
"version": "v1"
},
{
"created": "Fri, 15 Jul 2016 07:51:28 GMT",
"version": "v2"
}
] | 2016-07-18 | [
[
"Sudholt",
"Dirk",
""
],
[
"Witt",
"Carsten",
""
]
] | We provide a rigorous runtime analysis concerning the update strength, a vital parameter in probabilistic model-building GAs such as the step size $1/K$ in the compact Genetic Algorithm (cGA) and the evaporation factor $\rho$ in ACO. While a large update strength is desirable for exploitation, there is a general trade-off: too strong updates can lead to genetic drift and poor performance. We demonstrate this trade-off for the cGA and a simple MMAS ACO algorithm on the OneMax function. More precisely, we obtain lower bounds on the expected runtime of $\Omega(K\sqrt{n} + n \log n)$ and $\Omega(\sqrt{n}/\rho + n \log n)$, respectively, showing that the update strength should be limited to $1/K, \rho = O(1/(\sqrt{n} \log n))$. In fact, choosing $1/K, \rho \sim 1/(\sqrt{n}\log n)$ both algorithms efficiently optimize OneMax in expected time $O(n \log n)$. Our analyses provide new insights into the stochastic behavior of probabilistic model-building GAs and propose new guidelines for setting the update strength in global optimization. |
2205.08664 | Taro L. Saito | Taro L. Saito, Naoki Takezoe, Yukihiro Okada, Takako Shimamoto,
Dongmin Yu, Suprith Chandrashekharachar, Kai Sasaki, Shohei Okumiya, Yan
Wang, Takashi Kurihara, Ryu Kobayashi, Keisuke Suzuki, Zhenghong Yang, Makoto
Onizuka | Journey of Migrating Millions of Queries on The Cloud | This version is published in DBTest '22: Proceedings of the 2022
workshop on 9th International Workshop of Testing Database Systems | null | 10.1145/3531348.3532177 | null | cs.DB | http://creativecommons.org/licenses/by/4.0/ | Treasure Data is processing millions of distributed SQL queries every day on
the cloud. Upgrading the query engine service at this scale is challenging
because we need to migrate all of the production queries of the customers to a
new version while preserving the correctness and performance of the data
processing pipelines. To ensure the quality of the query engines, we utilize
our query logs to build customer-specific benchmarks and replay these queries
with real customer data in a secure pre-production environment. To simulate
millions of queries, we need effective minimization of test query sets and
better reporting of the simulation results to proactively find incompatible
changes and performance regression of the new version. This paper describes the
overall design of our system and shares various challenges in maintaining the
quality of the query engine service on the cloud.
| [
{
"created": "Tue, 17 May 2022 23:48:26 GMT",
"version": "v1"
}
] | 2022-05-19 | [
[
"Saito",
"Taro L.",
""
],
[
"Takezoe",
"Naoki",
""
],
[
"Okada",
"Yukihiro",
""
],
[
"Shimamoto",
"Takako",
""
],
[
"Yu",
"Dongmin",
""
],
[
"Chandrashekharachar",
"Suprith",
""
],
[
"Sasaki",
"Kai",
""
]... | Treasure Data is processing millions of distributed SQL queries every day on the cloud. Upgrading the query engine service at this scale is challenging because we need to migrate all of the production queries of the customers to a new version while preserving the correctness and performance of the data processing pipelines. To ensure the quality of the query engines, we utilize our query logs to build customer-specific benchmarks and replay these queries with real customer data in a secure pre-production environment. To simulate millions of queries, we need effective minimization of test query sets and better reporting of the simulation results to proactively find incompatible changes and performance regression of the new version. This paper describes the overall design of our system and shares various challenges in maintaining the quality of the query engine service on the cloud. |
1403.0093 | Masoud Abbaszadeh | Masoud Abbaszadeh, Horacio J. Marquez | Robust Nonlinear L2 Filtering of Uncertain Lipschitz Systems via Pareto
Optimization | 21 pages, 5 figures. arXiv admin note: text overlap with
arXiv:1010.0696 | null | null | null | cs.SY math.OC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A new approach for robust Hinfty filtering for a class of Lipschitz nonlinear
systems with time-varying uncertainties both in the linear and nonlinear parts
of the system is proposed in an LMI framework. The admissible Lipschitz
constant of the system and the disturbance attenuation level are maximized
simultaneously through convex multiobjective optimization. The resulting Hinfty
filter guarantees asymptotic stability of the estimation error dynamics with
exponential convergence and is robust against nonlinear additive uncertainty
and time-varying parametric uncertainties. Explicit bounds on the nonlinear
uncertainty are derived based on norm-wise and element-wise robustness
analysis.
| [
{
"created": "Sat, 1 Mar 2014 15:08:19 GMT",
"version": "v1"
}
] | 2014-03-04 | [
[
"Abbaszadeh",
"Masoud",
""
],
[
"Marquez",
"Horacio J.",
""
]
] | A new approach for robust Hinfty filtering for a class of Lipschitz nonlinear systems with time-varying uncertainties both in the linear and nonlinear parts of the system is proposed in an LMI framework. The admissible Lipschitz constant of the system and the disturbance attenuation level are maximized simultaneously through convex multiobjective optimization. The resulting Hinfty filter guarantees asymptotic stability of the estimation error dynamics with exponential convergence and is robust against nonlinear additive uncertainty and time-varying parametric uncertainties. Explicit bounds on the nonlinear uncertainty are derived based on norm-wise and element-wise robustness analysis. |
1011.3278 | Xiao-Lun Wu | Tuba Altindal, Li Xie, Xiao-Lun Wu | Implications of 3-step swimming patterns in bacterial chemotaxis | 18 pages, 4 figures, submitted to biophysical journal | null | 10.1016/j.bpj.2010.11.029 | null | q-bio.CB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We recently found that marine bacteria Vibrio alginolyticus execute a cyclic
3-step (run- reverse-flick) motility pattern that is distinctively different
from the 2-step (run-tumble) pattern of Escherichia coli. How this novel
swimming pattern is regulated by cells of V. alginolyticus is not currently
known, but its significance for bacterial chemotaxis is self- evident and will
be delineated herein. Using an approach introduced by de Gennes, we calculated
the migration speed of a cell executing the 3-step pattern in a linear chemical
gradient, and found that a biphasic chemotactic response arises naturally. The
implication of such a response for the cells to adapt to ocean environments and
its possible connection to E. coli 's response are also discussed.
| [
{
"created": "Mon, 15 Nov 2010 01:42:38 GMT",
"version": "v1"
}
] | 2017-07-26 | [
[
"Altindal",
"Tuba",
""
],
[
"Xie",
"Li",
""
],
[
"Wu",
"Xiao-Lun",
""
]
] | We recently found that marine bacteria Vibrio alginolyticus execute a cyclic 3-step (run- reverse-flick) motility pattern that is distinctively different from the 2-step (run-tumble) pattern of Escherichia coli. How this novel swimming pattern is regulated by cells of V. alginolyticus is not currently known, but its significance for bacterial chemotaxis is self- evident and will be delineated herein. Using an approach introduced by de Gennes, we calculated the migration speed of a cell executing the 3-step pattern in a linear chemical gradient, and found that a biphasic chemotactic response arises naturally. The implication of such a response for the cells to adapt to ocean environments and its possible connection to E. coli 's response are also discussed. |
1906.11878 | Hossein Ghayoumi Zadeh | Mehdi Abbaszadeh, Aliakbar Rahimifard, Mohammadali Eftekhari, Hossein
Ghayoumi Zadeh, Ali Fayazi, Ali Dini, Mostafa Danaeian | Deep Learning-Based Classification Of the Defective Pistachios Via Deep
Autoencoder Neural Networks | null | null | null | null | cs.CV cs.LG cs.NE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Pistachio nut is mainly consumed as raw, salted or roasted because of its
high nutritional properties and favorable taste. Pistachio nuts with shell and
kernel defects, besides not being acceptable for a consumer, are also prone to
insects damage, mold decay, and aflatoxin contamination. In this research, a
deep learning-based imaging algorithm was developed to improve the sorting of
nuts with shell and kernel defects that indicate the risk of aflatoxin
contamination, such as dark stains, oily stains, adhering hull, fungal decay
and Aspergillus molds. This paper presents an unsupervised learning method to
classify defective and unpleasant pistachios based on deep Auto-encoder neural
networks. The testing of the designed neural network on a validation dataset
showed that nuts having dark stain, oily stain or adhering hull with an
accuracy of 80.3% can be distinguished from normal nuts. Due to the limited
memory available in the HPC of university, the results are reasonable and
justifiable.
| [
{
"created": "Mon, 10 Jun 2019 13:02:50 GMT",
"version": "v1"
}
] | 2019-07-01 | [
[
"Abbaszadeh",
"Mehdi",
""
],
[
"Rahimifard",
"Aliakbar",
""
],
[
"Eftekhari",
"Mohammadali",
""
],
[
"Zadeh",
"Hossein Ghayoumi",
""
],
[
"Fayazi",
"Ali",
""
],
[
"Dini",
"Ali",
""
],
[
"Danaeian",
"Mostafa",
... | Pistachio nut is mainly consumed as raw, salted or roasted because of its high nutritional properties and favorable taste. Pistachio nuts with shell and kernel defects, besides not being acceptable for a consumer, are also prone to insects damage, mold decay, and aflatoxin contamination. In this research, a deep learning-based imaging algorithm was developed to improve the sorting of nuts with shell and kernel defects that indicate the risk of aflatoxin contamination, such as dark stains, oily stains, adhering hull, fungal decay and Aspergillus molds. This paper presents an unsupervised learning method to classify defective and unpleasant pistachios based on deep Auto-encoder neural networks. The testing of the designed neural network on a validation dataset showed that nuts having dark stain, oily stain or adhering hull with an accuracy of 80.3% can be distinguished from normal nuts. Due to the limited memory available in the HPC of university, the results are reasonable and justifiable. |
1912.13382 | Jose Del Aguila Ferrandis Mr | Jos\'e del \'Aguila Ferrandis, Michael Triantafyllou, Chryssostomos
Chryssostomidis, George Karniadakis | Learning functionals via LSTM neural networks for predicting vessel
dynamics in extreme sea states | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Predicting motions of vessels in extreme sea states represents one of the
most challenging problems in naval hydrodynamics. It involves computing complex
nonlinear wave-body interactions, hence taxing heavily computational resources.
Here, we put forward a new simulation paradigm by training recurrent type
neural networks (RNNs) that take as input the stochastic wave elevation at a
certain sea state and output the main vessel motions, e.g., pitch, heave and
roll. We first compare the performance of standard RNNs versus GRU and LSTM
neural networks (NNs) and show that LSTM NNs lead to the best performance. We
then examine the testing error of two representative vessels, a catamaran in
sea state 1 and a battleship in sea state 8. We demonstrate that good accuracy
is achieved for both cases in predicting the vessel motions for unseen wave
elevations. We train the NNs with expensive CFD simulations offline, but upon
training, the prediction of the vessel dynamics online can be obtained at a
fraction of a second. This work is motivated by the universal approximation
theorem for functionals [1], and it is the first implementation of such theory
to realistic engineering problems.
| [
{
"created": "Mon, 23 Dec 2019 18:39:12 GMT",
"version": "v1"
}
] | 2020-01-01 | [
[
"Ferrandis",
"José del Águila",
""
],
[
"Triantafyllou",
"Michael",
""
],
[
"Chryssostomidis",
"Chryssostomos",
""
],
[
"Karniadakis",
"George",
""
]
] | Predicting motions of vessels in extreme sea states represents one of the most challenging problems in naval hydrodynamics. It involves computing complex nonlinear wave-body interactions, hence taxing heavily computational resources. Here, we put forward a new simulation paradigm by training recurrent type neural networks (RNNs) that take as input the stochastic wave elevation at a certain sea state and output the main vessel motions, e.g., pitch, heave and roll. We first compare the performance of standard RNNs versus GRU and LSTM neural networks (NNs) and show that LSTM NNs lead to the best performance. We then examine the testing error of two representative vessels, a catamaran in sea state 1 and a battleship in sea state 8. We demonstrate that good accuracy is achieved for both cases in predicting the vessel motions for unseen wave elevations. We train the NNs with expensive CFD simulations offline, but upon training, the prediction of the vessel dynamics online can be obtained at a fraction of a second. This work is motivated by the universal approximation theorem for functionals [1], and it is the first implementation of such theory to realistic engineering problems. |
0805.2621 | Jae-Suk Park | Hyun-Keun Jun and Jae-Suk Park | Topological Sigma B Model in 4-Dimensions | 16 pages, JHEP style (minor corrections) | JHEP 0811:005,2008 | 10.1088/1126-6708/2008/11/005 | null | hep-th | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose a 4-dimensional version of topological sigma B-model, governing
maps from a smooth compact 4-manifold M to a Calabi-Yau target manifold X. The
theory depends on on complex structure of X, while is independent of Kaehler
metric of X. The theory is also a 4-dimensiona topological field theory in the
sense that the theory is independent of variation of Riemannian metric of the
source 4-manifold M, potentially leading to new smooth invariant of
4-manifolds. We argue that the theory also comes with a topological family
parametrized by the extended moduli space of complex structures.
| [
{
"created": "Fri, 16 May 2008 21:09:06 GMT",
"version": "v1"
},
{
"created": "Wed, 10 Sep 2008 05:36:02 GMT",
"version": "v2"
}
] | 2009-12-07 | [
[
"Jun",
"Hyun-Keun",
""
],
[
"Park",
"Jae-Suk",
""
]
] | We propose a 4-dimensional version of topological sigma B-model, governing maps from a smooth compact 4-manifold M to a Calabi-Yau target manifold X. The theory depends on on complex structure of X, while is independent of Kaehler metric of X. The theory is also a 4-dimensiona topological field theory in the sense that the theory is independent of variation of Riemannian metric of the source 4-manifold M, potentially leading to new smooth invariant of 4-manifolds. We argue that the theory also comes with a topological family parametrized by the extended moduli space of complex structures. |
1602.03638 | Mikael Mortensen | Mikael Mortensen and Hans Petter Langtangen | High performance Python for direct numerical simulations of turbulent
flows | null | null | 10.1016/j.cpc.2016.02.005 | null | cs.MS cs.DC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Direct Numerical Simulations (DNS) of the Navier Stokes equations is an
invaluable research tool in fluid dynamics. Still, there are few publicly
available research codes and, due to the heavy number crunching implied,
available codes are usually written in low-level languages such as C/C++ or
Fortran. In this paper we describe a pure scientific Python pseudo-spectral DNS
code that nearly matches the performance of C++ for thousands of processors and
billions of unknowns. We also describe a version optimized through Cython, that
is found to match the speed of C++. The solvers are written from scratch in
Python, both the mesh, the MPI domain decomposition, and the temporal
integrators. The solvers have been verified and benchmarked on the Shaheen
supercomputer at the KAUST supercomputing laboratory, and we are able to show
very good scaling up to several thousand cores.
A very important part of the implementation is the mesh decomposition (we
implement both slab and pencil decompositions) and 3D parallel Fast Fourier
Transforms (FFT). The mesh decomposition and FFT routines have been implemented
in Python using serial FFT routines (either NumPy, pyFFTW or any other serial
FFT module), NumPy array manipulations and with MPI communications handled by
MPI for Python (mpi4py). We show how we are able to execute a 3D parallel FFT
in Python for a slab mesh decomposition using 4 lines of compact Python code,
for which the parallel performance on Shaheen is found to be slightly better
than similar routines provided through the FFTW library. For a pencil mesh
decomposition 7 lines of code is required to execute a transform.
| [
{
"created": "Thu, 11 Feb 2016 08:12:37 GMT",
"version": "v1"
}
] | 2016-05-04 | [
[
"Mortensen",
"Mikael",
""
],
[
"Langtangen",
"Hans Petter",
""
]
] | Direct Numerical Simulations (DNS) of the Navier Stokes equations is an invaluable research tool in fluid dynamics. Still, there are few publicly available research codes and, due to the heavy number crunching implied, available codes are usually written in low-level languages such as C/C++ or Fortran. In this paper we describe a pure scientific Python pseudo-spectral DNS code that nearly matches the performance of C++ for thousands of processors and billions of unknowns. We also describe a version optimized through Cython, that is found to match the speed of C++. The solvers are written from scratch in Python, both the mesh, the MPI domain decomposition, and the temporal integrators. The solvers have been verified and benchmarked on the Shaheen supercomputer at the KAUST supercomputing laboratory, and we are able to show very good scaling up to several thousand cores. A very important part of the implementation is the mesh decomposition (we implement both slab and pencil decompositions) and 3D parallel Fast Fourier Transforms (FFT). The mesh decomposition and FFT routines have been implemented in Python using serial FFT routines (either NumPy, pyFFTW or any other serial FFT module), NumPy array manipulations and with MPI communications handled by MPI for Python (mpi4py). We show how we are able to execute a 3D parallel FFT in Python for a slab mesh decomposition using 4 lines of compact Python code, for which the parallel performance on Shaheen is found to be slightly better than similar routines provided through the FFTW library. For a pencil mesh decomposition 7 lines of code is required to execute a transform. |
1610.06070 | Norihiro Tanahashi | Koji Hashimoto and Norihiro Tanahashi | Universality in Chaos of Particle Motion near Black Hole Horizon | 12 pages, 4 figures; v2: references added, numerical plots in Fig. 4
corrected | Phys. Rev. D 95, 024007 (2017) | 10.1103/PhysRevD.95.024007 | OU-HET-911 | hep-th gr-qc | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Motion of a particle near a horizon of a spherically symmetric black hole is
shown to possess a universal Lyapunov exponent of a chaos provided by its
surface gravity. To probe the horizon, we introduce electromagnetic or scalar
force to the particle so that it does not fall into the horizon. There appears
an unstable maximum of the total potential where the evaluated maximal Lyapunov
exponent is found to be independent of the external forces and the particle
mass. The Lyapunov exponent is universally given by the surface gravity of the
black hole. Unless there are other sources of a chaos, the Lyapunov exponent is
subject to an inequality $\lambda \leq 2\pi T_{\rm BH}/\hbar$, which is
identical to the bound recently discovered by Maldacena, Shenker and Stanford.
| [
{
"created": "Wed, 19 Oct 2016 15:44:35 GMT",
"version": "v1"
},
{
"created": "Tue, 1 Nov 2016 01:18:58 GMT",
"version": "v2"
}
] | 2017-01-11 | [
[
"Hashimoto",
"Koji",
""
],
[
"Tanahashi",
"Norihiro",
""
]
] | Motion of a particle near a horizon of a spherically symmetric black hole is shown to possess a universal Lyapunov exponent of a chaos provided by its surface gravity. To probe the horizon, we introduce electromagnetic or scalar force to the particle so that it does not fall into the horizon. There appears an unstable maximum of the total potential where the evaluated maximal Lyapunov exponent is found to be independent of the external forces and the particle mass. The Lyapunov exponent is universally given by the surface gravity of the black hole. Unless there are other sources of a chaos, the Lyapunov exponent is subject to an inequality $\lambda \leq 2\pi T_{\rm BH}/\hbar$, which is identical to the bound recently discovered by Maldacena, Shenker and Stanford. |
2310.03091 | Daile Osorio-Roig | Daile Osorio-Roig, Lazaro J. Gonzalez-Soler, Christian Rathgeb,
Christoph Busch | Privacy-preserving Multi-biometric Indexing based on Frequent Binary
Patterns | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | The development of large-scale identification systems that ensure the privacy
protection of enrolled subjects represents a major challenge. Biometric
deployments that provide interoperability and usability by including efficient
multi-biometric solutions are a recent requirement. In the context of privacy
protection, several template protection schemes have been proposed in the past.
However, these schemes seem inadequate for indexing (workload reduction) in
biometric identification systems. More specifically, they have been used in
identification systems that perform exhaustive searches, leading to a
degradation of computational efficiency. To overcome these limitations, we
propose an efficient privacy-preserving multi-biometric identification system
that retrieves protected deep cancelable templates and is agnostic with respect
to biometric characteristics and biometric template protection schemes. To this
end, a multi-biometric binning scheme is designed to exploit the low
intra-class variation properties contained in the frequent binary patterns
extracted from different types of biometric characteristics. Experimental
results reported on publicly available databases using state-of-the-art Deep
Neural Network (DNN)-based embedding extractors show that the protected
multi-biometric identification system can reduce the computational workload to
approximately 57\% (indexing up to three types of biometric characteristics)
and 53% (indexing up to two types of biometric characteristics), while
simultaneously improving the biometric performance of the baseline biometric
system at the high-security thresholds. The source code of the proposed
multi-biometric indexing approach together with the composed multi-biometric
dataset, will be made available to the research community once the article is
accepted.
| [
{
"created": "Wed, 4 Oct 2023 18:18:24 GMT",
"version": "v1"
}
] | 2023-10-06 | [
[
"Osorio-Roig",
"Daile",
""
],
[
"Gonzalez-Soler",
"Lazaro J.",
""
],
[
"Rathgeb",
"Christian",
""
],
[
"Busch",
"Christoph",
""
]
] | The development of large-scale identification systems that ensure the privacy protection of enrolled subjects represents a major challenge. Biometric deployments that provide interoperability and usability by including efficient multi-biometric solutions are a recent requirement. In the context of privacy protection, several template protection schemes have been proposed in the past. However, these schemes seem inadequate for indexing (workload reduction) in biometric identification systems. More specifically, they have been used in identification systems that perform exhaustive searches, leading to a degradation of computational efficiency. To overcome these limitations, we propose an efficient privacy-preserving multi-biometric identification system that retrieves protected deep cancelable templates and is agnostic with respect to biometric characteristics and biometric template protection schemes. To this end, a multi-biometric binning scheme is designed to exploit the low intra-class variation properties contained in the frequent binary patterns extracted from different types of biometric characteristics. Experimental results reported on publicly available databases using state-of-the-art Deep Neural Network (DNN)-based embedding extractors show that the protected multi-biometric identification system can reduce the computational workload to approximately 57\% (indexing up to three types of biometric characteristics) and 53% (indexing up to two types of biometric characteristics), while simultaneously improving the biometric performance of the baseline biometric system at the high-security thresholds. The source code of the proposed multi-biometric indexing approach together with the composed multi-biometric dataset, will be made available to the research community once the article is accepted. |
2010.10468 | Sherif Abdulatif | Sherif Abdulatif, Karim Armanious, Jayasankar T. Sajeev, Karim
Guirguis, Bin Yang | Investigating Cross-Domain Losses for Speech Enhancement | 5 pages, 3 figures and 1 table | null | null | null | cs.SD cs.LG eess.AS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent years have seen a surge in the number of available frameworks for
speech enhancement (SE) and recognition. Whether model-based or constructed via
deep learning, these frameworks often rely in isolation on either time-domain
signals or time-frequency (TF) representations of speech data. In this study,
we investigate the advantages of each set of approaches by separately examining
their impact on speech intelligibility and quality. Furthermore, we combine the
fragmented benefits of time-domain and TF speech representations by introducing
two new cross-domain SE frameworks. A quantitative comparative analysis against
recent model-based and deep learning SE approaches is performed to illustrate
the merit of the proposed frameworks.
| [
{
"created": "Tue, 20 Oct 2020 17:28:07 GMT",
"version": "v1"
},
{
"created": "Sun, 30 May 2021 01:56:54 GMT",
"version": "v2"
}
] | 2021-06-01 | [
[
"Abdulatif",
"Sherif",
""
],
[
"Armanious",
"Karim",
""
],
[
"Sajeev",
"Jayasankar T.",
""
],
[
"Guirguis",
"Karim",
""
],
[
"Yang",
"Bin",
""
]
] | Recent years have seen a surge in the number of available frameworks for speech enhancement (SE) and recognition. Whether model-based or constructed via deep learning, these frameworks often rely in isolation on either time-domain signals or time-frequency (TF) representations of speech data. In this study, we investigate the advantages of each set of approaches by separately examining their impact on speech intelligibility and quality. Furthermore, we combine the fragmented benefits of time-domain and TF speech representations by introducing two new cross-domain SE frameworks. A quantitative comparative analysis against recent model-based and deep learning SE approaches is performed to illustrate the merit of the proposed frameworks. |
1207.4258 | Lin Chen | Lin Chen, Athanasios V. Vasilakos | Joint Rate Adaptation and Medium Access in Wireless LANs: a
Non-cooperative Game Theoretic Perspective | null | null | null | null | cs.GT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Wireless local area networks (WLANs) based on IEEE 802.11 standards are
becoming ubiquitous today and typically support multiple data rates. In such
multi-rate WLANs, distributed medium access and rate adaptation are two key
elements to achieve efficient radio resource utilization, especially in
non-cooperative environments. In this paper, we present an analytical study on
the non-cooperative multi-rate WLANs composed of selfish users jointly
adjusting their data rate and contention window size at the medium access level
to maximize their own throughput, irrespective of the impact of their selfish
behaviors on overall system performance. Specifically, we develop an adapted
Tit-For-Tat (TFT) strategy to guide the system to an efficient equilibrium in
non-cooperative environments. We model the interactions among selfish users
under the adapted TFT framework as a non-cooperative joint medium access and
rate adaptation game. A systematic analysis is conducted on the structural
properties of the game to provide insights on the interaction between rate
adaptation and 802.11 medium access control in a competitive setting. We show
that the game has multiple equilibria, which, after the equilibrium refinement
process that we develop, reduce to a unique efficient equilibrium. We further
develop a distributed algorithm to achieve this equilibrium and demonstrate
that the equilibrium achieves the performance very close to the system optimum
in a social perspective.
| [
{
"created": "Wed, 18 Jul 2012 04:04:35 GMT",
"version": "v1"
}
] | 2012-07-19 | [
[
"Chen",
"Lin",
""
],
[
"Vasilakos",
"Athanasios V.",
""
]
] | Wireless local area networks (WLANs) based on IEEE 802.11 standards are becoming ubiquitous today and typically support multiple data rates. In such multi-rate WLANs, distributed medium access and rate adaptation are two key elements to achieve efficient radio resource utilization, especially in non-cooperative environments. In this paper, we present an analytical study on the non-cooperative multi-rate WLANs composed of selfish users jointly adjusting their data rate and contention window size at the medium access level to maximize their own throughput, irrespective of the impact of their selfish behaviors on overall system performance. Specifically, we develop an adapted Tit-For-Tat (TFT) strategy to guide the system to an efficient equilibrium in non-cooperative environments. We model the interactions among selfish users under the adapted TFT framework as a non-cooperative joint medium access and rate adaptation game. A systematic analysis is conducted on the structural properties of the game to provide insights on the interaction between rate adaptation and 802.11 medium access control in a competitive setting. We show that the game has multiple equilibria, which, after the equilibrium refinement process that we develop, reduce to a unique efficient equilibrium. We further develop a distributed algorithm to achieve this equilibrium and demonstrate that the equilibrium achieves the performance very close to the system optimum in a social perspective. |
2204.03508 | Zhihan Zhang | Zhihan Zhang, Wenhao Yu, Mengxia Yu, Zhichun Guo, Meng Jiang | A Survey of Multi-task Learning in Natural Language Processing:
Regarding Task Relatedness and Training Methods | Accepted to EACL 2023 as regular long paper | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | Multi-task learning (MTL) has become increasingly popular in natural language
processing (NLP) because it improves the performance of related tasks by
exploiting their commonalities and differences. Nevertheless, it is still not
understood very well how multi-task learning can be implemented based on the
relatedness of training tasks. In this survey, we review recent advances of
multi-task learning methods in NLP, with the aim of summarizing them into two
general multi-task training methods based on their task relatedness: (i) joint
training and (ii) multi-step training. We present examples in various NLP
downstream applications, summarize the task relationships and discuss future
directions of this promising topic.
| [
{
"created": "Thu, 7 Apr 2022 15:22:19 GMT",
"version": "v1"
},
{
"created": "Tue, 14 Feb 2023 19:58:57 GMT",
"version": "v2"
}
] | 2023-02-16 | [
[
"Zhang",
"Zhihan",
""
],
[
"Yu",
"Wenhao",
""
],
[
"Yu",
"Mengxia",
""
],
[
"Guo",
"Zhichun",
""
],
[
"Jiang",
"Meng",
""
]
] | Multi-task learning (MTL) has become increasingly popular in natural language processing (NLP) because it improves the performance of related tasks by exploiting their commonalities and differences. Nevertheless, it is still not understood very well how multi-task learning can be implemented based on the relatedness of training tasks. In this survey, we review recent advances of multi-task learning methods in NLP, with the aim of summarizing them into two general multi-task training methods based on their task relatedness: (i) joint training and (ii) multi-step training. We present examples in various NLP downstream applications, summarize the task relationships and discuss future directions of this promising topic. |
2007.06402 | Raphael Achddou | Rapha\"el Achddou, J.Matias di Martino, Guillermo Sapiro | Nested Learning For Multi-Granular Tasks | null | null | null | null | cs.CV cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Standard deep neural networks (DNNs) are commonly trained in an end-to-end
fashion for specific tasks such as object recognition, face identification, or
character recognition, among many examples. This specificity often leads to
overconfident models that generalize poorly to samples that are not from the
original training distribution. Moreover, such standard DNNs do not allow to
leverage information from heterogeneously annotated training data, where for
example, labels may be provided with different levels of granularity.
Furthermore, DNNs do not produce results with simultaneous different levels of
confidence for different levels of detail, they are most commonly an all or
nothing approach. To address these challenges, we introduce the concept of
nested learning: how to obtain a hierarchical representation of the input such
that a coarse label can be extracted first, and sequentially refine this
representation, if the sample permits, to obtain successively refined
predictions, all of them with the corresponding confidence. We explicitly
enforce this behavior by creating a sequence of nested information bottlenecks.
Looking at the problem of nested learning from an information theory
perspective, we design a network topology with two important properties. First,
a sequence of low dimensional (nested) feature embeddings are enforced. Then we
show how the explicit combination of nested outputs can improve both the
robustness and the accuracy of finer predictions. Experimental results on
Cifar-10, Cifar-100, MNIST, Fashion-MNIST, Dbpedia, and Plantvillage
demonstrate that nested learning outperforms the same network trained in the
standard end-to-end fashion.
| [
{
"created": "Mon, 13 Jul 2020 14:27:14 GMT",
"version": "v1"
}
] | 2020-07-14 | [
[
"Achddou",
"Raphaël",
""
],
[
"di Martino",
"J. Matias",
""
],
[
"Sapiro",
"Guillermo",
""
]
] | Standard deep neural networks (DNNs) are commonly trained in an end-to-end fashion for specific tasks such as object recognition, face identification, or character recognition, among many examples. This specificity often leads to overconfident models that generalize poorly to samples that are not from the original training distribution. Moreover, such standard DNNs do not allow to leverage information from heterogeneously annotated training data, where for example, labels may be provided with different levels of granularity. Furthermore, DNNs do not produce results with simultaneous different levels of confidence for different levels of detail, they are most commonly an all or nothing approach. To address these challenges, we introduce the concept of nested learning: how to obtain a hierarchical representation of the input such that a coarse label can be extracted first, and sequentially refine this representation, if the sample permits, to obtain successively refined predictions, all of them with the corresponding confidence. We explicitly enforce this behavior by creating a sequence of nested information bottlenecks. Looking at the problem of nested learning from an information theory perspective, we design a network topology with two important properties. First, a sequence of low dimensional (nested) feature embeddings are enforced. Then we show how the explicit combination of nested outputs can improve both the robustness and the accuracy of finer predictions. Experimental results on Cifar-10, Cifar-100, MNIST, Fashion-MNIST, Dbpedia, and Plantvillage demonstrate that nested learning outperforms the same network trained in the standard end-to-end fashion. |
1904.02755 | Soham Ghosh | Soham Ghosh, Anuva Agarwal, Zarana Parekh, Alexander Hauptmann | ExCL: Extractive Clip Localization Using Natural Language Descriptions | Accepted at NAACL 2019, Short Paper | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The task of retrieving clips within videos based on a given natural language
query requires cross-modal reasoning over multiple frames. Prior approaches
such as sliding window classifiers are inefficient, while text-clip similarity
driven ranking-based approaches such as segment proposal networks are far more
complicated. In order to select the most relevant video clip corresponding to
the given text description, we propose a novel extractive approach that
predicts the start and end frames by leveraging cross-modal interactions
between the text and video - this removes the need to retrieve and re-rank
multiple proposal segments. Using recurrent networks we encode the two
modalities into a joint representation which is then used in different variants
of start-end frame predictor networks. Through extensive experimentation and
ablative analysis, we demonstrate that our simple and elegant approach
significantly outperforms state of the art on two datasets and has comparable
performance on a third.
| [
{
"created": "Thu, 4 Apr 2019 19:17:04 GMT",
"version": "v1"
}
] | 2019-04-08 | [
[
"Ghosh",
"Soham",
""
],
[
"Agarwal",
"Anuva",
""
],
[
"Parekh",
"Zarana",
""
],
[
"Hauptmann",
"Alexander",
""
]
] | The task of retrieving clips within videos based on a given natural language query requires cross-modal reasoning over multiple frames. Prior approaches such as sliding window classifiers are inefficient, while text-clip similarity driven ranking-based approaches such as segment proposal networks are far more complicated. In order to select the most relevant video clip corresponding to the given text description, we propose a novel extractive approach that predicts the start and end frames by leveraging cross-modal interactions between the text and video - this removes the need to retrieve and re-rank multiple proposal segments. Using recurrent networks we encode the two modalities into a joint representation which is then used in different variants of start-end frame predictor networks. Through extensive experimentation and ablative analysis, we demonstrate that our simple and elegant approach significantly outperforms state of the art on two datasets and has comparable performance on a third. |
2009.08716 | Zhengjie Yang | Zhengjie Yang, Wei Bao, Dong Yuan, Nguyen H. Tran, and Albert Y.
Zomaya | Federated Learning with Nesterov Accelerated Gradient | publised in TPDS. 18 pages, 6 figures | null | 10.1109/TPDS.2022.3206480 | null | cs.LG cs.DC stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Federated learning (FL) is a fast-developing technique that allows multiple
workers to train a global model based on a distributed dataset. Conventional FL
(FedAvg) employs gradient descent algorithm, which may not be efficient enough.
Momentum is able to improve the situation by adding an additional momentum step
to accelerate the convergence and has demonstrated its benefits in both
centralized and FL environments. It is well-known that Nesterov Accelerated
Gradient (NAG) is a more advantageous form of momentum, but it is not clear how
to quantify the benefits of NAG in FL so far. This motives us to propose
FedNAG, which employs NAG in each worker as well as NAG momentum and model
aggregation in the aggregator. We provide a detailed convergence analysis of
FedNAG and compare it with FedAvg. Extensive experiments based on real-world
datasets and trace-driven simulation are conducted, demonstrating that FedNAG
increases the learning accuracy by 3-24% and decreases the total training time
by 11-70% compared with the benchmarks under a wide range of settings.
| [
{
"created": "Fri, 18 Sep 2020 09:38:11 GMT",
"version": "v1"
},
{
"created": "Wed, 26 Oct 2022 02:46:51 GMT",
"version": "v2"
}
] | 2022-10-27 | [
[
"Yang",
"Zhengjie",
""
],
[
"Bao",
"Wei",
""
],
[
"Yuan",
"Dong",
""
],
[
"Tran",
"Nguyen H.",
""
],
[
"Zomaya",
"Albert Y.",
""
]
] | Federated learning (FL) is a fast-developing technique that allows multiple workers to train a global model based on a distributed dataset. Conventional FL (FedAvg) employs gradient descent algorithm, which may not be efficient enough. Momentum is able to improve the situation by adding an additional momentum step to accelerate the convergence and has demonstrated its benefits in both centralized and FL environments. It is well-known that Nesterov Accelerated Gradient (NAG) is a more advantageous form of momentum, but it is not clear how to quantify the benefits of NAG in FL so far. This motives us to propose FedNAG, which employs NAG in each worker as well as NAG momentum and model aggregation in the aggregator. We provide a detailed convergence analysis of FedNAG and compare it with FedAvg. Extensive experiments based on real-world datasets and trace-driven simulation are conducted, demonstrating that FedNAG increases the learning accuracy by 3-24% and decreases the total training time by 11-70% compared with the benchmarks under a wide range of settings. |
0909.2377 | Nevin Vunka Jungum | Soumaya Zirari, Philippe Canalda and Francois Spies | Geometric and Signal Strength Dilution of Precision (DoP)Wi-Fi | International Journal of Computer Science Issues (IJCSI), Volume 3,
pp35-44, August 2009 | S.Zirari,P. Canalda and F.Spies, " Geometric and Signal Strength
Dilution of Precision (DoP)Wi-Fi", International Journal of Computer Science
Issues (IJCSI), Volume 3, pp35-44, August 2009 | null | null | cs.NI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The democratization of wireless networks combined to the emergence of mobile
devices increasingly autonomous and efficient lead to new services. Positioning
services become overcrowded. Accuracy is the main quality criteria in
positioning. But to better appreciate this one a coefficient is needed. In this
paper we present Geometric and Signal Strength Dilution of Precision (DOP) for
positioning systems based on Wi-Fi and Signal Strength measurements.
| [
{
"created": "Sat, 12 Sep 2009 22:24:52 GMT",
"version": "v1"
}
] | 2009-09-15 | [
[
"Zirari",
"Soumaya",
""
],
[
"Canalda",
"Philippe",
""
],
[
"Spies",
"Francois",
""
]
] | The democratization of wireless networks combined to the emergence of mobile devices increasingly autonomous and efficient lead to new services. Positioning services become overcrowded. Accuracy is the main quality criteria in positioning. But to better appreciate this one a coefficient is needed. In this paper we present Geometric and Signal Strength Dilution of Precision (DOP) for positioning systems based on Wi-Fi and Signal Strength measurements. |
1710.01416 | Saed Khawaldeh | Vu Hoang Minh, Tajwar Abrar Aleef, Usama Pervaiz, Yeman Brhane Hagos,
Saed Khawaldeh | Smoothness-based Edge Detection using Low-SNR Camera for Robot
Navigation | null | null | null | null | cs.CV cs.RO stat.AP stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In the emerging advancement in the branch of autonomous robotics, the ability
of a robot to efficiently localize and construct maps of its surrounding is
crucial. This paper deals with utilizing thermal-infrared cameras, as opposed
to conventional cameras as the primary sensor to capture images of the robot's
surroundings. For localization, the images need to be further processed before
feeding them to a navigational system. The main motivation of this paper was to
develop an edge detection methodology capable of utilizing the low-SNR poor
output from such a thermal camera and effectively detect smooth edges of the
surrounding environment. The enhanced edge detector proposed in this paper
takes the raw image from the thermal sensor, denoises the images, applies Canny
edge detection followed by CSS method. The edges are ranked to remove any noise
and only edges of the highest rank are kept. Then, the broken edges are linked
by computing edge metrics and a smooth edge of the surrounding is displayed in
a binary image. Several comparisons are also made in the paper between the
proposed technique and the existing techniques.
| [
{
"created": "Tue, 3 Oct 2017 22:48:41 GMT",
"version": "v1"
}
] | 2017-10-05 | [
[
"Minh",
"Vu Hoang",
""
],
[
"Aleef",
"Tajwar Abrar",
""
],
[
"Pervaiz",
"Usama",
""
],
[
"Hagos",
"Yeman Brhane",
""
],
[
"Khawaldeh",
"Saed",
""
]
] | In the emerging advancement in the branch of autonomous robotics, the ability of a robot to efficiently localize and construct maps of its surrounding is crucial. This paper deals with utilizing thermal-infrared cameras, as opposed to conventional cameras as the primary sensor to capture images of the robot's surroundings. For localization, the images need to be further processed before feeding them to a navigational system. The main motivation of this paper was to develop an edge detection methodology capable of utilizing the low-SNR poor output from such a thermal camera and effectively detect smooth edges of the surrounding environment. The enhanced edge detector proposed in this paper takes the raw image from the thermal sensor, denoises the images, applies Canny edge detection followed by CSS method. The edges are ranked to remove any noise and only edges of the highest rank are kept. Then, the broken edges are linked by computing edge metrics and a smooth edge of the surrounding is displayed in a binary image. Several comparisons are also made in the paper between the proposed technique and the existing techniques. |
2303.18119 | Juan M. Gandarias | Luca Fortini (1,2), Mattia Leonori (1), Juan M. Gandarias (1), Elena
de Momi (2), Arash Ajoudani (1) ((1) Human-Robot Interfaces and Interaction,
Istituto Italiano di Tecnologia, Genoa, Italy (2) Department of Electronics,
Information and Bioengineering, Politecnico di Milano, Milan, Italy) | Markerless 3D human pose tracking through multiple cameras and AI:
Enabling high accuracy, robustness, and real-time performance | 19 pages, 7 figures | null | null | null | cs.CV cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Tracking 3D human motion in real-time is crucial for numerous applications
across many fields. Traditional approaches involve attaching artificial
fiducial objects or sensors to the body, limiting their usability and
comfort-of-use and consequently narrowing their application fields. Recent
advances in Artificial Intelligence (AI) have allowed for markerless solutions.
However, most of these methods operate in 2D, while those providing 3D
solutions compromise accuracy and real-time performance. To address this
challenge and unlock the potential of visual pose estimation methods in
real-world scenarios, we propose a markerless framework that combines
multi-camera views and 2D AI-based pose estimation methods to track 3D human
motion. Our approach integrates a Weighted Least Square (WLS) algorithm that
computes 3D human motion from multiple 2D pose estimations provided by an
AI-driven method. The method is integrated within the Open-VICO framework
allowing simulation and real-world execution. Several experiments have been
conducted, which have shown high accuracy and real-time performance,
demonstrating the high level of readiness for real-world applications and the
potential to revolutionize human motion capture.
| [
{
"created": "Fri, 31 Mar 2023 15:06:50 GMT",
"version": "v1"
}
] | 2023-04-03 | [
[
"Fortini",
"Luca",
""
],
[
"Leonori",
"Mattia",
""
],
[
"Gandarias",
"Juan M.",
""
],
[
"de Momi",
"Elena",
""
],
[
"Ajoudani",
"Arash",
""
]
] | Tracking 3D human motion in real-time is crucial for numerous applications across many fields. Traditional approaches involve attaching artificial fiducial objects or sensors to the body, limiting their usability and comfort-of-use and consequently narrowing their application fields. Recent advances in Artificial Intelligence (AI) have allowed for markerless solutions. However, most of these methods operate in 2D, while those providing 3D solutions compromise accuracy and real-time performance. To address this challenge and unlock the potential of visual pose estimation methods in real-world scenarios, we propose a markerless framework that combines multi-camera views and 2D AI-based pose estimation methods to track 3D human motion. Our approach integrates a Weighted Least Square (WLS) algorithm that computes 3D human motion from multiple 2D pose estimations provided by an AI-driven method. The method is integrated within the Open-VICO framework allowing simulation and real-world execution. Several experiments have been conducted, which have shown high accuracy and real-time performance, demonstrating the high level of readiness for real-world applications and the potential to revolutionize human motion capture. |
1002.1846 | Goran Duplancic | G. Duplancic, D. Glavan, H. Stefancic | Probability distribution of the vacuum energy density | 5 pages, 2 figures, revised version to appear in Phys.Rev.D | Phys.Rev.D82:125008,2010 | 10.1103/PhysRevD.82.125008 | null | hep-th astro-ph.CO gr-qc hep-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | As the vacuum state of a quantum field is not an eigenstate of the
Hamiltonian density, the vacuum energy density can be represented as a random
variable. We present an analytical calculation of the probability distribution
of the vacuum energy density for real and complex massless scalar fields in
Minkowski space. The obtained probability distributions are broad and the
vacuum expectation value of the Hamiltonian density is not fully representative
of the vacuum energy density.
| [
{
"created": "Tue, 9 Feb 2010 13:13:29 GMT",
"version": "v1"
},
{
"created": "Fri, 26 Nov 2010 13:52:49 GMT",
"version": "v2"
}
] | 2010-12-24 | [
[
"Duplancic",
"G.",
""
],
[
"Glavan",
"D.",
""
],
[
"Stefancic",
"H.",
""
]
] | As the vacuum state of a quantum field is not an eigenstate of the Hamiltonian density, the vacuum energy density can be represented as a random variable. We present an analytical calculation of the probability distribution of the vacuum energy density for real and complex massless scalar fields in Minkowski space. The obtained probability distributions are broad and the vacuum expectation value of the Hamiltonian density is not fully representative of the vacuum energy density. |
2210.15042 | Mohammadreza Ebrahimi | Rouzbeh Behnia, Mohamamdreza Ebrahimi, Jason Pacheco, Balaji
Padmanabhan | Privately Fine-Tuning Large Language Models with Differential Privacy | Publised at IEEE ICDM Workshop on Machine Learning for Cybersecurity
(MLC) 2022 | 2022 IEEE International Conference on Data Mining Workshops
(ICDMW), pp. 560-566 | 10.1109/ICDMW58026.2022.00078 | null | cs.CR cs.CL | http://creativecommons.org/licenses/by/4.0/ | Pre-trained Large Language Models (LLMs) are an integral part of modern AI
that have led to breakthrough performances in complex AI tasks. Major AI
companies with expensive infrastructures are able to develop and train these
large models with billions and millions of parameters from scratch. Third
parties, researchers, and practitioners are increasingly adopting these
pre-trained models and fine-tuning them on their private data to accomplish
their downstream AI tasks. However, it has been shown that an adversary can
extract/reconstruct the exact training samples from these LLMs, which can lead
to revealing personally identifiable information. The issue has raised deep
concerns about the privacy of LLMs. Differential privacy (DP) provides a
rigorous framework that allows adding noise in the process of training or
fine-tuning LLMs such that extracting the training data becomes infeasible
(i.e., with a cryptographically small success probability). While the
theoretical privacy guarantees offered in most extant studies assume learning
models from scratch through many training iterations in an asymptotic setting,
this assumption does not hold in fine-tuning scenarios in which the number of
training iterations is significantly smaller. To address the gap, we present
\ewtune, a DP framework for fine-tuning LLMs based on Edgeworth accountant with
finite-sample privacy guarantees. Our results across four well-established
natural language understanding (NLU) tasks show that while \ewtune~adds privacy
guarantees to LLM fine-tuning process, it directly contributes to decreasing
the induced noise to up to 5.6\% and improves the state-of-the-art LLMs
performance by up to 1.1\% across all NLU tasks. We have open-sourced our
implementations for wide adoption and public testing purposes.
| [
{
"created": "Wed, 26 Oct 2022 21:18:31 GMT",
"version": "v1"
},
{
"created": "Fri, 17 Mar 2023 00:55:42 GMT",
"version": "v2"
},
{
"created": "Mon, 20 Mar 2023 01:33:23 GMT",
"version": "v3"
}
] | 2023-05-02 | [
[
"Behnia",
"Rouzbeh",
""
],
[
"Ebrahimi",
"Mohamamdreza",
""
],
[
"Pacheco",
"Jason",
""
],
[
"Padmanabhan",
"Balaji",
""
]
] | Pre-trained Large Language Models (LLMs) are an integral part of modern AI that have led to breakthrough performances in complex AI tasks. Major AI companies with expensive infrastructures are able to develop and train these large models with billions and millions of parameters from scratch. Third parties, researchers, and practitioners are increasingly adopting these pre-trained models and fine-tuning them on their private data to accomplish their downstream AI tasks. However, it has been shown that an adversary can extract/reconstruct the exact training samples from these LLMs, which can lead to revealing personally identifiable information. The issue has raised deep concerns about the privacy of LLMs. Differential privacy (DP) provides a rigorous framework that allows adding noise in the process of training or fine-tuning LLMs such that extracting the training data becomes infeasible (i.e., with a cryptographically small success probability). While the theoretical privacy guarantees offered in most extant studies assume learning models from scratch through many training iterations in an asymptotic setting, this assumption does not hold in fine-tuning scenarios in which the number of training iterations is significantly smaller. To address the gap, we present \ewtune, a DP framework for fine-tuning LLMs based on Edgeworth accountant with finite-sample privacy guarantees. Our results across four well-established natural language understanding (NLU) tasks show that while \ewtune~adds privacy guarantees to LLM fine-tuning process, it directly contributes to decreasing the induced noise to up to 5.6\% and improves the state-of-the-art LLMs performance by up to 1.1\% across all NLU tasks. We have open-sourced our implementations for wide adoption and public testing purposes. |
2301.10903 | Tom Steudtner | Ian Jack, Hugh Osborn, Tom Steudtner | Explorations in Scalar Fermion Theories: $\beta$-functions,
Supersymmetry and Fixed Points | 76 pages, 3 external figures | null | null | DO-TH 22/06 | hep-th hep-ph | http://creativecommons.org/licenses/by/4.0/ | Results for $\beta$-functions and anomalous dimensions in general scalar
fermion theories are presented to three loops. Various constraints on the
individual coefficients for each diagram following from supersymmetry are
analysed. The results are used to discuss potential fixed points in the
$\varepsilon$-expansion for scalar fermion theories, with arbitrary numbers of
scalar fields, and where there are just two scalar couplings and one Yukawa
coupling. For different examples the fixed points follow a similar pattern as
the numbers of fermions is varied. For diagrams with subdivergences there are
extensive consistency constraints arising from the existence of a perturbative
$a$-function and these are analysed in detail. Further arbitrary scheme
variations which preserve the form of $\beta$ functions and anomalous
dimensions in terms of 1PI diagrams are also discussed. The existence of linear
and quadratic scheme invariants is demonstrated and the consistency condition
are shown to be expressible in terms of these invariants.
| [
{
"created": "Thu, 26 Jan 2023 02:23:13 GMT",
"version": "v1"
},
{
"created": "Thu, 4 Jan 2024 11:48:31 GMT",
"version": "v2"
}
] | 2024-01-05 | [
[
"Jack",
"Ian",
""
],
[
"Osborn",
"Hugh",
""
],
[
"Steudtner",
"Tom",
""
]
] | Results for $\beta$-functions and anomalous dimensions in general scalar fermion theories are presented to three loops. Various constraints on the individual coefficients for each diagram following from supersymmetry are analysed. The results are used to discuss potential fixed points in the $\varepsilon$-expansion for scalar fermion theories, with arbitrary numbers of scalar fields, and where there are just two scalar couplings and one Yukawa coupling. For different examples the fixed points follow a similar pattern as the numbers of fermions is varied. For diagrams with subdivergences there are extensive consistency constraints arising from the existence of a perturbative $a$-function and these are analysed in detail. Further arbitrary scheme variations which preserve the form of $\beta$ functions and anomalous dimensions in terms of 1PI diagrams are also discussed. The existence of linear and quadratic scheme invariants is demonstrated and the consistency condition are shown to be expressible in terms of these invariants. |
0806.3704 | Matteo Beccaria | Matteo Beccaria | The generalized scaling function of AdS/CFT and semiclassical string
theory | 31 pages, 8 eps figures | JHEP 0807:082,2008 | 10.1088/1126-6708/2008/07/082 | null | hep-th | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recently, Freyhult, Rej and Staudacher (FRS) proposed an integral equation
determining the leading logarithmic term of the anomalous dimension of sl(2)
twist-operators in N=4 SYM for large Lorentz spin M and twist L at fixed j =
L/log(M). We discuss the large j limit of the FRS equation. This limit can be
matched with the {\em fast long string} limit of AdS_5 X S^5 superstring
perturbation theory at all couplings. In particular, a certain part of the
classical and one-loop string result is known to be protected and can be
computed in the weakly coupled large-j limit of the FRS equation. We present
various analytical and numerical results supporting agreement at one and two
loops in the gauge theory.
| [
{
"created": "Mon, 23 Jun 2008 15:16:55 GMT",
"version": "v1"
}
] | 2011-06-02 | [
[
"Beccaria",
"Matteo",
""
]
] | Recently, Freyhult, Rej and Staudacher (FRS) proposed an integral equation determining the leading logarithmic term of the anomalous dimension of sl(2) twist-operators in N=4 SYM for large Lorentz spin M and twist L at fixed j = L/log(M). We discuss the large j limit of the FRS equation. This limit can be matched with the {\em fast long string} limit of AdS_5 X S^5 superstring perturbation theory at all couplings. In particular, a certain part of the classical and one-loop string result is known to be protected and can be computed in the weakly coupled large-j limit of the FRS equation. We present various analytical and numerical results supporting agreement at one and two loops in the gauge theory. |
2308.05038 | Himarsha R Jayanetti | Himarsha R. Jayanetti, Erika Frydenlund, Michele C. Weigle | Xenophobic Events vs. Refugee Population -- Using GDELT to Identify
Countries with Disproportionate Coverage | 10 pages, 2 figures, accepted as a Working Paper at SBP-BRiMS 2023.
arXiv admin note: text overlap with arXiv:2305.01708 | null | null | null | cs.CY | http://creativecommons.org/licenses/by-nc-sa/4.0/ | In this preliminary study, we used the Global Database of Events, Language,
and Tone (GDELT) database to examine xenophobic events reported in the media
during 2022. We collected a dataset of 2,778 unique events and created a
choropleth map illustrating the frequency of events scaled by the refugee
population's proportion in each host country. We identified the top 10
countries with the highest scaled event frequencies among those with more than
50,000 refugees. Contrary to the belief that hosting a significant number of
forced migrants results in higher xenophobic incidents, our findings indicate a
potential connection to political factors. We also categorized the 20 root
event codes in the CAMEO event data as either "Direct" or "Indirect". Almost
90% of the events related to refugees in 2022 were classified as "Indirect".
| [
{
"created": "Wed, 9 Aug 2023 16:10:05 GMT",
"version": "v1"
}
] | 2023-08-10 | [
[
"Jayanetti",
"Himarsha R.",
""
],
[
"Frydenlund",
"Erika",
""
],
[
"Weigle",
"Michele C.",
""
]
] | In this preliminary study, we used the Global Database of Events, Language, and Tone (GDELT) database to examine xenophobic events reported in the media during 2022. We collected a dataset of 2,778 unique events and created a choropleth map illustrating the frequency of events scaled by the refugee population's proportion in each host country. We identified the top 10 countries with the highest scaled event frequencies among those with more than 50,000 refugees. Contrary to the belief that hosting a significant number of forced migrants results in higher xenophobic incidents, our findings indicate a potential connection to political factors. We also categorized the 20 root event codes in the CAMEO event data as either "Direct" or "Indirect". Almost 90% of the events related to refugees in 2022 were classified as "Indirect". |
1711.00296 | Linus Wulff | Linus Wulff | Classifying integrable symmetric space strings via factorized scattering | 17 pages; v2: Improvements to sec 1, results now summarized in Tab 1.
Matches published version | null | 10.1007/JHEP02(2018)106 | null | hep-th | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | All symmetric space $AdS_n$ solutions of type II supergravity have recently
been found for $n>2$. For the supersymmetric solutions (and their T-duals) it
is known that the Green-Schwarz string is classically integrable. We complete
the classification by ruling out integrability for the remaining
non-supersymmetric solutions. This is achieved by showing that tree-level
scattering on the worldsheet of a GKP or BMN string fails to factorize for
these cases.
| [
{
"created": "Wed, 1 Nov 2017 11:35:41 GMT",
"version": "v1"
},
{
"created": "Mon, 26 Feb 2018 14:04:07 GMT",
"version": "v2"
}
] | 2018-04-04 | [
[
"Wulff",
"Linus",
""
]
] | All symmetric space $AdS_n$ solutions of type II supergravity have recently been found for $n>2$. For the supersymmetric solutions (and their T-duals) it is known that the Green-Schwarz string is classically integrable. We complete the classification by ruling out integrability for the remaining non-supersymmetric solutions. This is achieved by showing that tree-level scattering on the worldsheet of a GKP or BMN string fails to factorize for these cases. |
2008.13478 | Emilio Torrente-Lujan | A. Belhaj, H. Belmahi, M. Benali, W. El Hadri, H. El Moumni, E.
Torrente-Lujan | Shadows of 5D Black Holes from String Theory | null | null | 10.1016/j.physletb.2020.136025 | FISPAC-TH/271-20, UQBAR-TH/314-203 | hep-th gr-qc | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We study the shadow behaviors of five dimensional (5D) black holes embedded
in type IIB superstring/supergravity inspired spacetimes by considering
solutions with and without rotations. Geometrical properties as shapes and
sizes are analyzed in terms of the D3-brane number and the rotation parameter.
Concretely, we find that the shapes are indeed significantly distorted by such
physical parameters and the size of the shadows decreases with the brane or
"color" number and the rotation. Then, we investigate geometrical observables
and energy emission rate aspects.
| [
{
"created": "Mon, 31 Aug 2020 10:43:11 GMT",
"version": "v1"
}
] | 2020-12-23 | [
[
"Belhaj",
"A.",
""
],
[
"Belmahi",
"H.",
""
],
[
"Benali",
"M.",
""
],
[
"Hadri",
"W. El",
""
],
[
"Moumni",
"H. El",
""
],
[
"Torrente-Lujan",
"E.",
""
]
] | We study the shadow behaviors of five dimensional (5D) black holes embedded in type IIB superstring/supergravity inspired spacetimes by considering solutions with and without rotations. Geometrical properties as shapes and sizes are analyzed in terms of the D3-brane number and the rotation parameter. Concretely, we find that the shapes are indeed significantly distorted by such physical parameters and the size of the shadows decreases with the brane or "color" number and the rotation. Then, we investigate geometrical observables and energy emission rate aspects. |
2308.04333 | George Boateng | George Boateng, Jonathan Abrefah Mensah, Kevin Takyi Yeboah, William
Edor, Andrew Kojo Mensah-Onumah, Naafi Dasana Ibrahim, Nana Sam Yeboah | Towards an AI to Win Ghana's National Science and Maths Quiz | 7 pages. Under review at Deep Learning Indaba and Black in AI
Workshop @NeurIPS 2023 | null | null | null | cs.HC cs.CL cs.CY cs.SD eess.AS | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Can an AI win Ghana's National Science and Maths Quiz (NSMQ)? That is the
question we seek to answer in the NSMQ AI project, an open-source project that
is building AI to compete live in the NSMQ and win. The NSMQ is an annual live
science and mathematics competition for senior secondary school students in
Ghana in which 3 teams of 2 students compete by answering questions across
biology, chemistry, physics, and math in 5 rounds over 5 progressive stages
until a winning team is crowned for that year. The NSMQ is an exciting live
quiz competition with interesting technical challenges across speech-to-text,
text-to-speech, question-answering, and human-computer interaction. In this
ongoing work that began in January 2023, we give an overview of the project,
describe each of the teams, progress made thus far, and the next steps toward
our planned launch and debut of the AI in October for NSMQ 2023. An AI that
conquers this grand challenge can have real-world impact on education such as
enabling millions of students across Africa to have one-on-one learning support
from this AI.
| [
{
"created": "Tue, 8 Aug 2023 15:26:58 GMT",
"version": "v1"
}
] | 2023-08-09 | [
[
"Boateng",
"George",
""
],
[
"Mensah",
"Jonathan Abrefah",
""
],
[
"Yeboah",
"Kevin Takyi",
""
],
[
"Edor",
"William",
""
],
[
"Mensah-Onumah",
"Andrew Kojo",
""
],
[
"Ibrahim",
"Naafi Dasana",
""
],
[
"Yeboah",
... | Can an AI win Ghana's National Science and Maths Quiz (NSMQ)? That is the question we seek to answer in the NSMQ AI project, an open-source project that is building AI to compete live in the NSMQ and win. The NSMQ is an annual live science and mathematics competition for senior secondary school students in Ghana in which 3 teams of 2 students compete by answering questions across biology, chemistry, physics, and math in 5 rounds over 5 progressive stages until a winning team is crowned for that year. The NSMQ is an exciting live quiz competition with interesting technical challenges across speech-to-text, text-to-speech, question-answering, and human-computer interaction. In this ongoing work that began in January 2023, we give an overview of the project, describe each of the teams, progress made thus far, and the next steps toward our planned launch and debut of the AI in October for NSMQ 2023. An AI that conquers this grand challenge can have real-world impact on education such as enabling millions of students across Africa to have one-on-one learning support from this AI. |
hep-th/9610176 | null | Anna Tollsten | String Solutions to Supergravity | 9 pages, latex, uses a4.sty, no figures, contribution to the
proceedings of the workshop Gauge Theory, Applied Supersymmetry and Quantum
Gravity, Imperial College, London 1996 | null | 10.1142/9781848160927_0030 | NBI-HE 96-61 | hep-th | null | We find the comlete solution to ten-dimensional supergravity coupled to a
three-form field strength, given the ``standard ansatz" for the fields, and
show that in addition to the well-known elementary and solitonic (heterotic)
string solutions, one of the possibilities is an (unstable) elementary type I
string solution.
| [
{
"created": "Wed, 23 Oct 1996 11:16:59 GMT",
"version": "v1"
}
] | 2016-12-21 | [
[
"Tollsten",
"Anna",
""
]
] | We find the comlete solution to ten-dimensional supergravity coupled to a three-form field strength, given the ``standard ansatz" for the fields, and show that in addition to the well-known elementary and solitonic (heterotic) string solutions, one of the possibilities is an (unstable) elementary type I string solution. |
2405.15223 | Jialong Wu | Jialong Wu, Shaofeng Yin, Ningya Feng, Xu He, Dong Li, Jianye Hao,
Mingsheng Long | iVideoGPT: Interactive VideoGPTs are Scalable World Models | Project website: https://thuml.github.io/iVideoGPT | null | null | null | cs.CV cs.LG cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | World models empower model-based agents to interactively explore, reason, and
plan within imagined environments for real-world decision-making. However, the
high demand for interactivity poses challenges in harnessing recent
advancements in video generative models for developing world models at scale.
This work introduces Interactive VideoGPT (iVideoGPT), a scalable
autoregressive transformer framework that integrates multimodal signals--visual
observations, actions, and rewards--into a sequence of tokens, facilitating an
interactive experience of agents via next-token prediction. iVideoGPT features
a novel compressive tokenization technique that efficiently discretizes
high-dimensional visual observations. Leveraging its scalable architecture, we
are able to pre-train iVideoGPT on millions of human and robotic manipulation
trajectories, establishing a versatile foundation that is adaptable to serve as
interactive world models for a wide range of downstream tasks. These include
action-conditioned video prediction, visual planning, and model-based
reinforcement learning, where iVideoGPT achieves competitive performance
compared with state-of-the-art methods. Our work advances the development of
interactive general world models, bridging the gap between generative video
models and practical model-based reinforcement learning applications.
| [
{
"created": "Fri, 24 May 2024 05:29:12 GMT",
"version": "v1"
},
{
"created": "Sun, 2 Jun 2024 09:44:20 GMT",
"version": "v2"
}
] | 2024-06-04 | [
[
"Wu",
"Jialong",
""
],
[
"Yin",
"Shaofeng",
""
],
[
"Feng",
"Ningya",
""
],
[
"He",
"Xu",
""
],
[
"Li",
"Dong",
""
],
[
"Hao",
"Jianye",
""
],
[
"Long",
"Mingsheng",
""
]
] | World models empower model-based agents to interactively explore, reason, and plan within imagined environments for real-world decision-making. However, the high demand for interactivity poses challenges in harnessing recent advancements in video generative models for developing world models at scale. This work introduces Interactive VideoGPT (iVideoGPT), a scalable autoregressive transformer framework that integrates multimodal signals--visual observations, actions, and rewards--into a sequence of tokens, facilitating an interactive experience of agents via next-token prediction. iVideoGPT features a novel compressive tokenization technique that efficiently discretizes high-dimensional visual observations. Leveraging its scalable architecture, we are able to pre-train iVideoGPT on millions of human and robotic manipulation trajectories, establishing a versatile foundation that is adaptable to serve as interactive world models for a wide range of downstream tasks. These include action-conditioned video prediction, visual planning, and model-based reinforcement learning, where iVideoGPT achieves competitive performance compared with state-of-the-art methods. Our work advances the development of interactive general world models, bridging the gap between generative video models and practical model-based reinforcement learning applications. |
2007.13440 | Niall Twomey | Niall Twomey, Mikhail Fain, Andrey Ponikar, Nadine Sarraf | Towards Multi-Language Recipe Personalisation and Recommendation | 5 tables | Fourteenth ACM Conference on Recommender Systems (RecSys 2020) | 10.1145/3383313.3418478 | null | cs.IR cs.SI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Multi-language recipe personalisation and recommendation is an under-explored
field of information retrieval in academic and production systems. The existing
gaps in our current understanding are numerous, even on fundamental questions
such as whether consistent and high-quality recipe recommendation can be
delivered across languages. In this paper, we introduce the multi-language
recipe recommendation setting and present grounding results that will help to
establish the potential and absolute value of future work in this area. Our
work draws on several billion events from millions of recipes and users from
Arabic, English, Indonesian, Russian, and Spanish. We represent recipes using a
combination of normalised ingredients, standardised skills and image embeddings
obtained without human intervention. In modelling, we take a classical approach
based on optimising an embedded bi-linear user-item metric space towards the
interactions that most strongly elicit cooking intent. For users without
interaction histories, a bespoke content-based cold-start model that predicts
context and recipe affinity is introduced. We show that our approach to
personalisation is stable and easily scales to new languages. A robust
cross-validation campaign is employed and consistently rejects baseline models
and representations, strongly favouring those we propose. Our results are
presented in a language-oriented (as opposed to model-oriented) fashion to
emphasise the language-based goals of this work. We believe that this is the
first large-scale work that comprehensively considers the value and potential
of multi-language recipe recommendation and personalisation as well as
delivering scalable and reliable models.
| [
{
"created": "Mon, 27 Jul 2020 11:26:49 GMT",
"version": "v1"
},
{
"created": "Tue, 18 Aug 2020 10:57:33 GMT",
"version": "v2"
}
] | 2020-08-19 | [
[
"Twomey",
"Niall",
""
],
[
"Fain",
"Mikhail",
""
],
[
"Ponikar",
"Andrey",
""
],
[
"Sarraf",
"Nadine",
""
]
] | Multi-language recipe personalisation and recommendation is an under-explored field of information retrieval in academic and production systems. The existing gaps in our current understanding are numerous, even on fundamental questions such as whether consistent and high-quality recipe recommendation can be delivered across languages. In this paper, we introduce the multi-language recipe recommendation setting and present grounding results that will help to establish the potential and absolute value of future work in this area. Our work draws on several billion events from millions of recipes and users from Arabic, English, Indonesian, Russian, and Spanish. We represent recipes using a combination of normalised ingredients, standardised skills and image embeddings obtained without human intervention. In modelling, we take a classical approach based on optimising an embedded bi-linear user-item metric space towards the interactions that most strongly elicit cooking intent. For users without interaction histories, a bespoke content-based cold-start model that predicts context and recipe affinity is introduced. We show that our approach to personalisation is stable and easily scales to new languages. A robust cross-validation campaign is employed and consistently rejects baseline models and representations, strongly favouring those we propose. Our results are presented in a language-oriented (as opposed to model-oriented) fashion to emphasise the language-based goals of this work. We believe that this is the first large-scale work that comprehensively considers the value and potential of multi-language recipe recommendation and personalisation as well as delivering scalable and reliable models. |
2212.05752 | Chengyu Zheng | Chengyu Zheng, Ning song, Ruoyu Zhang, Lei Huang, Zhiqiang Wei, Jie
Nie (corresponding author) | Scale-Semantic Joint Decoupling Network for Image-text Retrieval in
Remote Sensing | null | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Image-text retrieval in remote sensing aims to provide flexible information
for data analysis and application. In recent years, state-of-the-art methods
are dedicated to ``scale decoupling'' and ``semantic decoupling'' strategies to
further enhance the capability of representation. However, these previous
approaches focus on either the disentangling scale or semantics but ignore
merging these two ideas in a union model, which extremely limits the
performance of cross-modal retrieval models. To address these issues, we
propose a novel Scale-Semantic Joint Decoupling Network (SSJDN) for remote
sensing image-text retrieval. Specifically, we design the Bidirectional Scale
Decoupling (BSD) module, which exploits Salience Feature Extraction (SFE) and
Salience-Guided Suppression (SGS) units to adaptively extract potential
features and suppress cumbersome features at other scales in a bidirectional
pattern to yield different scale clues. Besides, we design the Label-supervised
Semantic Decoupling (LSD) module by leveraging the category semantic labels as
prior knowledge to supervise images and texts probing significant
semantic-related information. Finally, we design a Semantic-guided Triple Loss
(STL), which adaptively generates a constant to adjust the loss function to
improve the probability of matching the same semantic image and text and
shorten the convergence time of the retrieval model. Our proposed SSJDN
outperforms state-of-the-art approaches in numerical experiments conducted on
four benchmark remote sensing datasets.
| [
{
"created": "Mon, 12 Dec 2022 08:02:35 GMT",
"version": "v1"
}
] | 2022-12-13 | [
[
"Zheng",
"Chengyu",
"",
"corresponding author"
],
[
"song",
"Ning",
"",
"corresponding author"
],
[
"Zhang",
"Ruoyu",
"",
"corresponding author"
],
[
"Huang",
"Lei",
"",
"corresponding author"
],
[
"Wei",
"Zhiqiang",
"",
... | Image-text retrieval in remote sensing aims to provide flexible information for data analysis and application. In recent years, state-of-the-art methods are dedicated to ``scale decoupling'' and ``semantic decoupling'' strategies to further enhance the capability of representation. However, these previous approaches focus on either the disentangling scale or semantics but ignore merging these two ideas in a union model, which extremely limits the performance of cross-modal retrieval models. To address these issues, we propose a novel Scale-Semantic Joint Decoupling Network (SSJDN) for remote sensing image-text retrieval. Specifically, we design the Bidirectional Scale Decoupling (BSD) module, which exploits Salience Feature Extraction (SFE) and Salience-Guided Suppression (SGS) units to adaptively extract potential features and suppress cumbersome features at other scales in a bidirectional pattern to yield different scale clues. Besides, we design the Label-supervised Semantic Decoupling (LSD) module by leveraging the category semantic labels as prior knowledge to supervise images and texts probing significant semantic-related information. Finally, we design a Semantic-guided Triple Loss (STL), which adaptively generates a constant to adjust the loss function to improve the probability of matching the same semantic image and text and shorten the convergence time of the retrieval model. Our proposed SSJDN outperforms state-of-the-art approaches in numerical experiments conducted on four benchmark remote sensing datasets. |
hep-th/9710193 | Kim | Dae Kwan Kim, K.G. Klimenko | Finite Density Effect in the Gross-Neveu Model in a Weakly Curved
$R^1\times S^2$ Spacetime | RevTeX, minor changes, new references are added | J.Phys.A31:5565,1998 | 10.1088/0305-4470/31/25/007 | null | hep-th hep-ph | null | The three-dimensional Gross-Neveu model in $R^{1} \times S^{2}$ spacetime is
considered at finite particles number density. We evaluate an effective
potential of the composite scalar field $\sigma(x)$, which is expressed in
terms of a scalar curvature $R$ and nonzero chemical potential $\mu$. We then
derive the critical values of $(R,\mu)$ at which the system undergoes the first
order phase transition from the phase with broken chiral invariance to the
symmetric phase.
| [
{
"created": "Sat, 25 Oct 1997 17:14:17 GMT",
"version": "v1"
},
{
"created": "Sat, 8 Nov 1997 04:12:14 GMT",
"version": "v2"
}
] | 2008-11-26 | [
[
"Kim",
"Dae Kwan",
""
],
[
"Klimenko",
"K. G.",
""
]
] | The three-dimensional Gross-Neveu model in $R^{1} \times S^{2}$ spacetime is considered at finite particles number density. We evaluate an effective potential of the composite scalar field $\sigma(x)$, which is expressed in terms of a scalar curvature $R$ and nonzero chemical potential $\mu$. We then derive the critical values of $(R,\mu)$ at which the system undergoes the first order phase transition from the phase with broken chiral invariance to the symmetric phase. |
2312.16097 | Si Wen | Si Wen and Brandon D. Gallas | Expanding to Arbitrary Study Designs: ANOVA to Estimate Limits of
Agreement for MRMC Studies | null | null | null | null | q-bio.QM | http://creativecommons.org/publicdomain/zero/1.0/ | A multi-reader multi-case (MRMC) analysis is applied to account for both
reader and case variability when evaluating the clinical performance of a
medical imaging device or reader performance under different reading
modalities. For a clinical task that measures a quantitative biomarker an
agreement analysis, such as limits of agreement (LOA), can be used. In this
work, we decompose the total variation in the data using a three-way mixed
effect ANOVA model to estimate the MRMC variance of individual differences and
the LOA between different reading modalities. There are rules for writing down
the expectation of the mean squares in terms of the variance components for
fully-crossed data, i.e. data where all the readers read all the cases in all
modalities being studied. Sometimes the annotation task is labor-intensive and
time-consuming or distributed across sites, so that a fully-crossed study is
not practical. In this work, we focus on estimating the MRMC variance in the
within- and between-readers and within- and between-modalities LOA for an
arbitrary study design. Simulation studies were conducted to validate the LOA
variance estimates. The method was also applied to a dataset to compare
pathologist performance for assessing the density of stromal tumor infiltrating
lymphocytes on different platforms.
| [
{
"created": "Tue, 26 Dec 2023 15:49:42 GMT",
"version": "v1"
}
] | 2023-12-27 | [
[
"Wen",
"Si",
""
],
[
"Gallas",
"Brandon D.",
""
]
] | A multi-reader multi-case (MRMC) analysis is applied to account for both reader and case variability when evaluating the clinical performance of a medical imaging device or reader performance under different reading modalities. For a clinical task that measures a quantitative biomarker an agreement analysis, such as limits of agreement (LOA), can be used. In this work, we decompose the total variation in the data using a three-way mixed effect ANOVA model to estimate the MRMC variance of individual differences and the LOA between different reading modalities. There are rules for writing down the expectation of the mean squares in terms of the variance components for fully-crossed data, i.e. data where all the readers read all the cases in all modalities being studied. Sometimes the annotation task is labor-intensive and time-consuming or distributed across sites, so that a fully-crossed study is not practical. In this work, we focus on estimating the MRMC variance in the within- and between-readers and within- and between-modalities LOA for an arbitrary study design. Simulation studies were conducted to validate the LOA variance estimates. The method was also applied to a dataset to compare pathologist performance for assessing the density of stromal tumor infiltrating lymphocytes on different platforms. |
hep-th/9410194 | J{\o}rgen Rasmussen, Nbi | J. Rasmussen and M. Weis | Induced Topology on the Hoop Group | 4 pages in LaTeX, NBI-HE-94-46 | null | null | null | hep-th | null | A new topology is proposed on the space of holonomy equivalence classes of
loops, induced by the topology of the space $\Sigma$ in which the loops are
embedded. The possible role for the new topology in the context of the work by
Ashtekar et al. is discussed.
| [
{
"created": "Wed, 26 Oct 1994 14:14:21 GMT",
"version": "v1"
}
] | 2007-05-23 | [
[
"Rasmussen",
"J.",
""
],
[
"Weis",
"M.",
""
]
] | A new topology is proposed on the space of holonomy equivalence classes of loops, induced by the topology of the space $\Sigma$ in which the loops are embedded. The possible role for the new topology in the context of the work by Ashtekar et al. is discussed. |
0912.2090 | Balt van Rees | Kostas Skenderis, Balt C. van Rees | Holography and wormholes in 2+1 dimensions | 37+20 pages, 23 figures; CMP version | Commun.Math.Phys.301:583-626,2011 | 10.1007/s00220-010-1163-z | ITF-2009-27 | hep-th | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We provide a holographic interpretation of a class of three-dimensional
wormhole spacetimes. These spacetimes have multiple asymptotic regions which
are separated from each other by horizons. Each such region is isometric to the
BTZ black hole and there is non-trivial spacetime topology hidden behind the
horizons. We show that application of the real-time gauge/gravity duality
results in a complete holographic description of these spacetimes with the dual
state capturing the non-trivial topology behind the horizons. We also show that
these spacetimes are in correspondence with trivalent graphs and provide an
explicit metric description with all physical parameters appearing in the
metric.
| [
{
"created": "Fri, 11 Dec 2009 20:21:52 GMT",
"version": "v1"
},
{
"created": "Wed, 1 Sep 2010 09:59:26 GMT",
"version": "v2"
}
] | 2011-02-15 | [
[
"Skenderis",
"Kostas",
""
],
[
"van Rees",
"Balt C.",
""
]
] | We provide a holographic interpretation of a class of three-dimensional wormhole spacetimes. These spacetimes have multiple asymptotic regions which are separated from each other by horizons. Each such region is isometric to the BTZ black hole and there is non-trivial spacetime topology hidden behind the horizons. We show that application of the real-time gauge/gravity duality results in a complete holographic description of these spacetimes with the dual state capturing the non-trivial topology behind the horizons. We also show that these spacetimes are in correspondence with trivalent graphs and provide an explicit metric description with all physical parameters appearing in the metric. |
hep-th/9403142 | Robert Coquereaux | R. Coquereaux | Triangular dissections, aperiodic tilings and Jones algebras | 14 pages. Revised version. 18 Postcript figures, a 500 kb uuencoded
file called images.uu available by mosaic or gopher from
gopher://cpt.univ-mrs.fr/11/preprints/94/fundamental-interactions/94-P.3020 | Adv.Appl.Math. 16 (1995) 402-424 | null | CPT - 94 /P.3020 | hep-th funct-an math.MG math.QA | null | The Brattelli diagram associated with a given bicolored Dynkin-Coxeter graph
of type $A_n$ determines planar fractal sets obtained by infinite dissections
of a given triangle. All triangles appearing in the dissection process have
angles that are multiples of $\pi/ (n+1).$ There are usually several possible
infinite dissections compatible with a given $n$ but a given one makes use of
$n/2$ triangle types if $n$ is even. Jones algebra with index $[ 4 \ \cos^2{\pi
\over n+1}]^{-1}$ (values of the discrete range) act naturally on vector spaces
associated with those fractal sets. Triangles of a given type are always
congruent at each step of the dissection process. In the particular case $n=4$,
there are isometric and the whole structure lead, after proper inflation, to
aperiodic Penrose tilings. The ``tilings'' associated with other values of the
index are discussed and shown to be encoded by equivalence classes of infinite
sequences (with appropriate constraints) using $n/2$ digits (if $n$ is even)
and generalizing the Fibonacci numbers.
| [
{
"created": "Wed, 23 Mar 1994 15:07:30 GMT",
"version": "v1"
},
{
"created": "Mon, 27 Mar 1995 11:32:27 GMT",
"version": "v2"
}
] | 2008-02-03 | [
[
"Coquereaux",
"R.",
""
]
] | The Brattelli diagram associated with a given bicolored Dynkin-Coxeter graph of type $A_n$ determines planar fractal sets obtained by infinite dissections of a given triangle. All triangles appearing in the dissection process have angles that are multiples of $\pi/ (n+1).$ There are usually several possible infinite dissections compatible with a given $n$ but a given one makes use of $n/2$ triangle types if $n$ is even. Jones algebra with index $[ 4 \ \cos^2{\pi \over n+1}]^{-1}$ (values of the discrete range) act naturally on vector spaces associated with those fractal sets. Triangles of a given type are always congruent at each step of the dissection process. In the particular case $n=4$, there are isometric and the whole structure lead, after proper inflation, to aperiodic Penrose tilings. The ``tilings'' associated with other values of the index are discussed and shown to be encoded by equivalence classes of infinite sequences (with appropriate constraints) using $n/2$ digits (if $n$ is even) and generalizing the Fibonacci numbers. |
2307.14292 | Pedro Montealegre | Pierre Fraigniaud, Fr\'ed\'eric Mazoit, Pedro Montealegre, Ivan
Rapaport, Ioan Todinca | Distributed Certification for Classes of Dense Graphs | null | null | null | null | cs.DC | http://creativecommons.org/licenses/by/4.0/ | A proof-labeling scheme (PLS) for a boolean predicate $\Pi$ on labeled graphs
is a mechanism used for certifying the legality with respect to $\Pi$ of global
network states in a distributed manner. In a PLS, a certificate is assigned to
each processing node of the network, and the nodes are in charge of checking
that the collection of certificates forms a global proof that the system is in
a correct state, by exchanging the certificates once, between neighbors only.
The main measure of complexity is the size of the certificates. Many PLSs have
been designed for certifying specific predicates, including cycle-freeness,
minimum-weight spanning tree, planarity, etc.
In 2021, a breakthrough has been obtained, as a meta-theorem stating that a
large set of properties have compact PLSs in a large class of networks. Namely,
for every $\mathrm{MSO}_2$ property $\Pi$ on labeled graphs, there exists a PLS
for $\Pi$ with $O(\log n)$-bit certificates for all graphs of bounded
tree-depth. This result has been extended to the larger class of graphs with
bounded {tree-width}, using certificates on $O(\log^2 n)$ bits.
We extend this result even further, to the larger class of graphs with
bounded clique-width, which, as opposed to the other two aforementioned
classes, includes dense graphs. We show that, for every $\mathrm{MSO}_1$
property $\Pi$ on labeled graphs, there exists a PLS for $\Pi$ with $O(\log^2
n)$ bit certificates for all graphs of bounded clique-width.
| [
{
"created": "Wed, 26 Jul 2023 16:49:39 GMT",
"version": "v1"
}
] | 2023-07-27 | [
[
"Fraigniaud",
"Pierre",
""
],
[
"Mazoit",
"Frédéric",
""
],
[
"Montealegre",
"Pedro",
""
],
[
"Rapaport",
"Ivan",
""
],
[
"Todinca",
"Ioan",
""
]
] | A proof-labeling scheme (PLS) for a boolean predicate $\Pi$ on labeled graphs is a mechanism used for certifying the legality with respect to $\Pi$ of global network states in a distributed manner. In a PLS, a certificate is assigned to each processing node of the network, and the nodes are in charge of checking that the collection of certificates forms a global proof that the system is in a correct state, by exchanging the certificates once, between neighbors only. The main measure of complexity is the size of the certificates. Many PLSs have been designed for certifying specific predicates, including cycle-freeness, minimum-weight spanning tree, planarity, etc. In 2021, a breakthrough has been obtained, as a meta-theorem stating that a large set of properties have compact PLSs in a large class of networks. Namely, for every $\mathrm{MSO}_2$ property $\Pi$ on labeled graphs, there exists a PLS for $\Pi$ with $O(\log n)$-bit certificates for all graphs of bounded tree-depth. This result has been extended to the larger class of graphs with bounded {tree-width}, using certificates on $O(\log^2 n)$ bits. We extend this result even further, to the larger class of graphs with bounded clique-width, which, as opposed to the other two aforementioned classes, includes dense graphs. We show that, for every $\mathrm{MSO}_1$ property $\Pi$ on labeled graphs, there exists a PLS for $\Pi$ with $O(\log^2 n)$ bit certificates for all graphs of bounded clique-width. |
1903.06007 | Esteban Bautista | Esteban Bautista and Patrice Abry and Paulo Gon\c{c}alves | $L^\gamma$-PageRank for Semi-Supervised Learning | Submitted to Applied Network Science (special issue on machine
learning with graphs) | null | null | null | cs.SI cs.LG eess.SP stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | PageRank for Semi-Supervised Learning has shown to leverage data structures
and limited tagged examples to yield meaningful classification. Despite
successes, classification performance can still be improved, particularly in
cases of fuzzy graphs or unbalanced labeled data. To address such limitations,
a novel approach based on powers of the Laplacian matrix $L^\gamma$ ($\gamma >
0$), referred to as $L^\gamma$-PageRank, is proposed. Its theoretical study
shows that it operates on signed graphs, where nodes belonging to one same
class are more likely to share positive edges while nodes from different
classes are more likely to be connected with negative edges. It is shown that
by selecting an optimal $\gamma$, classification performance can be
significantly enhanced. A procedure for the automated estimation of the optimal
$\gamma$, from a unique observation of data, is devised and assessed.
Experiments on several datasets demonstrate the effectiveness of both
$L^\gamma$-PageRank classification and the optimal $\gamma$ estimation.
| [
{
"created": "Mon, 11 Mar 2019 16:31:37 GMT",
"version": "v1"
}
] | 2019-03-15 | [
[
"Bautista",
"Esteban",
""
],
[
"Abry",
"Patrice",
""
],
[
"Gonçalves",
"Paulo",
""
]
] | PageRank for Semi-Supervised Learning has shown to leverage data structures and limited tagged examples to yield meaningful classification. Despite successes, classification performance can still be improved, particularly in cases of fuzzy graphs or unbalanced labeled data. To address such limitations, a novel approach based on powers of the Laplacian matrix $L^\gamma$ ($\gamma > 0$), referred to as $L^\gamma$-PageRank, is proposed. Its theoretical study shows that it operates on signed graphs, where nodes belonging to one same class are more likely to share positive edges while nodes from different classes are more likely to be connected with negative edges. It is shown that by selecting an optimal $\gamma$, classification performance can be significantly enhanced. A procedure for the automated estimation of the optimal $\gamma$, from a unique observation of data, is devised and assessed. Experiments on several datasets demonstrate the effectiveness of both $L^\gamma$-PageRank classification and the optimal $\gamma$ estimation. |
2307.02575 | Hannah Kerner | Hannah Kerner, Catherine Nakalembe, Adam Yang, Ivan Zvonkov, Ryan
McWeeny, Gabriel Tseng, Inbal Becker-Reshef | How accurate are existing land cover maps for agriculture in Sub-Saharan
Africa? | null | Scientific Data, 11(1), 486 | 10.1038/s41597-024-03306-z | null | cs.LG cs.CY | http://creativecommons.org/licenses/by-sa/4.0/ | Satellite Earth observations (EO) can provide affordable and timely
information for assessing crop conditions and food production. Such monitoring
systems are essential in Africa, where there is high food insecurity and sparse
agricultural statistics. EO-based monitoring systems require accurate cropland
maps to provide information about croplands, but there is a lack of data to
determine which of the many available land cover maps most accurately identify
cropland in African countries. This study provides a quantitative evaluation
and intercomparison of 11 publicly available land cover maps to assess their
suitability for cropland classification and EO-based agriculture monitoring in
Africa using statistically rigorous reference datasets from 8 countries. We
hope the results of this study will help users determine the most suitable map
for their needs and encourage future work to focus on resolving inconsistencies
between maps and improving accuracy in low-accuracy regions.
| [
{
"created": "Wed, 5 Jul 2023 18:17:23 GMT",
"version": "v1"
},
{
"created": "Sun, 2 Jun 2024 11:42:03 GMT",
"version": "v2"
}
] | 2024-06-04 | [
[
"Kerner",
"Hannah",
""
],
[
"Nakalembe",
"Catherine",
""
],
[
"Yang",
"Adam",
""
],
[
"Zvonkov",
"Ivan",
""
],
[
"McWeeny",
"Ryan",
""
],
[
"Tseng",
"Gabriel",
""
],
[
"Becker-Reshef",
"Inbal",
""
]
] | Satellite Earth observations (EO) can provide affordable and timely information for assessing crop conditions and food production. Such monitoring systems are essential in Africa, where there is high food insecurity and sparse agricultural statistics. EO-based monitoring systems require accurate cropland maps to provide information about croplands, but there is a lack of data to determine which of the many available land cover maps most accurately identify cropland in African countries. This study provides a quantitative evaluation and intercomparison of 11 publicly available land cover maps to assess their suitability for cropland classification and EO-based agriculture monitoring in Africa using statistically rigorous reference datasets from 8 countries. We hope the results of this study will help users determine the most suitable map for their needs and encourage future work to focus on resolving inconsistencies between maps and improving accuracy in low-accuracy regions. |
2310.05710 | Erich Schikuta | Johannes Koppenwallner, Erich Schikuta | DiCE -- A Data Encryption Proxy for the Cloud | 12 pages | null | null | null | cs.CR cs.DB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Outsourcing a relational database to the cloud offers several benefits,
including scalability, availability, and cost-effectiveness. However, there are
concerns about the confidentiality and security of the outsourced data. A
general approach here would be to encrypt the data with a standardized
encryption algorithm and then store the data only encrypted in the cloud. The
problem with this approach, however, is that with encryption, important
properties of the data such as sorting, format or comparability, which are
essential for the functioning of database queries, are lost. One solution to
this problem is the use of (e.g. order-preserving) encryption algorithms, which
also preserve these properties in the encrypted data, thus enabling queries to
encrypted data. These algorithms range from simple algorithms like Caesar
encryption to secure algorithms like mOPE. In order to be able to use these
algorithms as easy as possible, ``DiCE'' a JDBC driver was developed, that
parses SQL queries as a proxy and transparently encrypts and decrypts these
queries. This allows to execute many queries on an encrypted database in the
cloud with (nearly) the performance as on unencrypted databases. The DiCE
driver can be used with any other JDBC driver and therefore supports a variety
of databases. The driver can be configured to support different encryption
algorithms. To keep track of the operations, the ``Dice Information Client''
has been developed to track the encryption and decryption of the driver.
Although the result of the performance analysis shows a certain overhead due to
the parsing and encryption of the SQL queries in the Dice driver, this overhead
is significantly reduced by other influencing factors such as the network and
parallel queries.
| [
{
"created": "Mon, 9 Oct 2023 13:33:59 GMT",
"version": "v1"
}
] | 2023-10-10 | [
[
"Koppenwallner",
"Johannes",
""
],
[
"Schikuta",
"Erich",
""
]
] | Outsourcing a relational database to the cloud offers several benefits, including scalability, availability, and cost-effectiveness. However, there are concerns about the confidentiality and security of the outsourced data. A general approach here would be to encrypt the data with a standardized encryption algorithm and then store the data only encrypted in the cloud. The problem with this approach, however, is that with encryption, important properties of the data such as sorting, format or comparability, which are essential for the functioning of database queries, are lost. One solution to this problem is the use of (e.g. order-preserving) encryption algorithms, which also preserve these properties in the encrypted data, thus enabling queries to encrypted data. These algorithms range from simple algorithms like Caesar encryption to secure algorithms like mOPE. In order to be able to use these algorithms as easy as possible, ``DiCE'' a JDBC driver was developed, that parses SQL queries as a proxy and transparently encrypts and decrypts these queries. This allows to execute many queries on an encrypted database in the cloud with (nearly) the performance as on unencrypted databases. The DiCE driver can be used with any other JDBC driver and therefore supports a variety of databases. The driver can be configured to support different encryption algorithms. To keep track of the operations, the ``Dice Information Client'' has been developed to track the encryption and decryption of the driver. Although the result of the performance analysis shows a certain overhead due to the parsing and encryption of the SQL queries in the Dice driver, this overhead is significantly reduced by other influencing factors such as the network and parallel queries. |
1209.6117 | A.G. Tsuchiya | A.G. Tsuchiya | On the pole structures of the disconnected part of hyper elliptic g loop
M point super string amplitudes | Genus one result is corrected and the proof modified | null | null | null | hep-th | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Structures of the disconnected part of higher genus superstring amplitudes
restricted to the hyper elliptic cases are investigated in the NSR formalism,
based on the DHoker Phong and recent results. A set of equations, which we can
regard as a basic tool to sum over the spin structures of any of g loop, M
point amplitudes systematically, is shown by using a classical result of
Abelian functions. We discuss structures of g loop, M point massless external
boson superstring amplitudes by assuming that the spin structure dependence of
any of the disconnected amplitudes is only on one kind of constants, the genus
g Weierstrass Pe function valued at the summation of g number of half periods
chosen out of 2g+1 half periods. This is a natural generalization of the case
of genus 1. This assumption will be validated by a conjectured theorem which
states that the spin structure dependent part of any string amplitude will be
naturally decomposed into two parts. One is composed of manifestly modular
invariant functions of positions of inserting operators, and the other is the
polynomial of Pe function constants related to the moduli of Riemann surfaces
only. It is shown that this is actually the case for any M for g=1, and M=1,2,3
for any g. Due to a technical problem, our consideration is at present
restricted to the case that g(g+1)divided by 2 is odd. Example calculations are
shown for the genus 2 by the method described here. In particular, our method
correctly reproduces biholomorphic 1 form of DHoker Phong result as for the
four point amplitudes of the disconnected parts.
| [
{
"created": "Thu, 27 Sep 2012 03:34:44 GMT",
"version": "v1"
},
{
"created": "Sun, 7 Oct 2012 11:23:41 GMT",
"version": "v2"
},
{
"created": "Fri, 3 Apr 2015 00:21:37 GMT",
"version": "v3"
}
] | 2015-04-06 | [
[
"Tsuchiya",
"A. G.",
""
]
] | Structures of the disconnected part of higher genus superstring amplitudes restricted to the hyper elliptic cases are investigated in the NSR formalism, based on the DHoker Phong and recent results. A set of equations, which we can regard as a basic tool to sum over the spin structures of any of g loop, M point amplitudes systematically, is shown by using a classical result of Abelian functions. We discuss structures of g loop, M point massless external boson superstring amplitudes by assuming that the spin structure dependence of any of the disconnected amplitudes is only on one kind of constants, the genus g Weierstrass Pe function valued at the summation of g number of half periods chosen out of 2g+1 half periods. This is a natural generalization of the case of genus 1. This assumption will be validated by a conjectured theorem which states that the spin structure dependent part of any string amplitude will be naturally decomposed into two parts. One is composed of manifestly modular invariant functions of positions of inserting operators, and the other is the polynomial of Pe function constants related to the moduli of Riemann surfaces only. It is shown that this is actually the case for any M for g=1, and M=1,2,3 for any g. Due to a technical problem, our consideration is at present restricted to the case that g(g+1)divided by 2 is odd. Example calculations are shown for the genus 2 by the method described here. In particular, our method correctly reproduces biholomorphic 1 form of DHoker Phong result as for the four point amplitudes of the disconnected parts. |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.