id
stringlengths
9
10
submitter
stringlengths
1
64
authors
stringlengths
4
20.7k
title
stringlengths
4
246
comments
stringlengths
1
523
journal-ref
stringlengths
4
404
doi
stringlengths
11
153
report-no
stringlengths
2
254
categories
stringlengths
5
98
license
stringclasses
9 values
orig_abstract
stringlengths
14
3.35k
versions
listlengths
1
60
update_date
stringlengths
10
10
authors_parsed
listlengths
1
1.35k
abstract
stringlengths
11
3.34k
2202.11940
Soumyabrata Pal
Arya Mazumdar, Soumyabrata Pal
Support Recovery in Mixture Models with Sparse Parameters
55 pages, Shorter version titled "On Learning Mixture Models with Sparse Parameters " accepted at AISTATS 2022
null
null
null
cs.LG cs.IT math.IT stat.ML
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Mixture models are widely used to fit complex and multimodal datasets. In this paper we study mixtures with high dimensional sparse latent parameter vectors and consider the problem of support recovery of those vectors. While parameter learning in mixture models is well-studied, the sparsity constraint remains relatively unexplored. Sparsity of parameter vectors is a natural constraint in variety of settings, and support recovery is a major step towards parameter estimation. We provide efficient algorithms for support recovery that have a logarithmic sample complexity dependence on the dimensionality of the latent space. Our algorithms are quite general, namely they are applicable to 1) mixtures of many different canonical distributions including Uniform, Poisson, Laplace, Gaussians, etc. 2) Mixtures of linear regressions and linear classifiers with Gaussian covariates under different assumptions on the unknown parameters. In most of these settings, our results are the first guarantees on the problem while in the rest, our results provide improvements on existing works.
[ { "created": "Thu, 24 Feb 2022 07:44:23 GMT", "version": "v1" }, { "created": "Sat, 10 Sep 2022 10:24:47 GMT", "version": "v2" } ]
2022-09-13
[ [ "Mazumdar", "Arya", "" ], [ "Pal", "Soumyabrata", "" ] ]
Mixture models are widely used to fit complex and multimodal datasets. In this paper we study mixtures with high dimensional sparse latent parameter vectors and consider the problem of support recovery of those vectors. While parameter learning in mixture models is well-studied, the sparsity constraint remains relatively unexplored. Sparsity of parameter vectors is a natural constraint in variety of settings, and support recovery is a major step towards parameter estimation. We provide efficient algorithms for support recovery that have a logarithmic sample complexity dependence on the dimensionality of the latent space. Our algorithms are quite general, namely they are applicable to 1) mixtures of many different canonical distributions including Uniform, Poisson, Laplace, Gaussians, etc. 2) Mixtures of linear regressions and linear classifiers with Gaussian covariates under different assumptions on the unknown parameters. In most of these settings, our results are the first guarantees on the problem while in the rest, our results provide improvements on existing works.
1612.09394
Kwonsoo Chae
Kwonsoo Chae and Hakjoo Oh and Kihong Heo and Hongseok Yang
Automatically generating features for learning program analysis heuristics
null
null
null
null
cs.PL
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
We present a technique for automatically generating features for data-driven program analyses. Recently data-driven approaches for building a program analysis have been proposed, which mine existing codebases and automatically learn heuristics for finding a cost-effective abstraction for a given analysis task. Such approaches reduce the burden of the analysis designers, but they do not remove it completely; they still leave the highly nontrivial task of designing so called features to the hands of the designers. Our technique automates this feature design process. The idea is to use programs as features after reducing and abstracting them. Our technique goes through selected program-query pairs in codebases, and it reduces and abstracts the program in each pair to a few lines of code, while ensuring that the analysis behaves similarly for the original and the new programs with respect to the query. Each reduced program serves as a boolean feature for program-query pairs. This feature evaluates to true for a given program-query pair when (as a program) it is included in the program part of the pair. We have implemented our approach for three real-world program analyses. Our experimental evaluation shows that these analyses with automatically-generated features perform comparably to those with manually crafted features.
[ { "created": "Fri, 30 Dec 2016 05:55:56 GMT", "version": "v1" } ]
2017-01-02
[ [ "Chae", "Kwonsoo", "" ], [ "Oh", "Hakjoo", "" ], [ "Heo", "Kihong", "" ], [ "Yang", "Hongseok", "" ] ]
We present a technique for automatically generating features for data-driven program analyses. Recently data-driven approaches for building a program analysis have been proposed, which mine existing codebases and automatically learn heuristics for finding a cost-effective abstraction for a given analysis task. Such approaches reduce the burden of the analysis designers, but they do not remove it completely; they still leave the highly nontrivial task of designing so called features to the hands of the designers. Our technique automates this feature design process. The idea is to use programs as features after reducing and abstracting them. Our technique goes through selected program-query pairs in codebases, and it reduces and abstracts the program in each pair to a few lines of code, while ensuring that the analysis behaves similarly for the original and the new programs with respect to the query. Each reduced program serves as a boolean feature for program-query pairs. This feature evaluates to true for a given program-query pair when (as a program) it is included in the program part of the pair. We have implemented our approach for three real-world program analyses. Our experimental evaluation shows that these analyses with automatically-generated features perform comparably to those with manually crafted features.
2009.10047
Congcong Wang
Congcong Wang and David Lillis
UCD-CS at W-NUT 2020 Shared Task-3: A Text to Text Approach for COVID-19 Event Extraction on Social Media
8 pages, 2 figures
null
null
null
cs.CL
http://creativecommons.org/licenses/by/4.0/
In this paper, we describe our approach in the shared task: COVID-19 event extraction from Twitter. The objective of this task is to extract answers from COVID-related tweets to a set of predefined slot-filling questions. Our approach treats the event extraction task as a question answering task by leveraging the transformer-based T5 text-to-text model. According to the official evaluation scores returned, namely F1, our submitted run achieves competitive performance compared to other participating runs (Top 3). However, we argue that this evaluation may underestimate the actual performance of runs based on text-generation. Although some such runs may answer the slot questions well, they may not be an exact string match for the gold standard answers. To measure the extent of this underestimation, we adopt a simple exact-answer transformation method aiming at converting the well-answered predictions to exactly-matched predictions. The results show that after this transformation our run overall reaches the same level of performance as the best participating run and state-of-the-art F1 scores in three of five COVID-related events. Our code is publicly available to aid reproducibility
[ { "created": "Mon, 21 Sep 2020 17:39:00 GMT", "version": "v1" }, { "created": "Mon, 12 Oct 2020 16:18:53 GMT", "version": "v2" } ]
2021-02-19
[ [ "Wang", "Congcong", "" ], [ "Lillis", "David", "" ] ]
In this paper, we describe our approach in the shared task: COVID-19 event extraction from Twitter. The objective of this task is to extract answers from COVID-related tweets to a set of predefined slot-filling questions. Our approach treats the event extraction task as a question answering task by leveraging the transformer-based T5 text-to-text model. According to the official evaluation scores returned, namely F1, our submitted run achieves competitive performance compared to other participating runs (Top 3). However, we argue that this evaluation may underestimate the actual performance of runs based on text-generation. Although some such runs may answer the slot questions well, they may not be an exact string match for the gold standard answers. To measure the extent of this underestimation, we adopt a simple exact-answer transformation method aiming at converting the well-answered predictions to exactly-matched predictions. The results show that after this transformation our run overall reaches the same level of performance as the best participating run and state-of-the-art F1 scores in three of five COVID-related events. Our code is publicly available to aid reproducibility
1702.08300
Merim Dzaferagic
Merim Dzaferagic, Nicholas Kaminski, Irene Macaluso, Nicola Marchetti
How Functional Complexity affects the Scalability-Energy Efficiency Trade-Off of HCC WSN Clustering
arXiv admin note: substantial text overlap with arXiv:1610.05970
null
null
null
cs.NI
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Even though clustering algorithms in Wireless Sensor Networks (WSN) are a well investigate subject, the increasing interest in the Internet of Things (IoT) and 5G technologies has precipitated the need of new ways to comprehend and overcome a new set of challenges. While studies mainly propose new algorithms and compare these algorithms based on a set of properties (e.g. energy efficiency, scalability), none of them focuses on the underlying mechanisms and organizational patterns that lead to these properties. We address this lack of understanding by applying a complex systems science approach to investigate the properties of WSNs arising from the communication patterns of the network nodes. We represent different implementations of clustering in WSNs with a functional topology graph. Moreover, we employ a complexity metric - functional complexity (CF) - to explain how local interactions give rise to the global behavior of the network. Our analysis shows that higher values of CF indicate higher scalability and lower energy efficiency.
[ { "created": "Mon, 27 Feb 2017 14:28:26 GMT", "version": "v1" } ]
2017-02-28
[ [ "Dzaferagic", "Merim", "" ], [ "Kaminski", "Nicholas", "" ], [ "Macaluso", "Irene", "" ], [ "Marchetti", "Nicola", "" ] ]
Even though clustering algorithms in Wireless Sensor Networks (WSN) are a well investigate subject, the increasing interest in the Internet of Things (IoT) and 5G technologies has precipitated the need of new ways to comprehend and overcome a new set of challenges. While studies mainly propose new algorithms and compare these algorithms based on a set of properties (e.g. energy efficiency, scalability), none of them focuses on the underlying mechanisms and organizational patterns that lead to these properties. We address this lack of understanding by applying a complex systems science approach to investigate the properties of WSNs arising from the communication patterns of the network nodes. We represent different implementations of clustering in WSNs with a functional topology graph. Moreover, we employ a complexity metric - functional complexity (CF) - to explain how local interactions give rise to the global behavior of the network. Our analysis shows that higher values of CF indicate higher scalability and lower energy efficiency.
1802.03064
Dirk Pfl\"uger
Markus K\"oppel and Fabian Franzelin and Ilja Kr\"oker and Sergey Oladyshkin and Gabriele Santin and Dominik Wittwar and Andrea Barth and Bernard Haasdonk and Wolfgang Nowak and Dirk Pfl\"uger and Christian Rohde
Comparison of data-driven uncertainty quantification methods for a carbon dioxide storage benchmark scenario
null
null
10.1007/s10596-018-9785-x
null
cs.CE cs.NA math.NA
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
A variety of methods is available to quantify uncertainties arising with\-in the modeling of flow and transport in carbon dioxide storage, but there is a lack of thorough comparisons. Usually, raw data from such storage sites can hardly be described by theoretical statistical distributions since only very limited data is available. Hence, exact information on distribution shapes for all uncertain parameters is very rare in realistic applications. We discuss and compare four different methods tested for data-driven uncertainty quantification based on a benchmark scenario of carbon dioxide storage. In the benchmark, for which we provide data and code, carbon dioxide is injected into a saline aquifer modeled by the nonlinear capillarity-free fractional flow formulation for two incompressible fluid phases, namely carbon dioxide and brine. To cover different aspects of uncertainty quantification, we incorporate various sources of uncertainty such as uncertainty of boundary conditions, of conceptual model definitions and of material properties. We consider recent versions of the following non-intrusive and intrusive uncertainty quantification methods: arbitary polynomial chaos, spatially adaptive sparse grids, kernel-based greedy interpolation and hybrid stochastic Galerkin. The performance of each approach is demonstrated assessing expectation value and standard deviation of the carbon dioxide saturation against a reference statistic based on Monte Carlo sampling. We compare the convergence of all methods reporting on accuracy with respect to the number of model runs and resolution. Finally we offer suggestions about the methods' advantages and disadvantages that can guide the modeler for uncertainty quantification in carbon dioxide storage and beyond.
[ { "created": "Thu, 8 Feb 2018 22:27:38 GMT", "version": "v1" } ]
2018-11-13
[ [ "Köppel", "Markus", "" ], [ "Franzelin", "Fabian", "" ], [ "Kröker", "Ilja", "" ], [ "Oladyshkin", "Sergey", "" ], [ "Santin", "Gabriele", "" ], [ "Wittwar", "Dominik", "" ], [ "Barth", "Andrea", "" ], [ "Haasdonk", "Bernard", "" ], [ "Nowak", "Wolfgang", "" ], [ "Pflüger", "Dirk", "" ], [ "Rohde", "Christian", "" ] ]
A variety of methods is available to quantify uncertainties arising with\-in the modeling of flow and transport in carbon dioxide storage, but there is a lack of thorough comparisons. Usually, raw data from such storage sites can hardly be described by theoretical statistical distributions since only very limited data is available. Hence, exact information on distribution shapes for all uncertain parameters is very rare in realistic applications. We discuss and compare four different methods tested for data-driven uncertainty quantification based on a benchmark scenario of carbon dioxide storage. In the benchmark, for which we provide data and code, carbon dioxide is injected into a saline aquifer modeled by the nonlinear capillarity-free fractional flow formulation for two incompressible fluid phases, namely carbon dioxide and brine. To cover different aspects of uncertainty quantification, we incorporate various sources of uncertainty such as uncertainty of boundary conditions, of conceptual model definitions and of material properties. We consider recent versions of the following non-intrusive and intrusive uncertainty quantification methods: arbitary polynomial chaos, spatially adaptive sparse grids, kernel-based greedy interpolation and hybrid stochastic Galerkin. The performance of each approach is demonstrated assessing expectation value and standard deviation of the carbon dioxide saturation against a reference statistic based on Monte Carlo sampling. We compare the convergence of all methods reporting on accuracy with respect to the number of model runs and resolution. Finally we offer suggestions about the methods' advantages and disadvantages that can guide the modeler for uncertainty quantification in carbon dioxide storage and beyond.
1904.03122
Stefan Larson
Stefan Larson, Anish Mahendran, Andrew Lee, Jonathan K. Kummerfeld, Parker Hill, Michael A. Laurenzano, Johann Hauswald, Lingjia Tang, Jason Mars
Outlier Detection for Improved Data Quality and Diversity in Dialog Systems
Accepted as long paper to NAACL 2019
null
null
null
cs.CL
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
In a corpus of data, outliers are either errors: mistakes in the data that are counterproductive, or are unique: informative samples that improve model robustness. Identifying outliers can lead to better datasets by (1) removing noise in datasets and (2) guiding collection of additional data to fill gaps. However, the problem of detecting both outlier types has received relatively little attention in NLP, particularly for dialog systems. We introduce a simple and effective technique for detecting both erroneous and unique samples in a corpus of short texts using neural sentence embeddings combined with distance-based outlier detection. We also present a novel data collection pipeline built atop our detection technique to automatically and iteratively mine unique data samples while discarding erroneous samples. Experiments show that our outlier detection technique is effective at finding errors while our data collection pipeline yields highly diverse corpora that in turn produce more robust intent classification and slot-filling models.
[ { "created": "Fri, 5 Apr 2019 15:31:28 GMT", "version": "v1" } ]
2019-04-08
[ [ "Larson", "Stefan", "" ], [ "Mahendran", "Anish", "" ], [ "Lee", "Andrew", "" ], [ "Kummerfeld", "Jonathan K.", "" ], [ "Hill", "Parker", "" ], [ "Laurenzano", "Michael A.", "" ], [ "Hauswald", "Johann", "" ], [ "Tang", "Lingjia", "" ], [ "Mars", "Jason", "" ] ]
In a corpus of data, outliers are either errors: mistakes in the data that are counterproductive, or are unique: informative samples that improve model robustness. Identifying outliers can lead to better datasets by (1) removing noise in datasets and (2) guiding collection of additional data to fill gaps. However, the problem of detecting both outlier types has received relatively little attention in NLP, particularly for dialog systems. We introduce a simple and effective technique for detecting both erroneous and unique samples in a corpus of short texts using neural sentence embeddings combined with distance-based outlier detection. We also present a novel data collection pipeline built atop our detection technique to automatically and iteratively mine unique data samples while discarding erroneous samples. Experiments show that our outlier detection technique is effective at finding errors while our data collection pipeline yields highly diverse corpora that in turn produce more robust intent classification and slot-filling models.
0912.3429
Pietro Sala Mr.
A. Montanari, G. Puppis, P. Sala, G. Sciavicco
Decidability of the interval temporal logic ABBar over the natural numbers
null
null
null
null
cs.LO
http://creativecommons.org/licenses/by/3.0/
In this paper, we focus our attention on the interval temporal logic of the Allen's relations "meets", "begins", and "begun by" (ABBar for short), interpreted over natural numbers. We first introduce the logic and we show that it is expressive enough to model distinctive interval properties,such as accomplishment conditions, to capture basic modalities of point-based temporal logic, such as the until operator, and to encode relevant metric constraints. Then, we prove that the satisfiability problem for ABBar over natural numbers is decidable by providing a small model theorem based on an original contraction method. Finally, we prove the EXPSPACE-completeness of the problem
[ { "created": "Thu, 17 Dec 2009 15:22:45 GMT", "version": "v1" }, { "created": "Wed, 3 Feb 2010 13:40:37 GMT", "version": "v2" } ]
2010-02-03
[ [ "Montanari", "A.", "" ], [ "Puppis", "G.", "" ], [ "Sala", "P.", "" ], [ "Sciavicco", "G.", "" ] ]
In this paper, we focus our attention on the interval temporal logic of the Allen's relations "meets", "begins", and "begun by" (ABBar for short), interpreted over natural numbers. We first introduce the logic and we show that it is expressive enough to model distinctive interval properties,such as accomplishment conditions, to capture basic modalities of point-based temporal logic, such as the until operator, and to encode relevant metric constraints. Then, we prove that the satisfiability problem for ABBar over natural numbers is decidable by providing a small model theorem based on an original contraction method. Finally, we prove the EXPSPACE-completeness of the problem
2404.00801
Ye Liu
Ye Liu, Jixuan He, Wanhua Li, Junsik Kim, Donglai Wei, Hanspeter Pfister, Chang Wen Chen
$R^2$-Tuning: Efficient Image-to-Video Transfer Learning for Video Temporal Grounding
ECCV 2024 Camera Ready
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Video temporal grounding (VTG) is a fine-grained video understanding problem that aims to ground relevant clips in untrimmed videos given natural language queries. Most existing VTG models are built upon frame-wise final-layer CLIP features, aided by additional temporal backbones (e.g., SlowFast) with sophisticated temporal reasoning mechanisms. In this work, we claim that CLIP itself already shows great potential for fine-grained spatial-temporal modeling, as each layer offers distinct yet useful information under different granularity levels. Motivated by this, we propose Reversed Recurrent Tuning ($R^2$-Tuning), a parameter- and memory-efficient transfer learning framework for video temporal grounding. Our method learns a lightweight $R^2$ Block containing only 1.5% of the total parameters to perform progressive spatial-temporal modeling. Starting from the last layer of CLIP, $R^2$ Block recurrently aggregates spatial features from earlier layers, then refines temporal correlation conditioning on the given query, resulting in a coarse-to-fine scheme. $R^2$-Tuning achieves state-of-the-art performance across three VTG tasks (i.e., moment retrieval, highlight detection, and video summarization) on six public benchmarks (i.e., QVHighlights, Charades-STA, Ego4D-NLQ, TACoS, YouTube Highlights, and TVSum) even without the additional backbone, demonstrating the significance and effectiveness of the proposed scheme. Our code is available at https://github.com/yeliudev/R2-Tuning.
[ { "created": "Sun, 31 Mar 2024 21:17:48 GMT", "version": "v1" }, { "created": "Sun, 21 Jul 2024 16:17:07 GMT", "version": "v2" } ]
2024-07-23
[ [ "Liu", "Ye", "" ], [ "He", "Jixuan", "" ], [ "Li", "Wanhua", "" ], [ "Kim", "Junsik", "" ], [ "Wei", "Donglai", "" ], [ "Pfister", "Hanspeter", "" ], [ "Chen", "Chang Wen", "" ] ]
Video temporal grounding (VTG) is a fine-grained video understanding problem that aims to ground relevant clips in untrimmed videos given natural language queries. Most existing VTG models are built upon frame-wise final-layer CLIP features, aided by additional temporal backbones (e.g., SlowFast) with sophisticated temporal reasoning mechanisms. In this work, we claim that CLIP itself already shows great potential for fine-grained spatial-temporal modeling, as each layer offers distinct yet useful information under different granularity levels. Motivated by this, we propose Reversed Recurrent Tuning ($R^2$-Tuning), a parameter- and memory-efficient transfer learning framework for video temporal grounding. Our method learns a lightweight $R^2$ Block containing only 1.5% of the total parameters to perform progressive spatial-temporal modeling. Starting from the last layer of CLIP, $R^2$ Block recurrently aggregates spatial features from earlier layers, then refines temporal correlation conditioning on the given query, resulting in a coarse-to-fine scheme. $R^2$-Tuning achieves state-of-the-art performance across three VTG tasks (i.e., moment retrieval, highlight detection, and video summarization) on six public benchmarks (i.e., QVHighlights, Charades-STA, Ego4D-NLQ, TACoS, YouTube Highlights, and TVSum) even without the additional backbone, demonstrating the significance and effectiveness of the proposed scheme. Our code is available at https://github.com/yeliudev/R2-Tuning.
1509.07513
Marco Antonio Valenzuela Esc\'arcega
Marco A. Valenzuela-Esc\'arcega, Gus Hahn-Powell, Mihai Surdeanu
Description of the Odin Event Extraction Framework and Rule Language
null
null
null
null
cs.CL
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
This document describes the Odin framework, which is a domain-independent platform for developing rule-based event extraction models. Odin aims to be powerful (the rule language allows the modeling of complex syntactic structures) and robust (to recover from syntactic parsing errors, syntactic patterns can be freely mixed with surface, token-based patterns), while remaining simple (some domain grammars can be up and running in minutes), and fast (Odin processes over 100 sentences/second in a real-world domain with over 200 rules). Here we include a thorough definition of the Odin rule language, together with a description of the Odin API in the Scala language, which allows one to apply these rules to arbitrary texts.
[ { "created": "Thu, 24 Sep 2015 20:10:27 GMT", "version": "v1" } ]
2015-09-28
[ [ "Valenzuela-Escárcega", "Marco A.", "" ], [ "Hahn-Powell", "Gus", "" ], [ "Surdeanu", "Mihai", "" ] ]
This document describes the Odin framework, which is a domain-independent platform for developing rule-based event extraction models. Odin aims to be powerful (the rule language allows the modeling of complex syntactic structures) and robust (to recover from syntactic parsing errors, syntactic patterns can be freely mixed with surface, token-based patterns), while remaining simple (some domain grammars can be up and running in minutes), and fast (Odin processes over 100 sentences/second in a real-world domain with over 200 rules). Here we include a thorough definition of the Odin rule language, together with a description of the Odin API in the Scala language, which allows one to apply these rules to arbitrary texts.
2110.14124
Wang Chen
Wang Chen, Jian Chen, Weitian Wu, Xinmin Yang, Hui Li
A novel multiobjective evolutionary algorithm based on decomposition and multi-reference points strategy
null
null
null
null
cs.NE math.OC
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Many real-world optimization problems such as engineering design can be eventually modeled as the corresponding multiobjective optimization problems (MOPs) which must be solved to obtain approximate Pareto optimal fronts. Multiobjective evolutionary algorithm based on decomposition (MOEA/D) has been regarded as a significantly promising approach for solving MOPs. Recent studies have shown that MOEA/D with uniform weight vectors is well-suited to MOPs with regular Pareto optimal fronts, but its performance in terms of diversity usually deteriorates when solving MOPs with irregular Pareto optimal fronts. In this way, the solution set obtained by the algorithm can not provide more reasonable choices for decision makers. In order to efficiently overcome this drawback, we propose an improved MOEA/D algorithm by virtue of the well-known Pascoletti-Serafini scalarization method and a new strategy of multi-reference points. Specifically, this strategy consists of the setting and adaptation of reference points generated by the techniques of equidistant partition and projection. For performance assessment, the proposed algorithm is compared with existing four state-of-the-art multiobjective evolutionary algorithms on benchmark test problems with various types of Pareto optimal fronts. According to the experimental results, the proposed algorithm exhibits better diversity performance than that of the other compared algorithms. Finally, our algorithm is applied to two real-world MOPs in engineering optimization successfully.
[ { "created": "Wed, 27 Oct 2021 02:07:08 GMT", "version": "v1" }, { "created": "Mon, 1 Nov 2021 13:31:17 GMT", "version": "v2" }, { "created": "Tue, 2 Nov 2021 07:01:08 GMT", "version": "v3" }, { "created": "Wed, 3 Nov 2021 11:03:40 GMT", "version": "v4" }, { "created": "Mon, 8 Nov 2021 16:07:28 GMT", "version": "v5" }, { "created": "Thu, 11 Nov 2021 08:21:35 GMT", "version": "v6" } ]
2021-11-12
[ [ "Chen", "Wang", "" ], [ "Chen", "Jian", "" ], [ "Wu", "Weitian", "" ], [ "Yang", "Xinmin", "" ], [ "Li", "Hui", "" ] ]
Many real-world optimization problems such as engineering design can be eventually modeled as the corresponding multiobjective optimization problems (MOPs) which must be solved to obtain approximate Pareto optimal fronts. Multiobjective evolutionary algorithm based on decomposition (MOEA/D) has been regarded as a significantly promising approach for solving MOPs. Recent studies have shown that MOEA/D with uniform weight vectors is well-suited to MOPs with regular Pareto optimal fronts, but its performance in terms of diversity usually deteriorates when solving MOPs with irregular Pareto optimal fronts. In this way, the solution set obtained by the algorithm can not provide more reasonable choices for decision makers. In order to efficiently overcome this drawback, we propose an improved MOEA/D algorithm by virtue of the well-known Pascoletti-Serafini scalarization method and a new strategy of multi-reference points. Specifically, this strategy consists of the setting and adaptation of reference points generated by the techniques of equidistant partition and projection. For performance assessment, the proposed algorithm is compared with existing four state-of-the-art multiobjective evolutionary algorithms on benchmark test problems with various types of Pareto optimal fronts. According to the experimental results, the proposed algorithm exhibits better diversity performance than that of the other compared algorithms. Finally, our algorithm is applied to two real-world MOPs in engineering optimization successfully.
2202.11460
Pavel Hrab\'ak
Hana Najmanov\'a and Veronika Pe\v{s}kov\'a and Luk\'a\v{s} Kukl\'ik and Marek Buk\'a\v{c}ek and Pavel Hrab\'ak and Daniel Va\v{s}ata
Evacuation trials from a double-deck electric train unit: Experimental data and sensitivity analysis
null
Safety Science, Volume 146, 2022, 105523, ISSN 0925-7535
10.1016/j.ssci.2021.105523
null
cs.MA physics.soc-ph
http://creativecommons.org/licenses/by-nc-nd/4.0/
Passenger trains represent a challenging environment in emergencies, with specific evacuation conditions resulting from the typical layout and interior design inherent to public transportation vehicles. This paper describes a dataset obtained in a full-scale controlled experiment emulating the emergency evacuation of a double-deck electric unit railcar carried out in Prague in 2018. 15 evacuation trials involving 91 participants were conducted under various evacuation scenarios considering different compositions of passenger crowd, exit widths, and exit types (e.g. egress to a high platform, to an open rail line using stairs, and a 750 mm jump without any supporting equipment). The study's main goals were to collect experimental data on the movement conditions in the railcar and to study the impact of various boundary conditions on evacuation process and total evacuation time. Movement characteristics (exit flows, speeds) and human behaviour (pre-movement activities, exiting behaviours) were also analysed. The data obtained was used to validate and adjust a Pathfinder model to capture important aspects of evacuation from the railcar. Furthermore, a series of simulations using this model was performed to provide sensitivity analysis of the influence of crowd composition, exit width, and exit type on total evacuation time. As a key finding, we can conclude that for the case of a standard exit path (platform or stairs) the width of the main exit had the greatest impact on total evacuation time, however, crowd composition played the prevailing role in evacuation scenarios involving a jump.
[ { "created": "Wed, 23 Feb 2022 12:25:06 GMT", "version": "v1" } ]
2022-02-24
[ [ "Najmanová", "Hana", "" ], [ "Pešková", "Veronika", "" ], [ "Kuklík", "Lukáš", "" ], [ "Bukáček", "Marek", "" ], [ "Hrabák", "Pavel", "" ], [ "Vašata", "Daniel", "" ] ]
Passenger trains represent a challenging environment in emergencies, with specific evacuation conditions resulting from the typical layout and interior design inherent to public transportation vehicles. This paper describes a dataset obtained in a full-scale controlled experiment emulating the emergency evacuation of a double-deck electric unit railcar carried out in Prague in 2018. 15 evacuation trials involving 91 participants were conducted under various evacuation scenarios considering different compositions of passenger crowd, exit widths, and exit types (e.g. egress to a high platform, to an open rail line using stairs, and a 750 mm jump without any supporting equipment). The study's main goals were to collect experimental data on the movement conditions in the railcar and to study the impact of various boundary conditions on evacuation process and total evacuation time. Movement characteristics (exit flows, speeds) and human behaviour (pre-movement activities, exiting behaviours) were also analysed. The data obtained was used to validate and adjust a Pathfinder model to capture important aspects of evacuation from the railcar. Furthermore, a series of simulations using this model was performed to provide sensitivity analysis of the influence of crowd composition, exit width, and exit type on total evacuation time. As a key finding, we can conclude that for the case of a standard exit path (platform or stairs) the width of the main exit had the greatest impact on total evacuation time, however, crowd composition played the prevailing role in evacuation scenarios involving a jump.
2401.11697
Sundar Narayanan
Sundaraparipurnan Narayanan, Mark Potkewitz
A risk-based approach to assessing liability risk for AI-driven harms considering EU liability directive
null
null
null
null
cs.CY
http://creativecommons.org/licenses/by/4.0/
Artificial intelligence can cause inconvenience, harm, or other unintended consequences in various ways, including those that arise from defects or malfunctions in the AI system itself or those caused by its use or misuse. Responsibility for AI harms or unintended consequences must be addressed to hold accountable the people who caused such harms and ensure that victims receive compensation for any damages or losses they may have sustained. Historical instances of harm caused by AI have led to European Union establishing an AI Liability Directive. The directive aims to lay down a uniform set of rules for access to information, delineate the duty and level of care required for AI development and use, and clarify the burden of proof for damages or harms caused by AI systems, establishing broader protection for victims. The future ability of provider to contest a product liability claim will depend on good practices adopted in designing, developing, and maintaining AI systems in the market. This paper provides a risk-based approach to examining liability for AI-driven injuries. It also provides an overview of existing liability approaches, insights into limitations and complexities in these approaches, and a detailed self-assessment questionnaire to assess the risk associated with liability for a specific AI system from a provider's perspective.
[ { "created": "Mon, 18 Dec 2023 15:52:43 GMT", "version": "v1" } ]
2024-01-23
[ [ "Narayanan", "Sundaraparipurnan", "" ], [ "Potkewitz", "Mark", "" ] ]
Artificial intelligence can cause inconvenience, harm, or other unintended consequences in various ways, including those that arise from defects or malfunctions in the AI system itself or those caused by its use or misuse. Responsibility for AI harms or unintended consequences must be addressed to hold accountable the people who caused such harms and ensure that victims receive compensation for any damages or losses they may have sustained. Historical instances of harm caused by AI have led to European Union establishing an AI Liability Directive. The directive aims to lay down a uniform set of rules for access to information, delineate the duty and level of care required for AI development and use, and clarify the burden of proof for damages or harms caused by AI systems, establishing broader protection for victims. The future ability of provider to contest a product liability claim will depend on good practices adopted in designing, developing, and maintaining AI systems in the market. This paper provides a risk-based approach to examining liability for AI-driven injuries. It also provides an overview of existing liability approaches, insights into limitations and complexities in these approaches, and a detailed self-assessment questionnaire to assess the risk associated with liability for a specific AI system from a provider's perspective.
2405.00181
Binzhu Xie
Hang Du, Sicheng Zhang, Binzhu Xie, Guoshun Nan, Jiayang Zhang, Junrui Xu, Hangyu Liu, Sicong Leng, Jiangming Liu, Hehe Fan, Dajiu Huang, Jing Feng, Linli Chen, Can Zhang, Xuhuan Li, Hao Zhang, Jianhang Chen, Qimei Cui, Xiaofeng Tao
Uncovering What, Why and How: A Comprehensive Benchmark for Causation Understanding of Video Anomaly
Accepted in CVPR2024, Codebase: https://github.com/fesvhtr/CUVA
null
null
null
cs.CV cs.AI
http://creativecommons.org/licenses/by-nc-sa/4.0/
Video anomaly understanding (VAU) aims to automatically comprehend unusual occurrences in videos, thereby enabling various applications such as traffic surveillance and industrial manufacturing. While existing VAU benchmarks primarily concentrate on anomaly detection and localization, our focus is on more practicality, prompting us to raise the following crucial questions: "what anomaly occurred?", "why did it happen?", and "how severe is this abnormal event?". In pursuit of these answers, we present a comprehensive benchmark for Causation Understanding of Video Anomaly (CUVA). Specifically, each instance of the proposed benchmark involves three sets of human annotations to indicate the "what", "why" and "how" of an anomaly, including 1) anomaly type, start and end times, and event descriptions, 2) natural language explanations for the cause of an anomaly, and 3) free text reflecting the effect of the abnormality. In addition, we also introduce MMEval, a novel evaluation metric designed to better align with human preferences for CUVA, facilitating the measurement of existing LLMs in comprehending the underlying cause and corresponding effect of video anomalies. Finally, we propose a novel prompt-based method that can serve as a baseline approach for the challenging CUVA. We conduct extensive experiments to show the superiority of our evaluation metric and the prompt-based approach. Our code and dataset are available at https://github.com/fesvhtr/CUVA.
[ { "created": "Tue, 30 Apr 2024 20:11:49 GMT", "version": "v1" }, { "created": "Mon, 6 May 2024 14:57:50 GMT", "version": "v2" } ]
2024-05-07
[ [ "Du", "Hang", "" ], [ "Zhang", "Sicheng", "" ], [ "Xie", "Binzhu", "" ], [ "Nan", "Guoshun", "" ], [ "Zhang", "Jiayang", "" ], [ "Xu", "Junrui", "" ], [ "Liu", "Hangyu", "" ], [ "Leng", "Sicong", "" ], [ "Liu", "Jiangming", "" ], [ "Fan", "Hehe", "" ], [ "Huang", "Dajiu", "" ], [ "Feng", "Jing", "" ], [ "Chen", "Linli", "" ], [ "Zhang", "Can", "" ], [ "Li", "Xuhuan", "" ], [ "Zhang", "Hao", "" ], [ "Chen", "Jianhang", "" ], [ "Cui", "Qimei", "" ], [ "Tao", "Xiaofeng", "" ] ]
Video anomaly understanding (VAU) aims to automatically comprehend unusual occurrences in videos, thereby enabling various applications such as traffic surveillance and industrial manufacturing. While existing VAU benchmarks primarily concentrate on anomaly detection and localization, our focus is on more practicality, prompting us to raise the following crucial questions: "what anomaly occurred?", "why did it happen?", and "how severe is this abnormal event?". In pursuit of these answers, we present a comprehensive benchmark for Causation Understanding of Video Anomaly (CUVA). Specifically, each instance of the proposed benchmark involves three sets of human annotations to indicate the "what", "why" and "how" of an anomaly, including 1) anomaly type, start and end times, and event descriptions, 2) natural language explanations for the cause of an anomaly, and 3) free text reflecting the effect of the abnormality. In addition, we also introduce MMEval, a novel evaluation metric designed to better align with human preferences for CUVA, facilitating the measurement of existing LLMs in comprehending the underlying cause and corresponding effect of video anomalies. Finally, we propose a novel prompt-based method that can serve as a baseline approach for the challenging CUVA. We conduct extensive experiments to show the superiority of our evaluation metric and the prompt-based approach. Our code and dataset are available at https://github.com/fesvhtr/CUVA.
2008.02146
Santosh Vempala
He Jia, Aditi Laddha, Yin Tat Lee, Santosh S. Vempala
Reducing Isotropy and Volume to KLS: An $O(n^3\psi^2)$ Volume Algorithm
23 pages, 1 figure; updated with current KLS bound and resulting complexity
null
null
null
cs.DS cs.CC math.FA
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
We show that the volume of a convex body in ${\bf R}^{n}$ in the general membership oracle model can be computed to within relative error $\varepsilon$ using $\widetilde{O}(n^{3}\psi^{2} + n^{3}/\varepsilon^{2})$ oracle queries, where $\psi$ is the KLS constant. With the current bound of $\psi=\widetilde{O}(1)$, this gives an $\widetilde{O}(n^{3}/\varepsilon^{2})$ algorithm, improving on the Lov\'{a}sz-Vempala $\widetilde{O}(n^{4}/\varepsilon^{2})$ algorithm from 2003. The main new ingredient is an $\widetilde{O}(n^{3}\psi^{2})$ algorithm for isotropic transformation, following which we can apply the $\widetilde{O}(n^{3}/\varepsilon^{2})$ volume algorithm of Cousins and Vempala for well-rounded convex bodies. We also give an efficient implementation of the new algorithm for convex polytopes defined by $m$ inequalities in ${\bf R}^{n}$: polytope volume can be estimated in time $\widetilde{O}(mn^{c}/\varepsilon^{2})$ where $c<3.2$ depends on the current matrix multiplication exponent; this improves known bounds.
[ { "created": "Wed, 5 Aug 2020 14:08:16 GMT", "version": "v1" }, { "created": "Sat, 3 Sep 2022 11:18:21 GMT", "version": "v2" } ]
2022-09-07
[ [ "Jia", "He", "" ], [ "Laddha", "Aditi", "" ], [ "Lee", "Yin Tat", "" ], [ "Vempala", "Santosh S.", "" ] ]
We show that the volume of a convex body in ${\bf R}^{n}$ in the general membership oracle model can be computed to within relative error $\varepsilon$ using $\widetilde{O}(n^{3}\psi^{2} + n^{3}/\varepsilon^{2})$ oracle queries, where $\psi$ is the KLS constant. With the current bound of $\psi=\widetilde{O}(1)$, this gives an $\widetilde{O}(n^{3}/\varepsilon^{2})$ algorithm, improving on the Lov\'{a}sz-Vempala $\widetilde{O}(n^{4}/\varepsilon^{2})$ algorithm from 2003. The main new ingredient is an $\widetilde{O}(n^{3}\psi^{2})$ algorithm for isotropic transformation, following which we can apply the $\widetilde{O}(n^{3}/\varepsilon^{2})$ volume algorithm of Cousins and Vempala for well-rounded convex bodies. We also give an efficient implementation of the new algorithm for convex polytopes defined by $m$ inequalities in ${\bf R}^{n}$: polytope volume can be estimated in time $\widetilde{O}(mn^{c}/\varepsilon^{2})$ where $c<3.2$ depends on the current matrix multiplication exponent; this improves known bounds.
2406.18310
Jie Liu
Wenting Chen, Jie Liu, Tommy W.S. Chow, Yixuan Yuan
Spatial-temporal Hierarchical Reinforcement Learning for Interpretable Pathology Image Super-Resolution
Accepted to IEEE TRANSACTIONS ON MEDICAL IMAGING (TMI)
null
null
null
cs.CV cs.LG eess.IV
http://creativecommons.org/licenses/by/4.0/
Pathology image are essential for accurately interpreting lesion cells in cytopathology screening, but acquiring high-resolution digital slides requires specialized equipment and long scanning times. Though super-resolution (SR) techniques can alleviate this problem, existing deep learning models recover pathology image in a black-box manner, which can lead to untruthful biological details and misdiagnosis. Additionally, current methods allocate the same computational resources to recover each pixel of pathology image, leading to the sub-optimal recovery issue due to the large variation of pathology image. In this paper, we propose the first hierarchical reinforcement learning framework named Spatial-Temporal hierARchical Reinforcement Learning (STAR-RL), mainly for addressing the aforementioned issues in pathology image super-resolution problem. We reformulate the SR problem as a Markov decision process of interpretable operations and adopt the hierarchical recovery mechanism in patch level, to avoid sub-optimal recovery. Specifically, the higher-level spatial manager is proposed to pick out the most corrupted patch for the lower-level patch worker. Moreover, the higher-level temporal manager is advanced to evaluate the selected patch and determine whether the optimization should be stopped earlier, thereby avoiding the over-processed problem. Under the guidance of spatial-temporal managers, the lower-level patch worker processes the selected patch with pixel-wise interpretable actions at each time step. Experimental results on medical images degraded by different kernels show the effectiveness of STAR-RL. Furthermore, STAR-RL validates the promotion in tumor diagnosis with a large margin and shows generalizability under various degradations. The source code is available at https://github.com/CUHK-AIM-Group/STAR-RL.
[ { "created": "Wed, 26 Jun 2024 12:50:10 GMT", "version": "v1" } ]
2024-06-27
[ [ "Chen", "Wenting", "" ], [ "Liu", "Jie", "" ], [ "Chow", "Tommy W. S.", "" ], [ "Yuan", "Yixuan", "" ] ]
Pathology image are essential for accurately interpreting lesion cells in cytopathology screening, but acquiring high-resolution digital slides requires specialized equipment and long scanning times. Though super-resolution (SR) techniques can alleviate this problem, existing deep learning models recover pathology image in a black-box manner, which can lead to untruthful biological details and misdiagnosis. Additionally, current methods allocate the same computational resources to recover each pixel of pathology image, leading to the sub-optimal recovery issue due to the large variation of pathology image. In this paper, we propose the first hierarchical reinforcement learning framework named Spatial-Temporal hierARchical Reinforcement Learning (STAR-RL), mainly for addressing the aforementioned issues in pathology image super-resolution problem. We reformulate the SR problem as a Markov decision process of interpretable operations and adopt the hierarchical recovery mechanism in patch level, to avoid sub-optimal recovery. Specifically, the higher-level spatial manager is proposed to pick out the most corrupted patch for the lower-level patch worker. Moreover, the higher-level temporal manager is advanced to evaluate the selected patch and determine whether the optimization should be stopped earlier, thereby avoiding the over-processed problem. Under the guidance of spatial-temporal managers, the lower-level patch worker processes the selected patch with pixel-wise interpretable actions at each time step. Experimental results on medical images degraded by different kernels show the effectiveness of STAR-RL. Furthermore, STAR-RL validates the promotion in tumor diagnosis with a large margin and shows generalizability under various degradations. The source code is available at https://github.com/CUHK-AIM-Group/STAR-RL.
2404.08472
Emadeldeen Eldele
Emadeldeen Eldele, Mohamed Ragab, Zhenghua Chen, Min Wu, Xiaoli Li
TSLANet: Rethinking Transformers for Time Series Representation Learning
Accepted in ICML 2024
null
null
null
cs.LG stat.ML
http://creativecommons.org/licenses/by/4.0/
Time series data, characterized by its intrinsic long and short-range dependencies, poses a unique challenge across analytical applications. While Transformer-based models excel at capturing long-range dependencies, they face limitations in noise sensitivity, computational efficiency, and overfitting with smaller datasets. In response, we introduce a novel Time Series Lightweight Adaptive Network (TSLANet), as a universal convolutional model for diverse time series tasks. Specifically, we propose an Adaptive Spectral Block, harnessing Fourier analysis to enhance feature representation and to capture both long-term and short-term interactions while mitigating noise via adaptive thresholding. Additionally, we introduce an Interactive Convolution Block and leverage self-supervised learning to refine the capacity of TSLANet for decoding complex temporal patterns and improve its robustness on different datasets. Our comprehensive experiments demonstrate that TSLANet outperforms state-of-the-art models in various tasks spanning classification, forecasting, and anomaly detection, showcasing its resilience and adaptability across a spectrum of noise levels and data sizes. The code is available at https://github.com/emadeldeen24/TSLANet.
[ { "created": "Fri, 12 Apr 2024 13:41:29 GMT", "version": "v1" }, { "created": "Mon, 6 May 2024 04:00:17 GMT", "version": "v2" } ]
2024-05-07
[ [ "Eldele", "Emadeldeen", "" ], [ "Ragab", "Mohamed", "" ], [ "Chen", "Zhenghua", "" ], [ "Wu", "Min", "" ], [ "Li", "Xiaoli", "" ] ]
Time series data, characterized by its intrinsic long and short-range dependencies, poses a unique challenge across analytical applications. While Transformer-based models excel at capturing long-range dependencies, they face limitations in noise sensitivity, computational efficiency, and overfitting with smaller datasets. In response, we introduce a novel Time Series Lightweight Adaptive Network (TSLANet), as a universal convolutional model for diverse time series tasks. Specifically, we propose an Adaptive Spectral Block, harnessing Fourier analysis to enhance feature representation and to capture both long-term and short-term interactions while mitigating noise via adaptive thresholding. Additionally, we introduce an Interactive Convolution Block and leverage self-supervised learning to refine the capacity of TSLANet for decoding complex temporal patterns and improve its robustness on different datasets. Our comprehensive experiments demonstrate that TSLANet outperforms state-of-the-art models in various tasks spanning classification, forecasting, and anomaly detection, showcasing its resilience and adaptability across a spectrum of noise levels and data sizes. The code is available at https://github.com/emadeldeen24/TSLANet.
1207.4166
Trey Smith
Trey Smith, Reid Simmons
Heuristic Search Value Iteration for POMDPs
Appears in Proceedings of the Twentieth Conference on Uncertainty in Artificial Intelligence (UAI2004)
null
null
UAI-P-2004-PG-520-527
cs.AI
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
We present a novel POMDP planning algorithm called heuristic search value iteration (HSVI).HSVI is an anytime algorithm that returns a policy and a provable bound on its regret with respect to the optimal policy. HSVI gets its power by combining two well-known techniques: attention-focusing search heuristics and piecewise linear convex representations of the value function. HSVI's soundness and convergence have been proven. On some benchmark problems from the literature, HSVI displays speedups of greater than 100 with respect to other state-of-the-art POMDP value iteration algorithms. We also apply HSVI to a new rover exploration problem 10 times larger than most POMDP problems in the literature.
[ { "created": "Wed, 11 Jul 2012 15:04:47 GMT", "version": "v1" } ]
2012-07-19
[ [ "Smith", "Trey", "" ], [ "Simmons", "Reid", "" ] ]
We present a novel POMDP planning algorithm called heuristic search value iteration (HSVI).HSVI is an anytime algorithm that returns a policy and a provable bound on its regret with respect to the optimal policy. HSVI gets its power by combining two well-known techniques: attention-focusing search heuristics and piecewise linear convex representations of the value function. HSVI's soundness and convergence have been proven. On some benchmark problems from the literature, HSVI displays speedups of greater than 100 with respect to other state-of-the-art POMDP value iteration algorithms. We also apply HSVI to a new rover exploration problem 10 times larger than most POMDP problems in the literature.
2211.07738
Amanda Calatrava Arroyo
Amanda Calatrava, Hern\'an Asorey, Jan Astalos, Alberto Azevedo, Francesco Benincasa, Ignacio Blanquer, Martin Bobak, Francisco Brasileiro, Laia Cod\'o, Laura del Cano, Borja Esteban, Meritxell Ferret, Josef Handl, Tobias Kerzenmacher, Valentin Kozlov, Ale\v{s} K\v{r}enek, Ricardo Martins, Manuel Pavesio, Antonio Juan Rubio-Montero, Juan S\'anchez-Ferrero
A survey of the European Open Science Cloud services for expanding the capacity and capabilities of multidisciplinary scientific applications
null
null
null
null
cs.DC
http://creativecommons.org/licenses/by-nc-nd/4.0/
Open Science is a paradigm in which scientific data, procedures, tools and results are shared transparently and reused by society as a whole. The initiative known as the European Open Science Cloud (EOSC) is an effort in Europe to provide an open, trusted, virtual and federated computing environment to execute scientific applications, and to store, share and re-use research data across borders and scientific disciplines. Additionally, scientific services are becoming increasingly data-intensive, not only in terms of computationally intensive tasks but also in terms of storage resources. Computing paradigms such as High Performance Computing (HPC) and Cloud Computing are applied to e-science applications to meet these demands. However, adapting applications and services to these paradigms is not a trivial task, commonly requiring a deep knowledge of the underlying technologies, which often constitutes a barrier for its uptake by scientists in general. In this context, EOSC-SYNERGY, a collaborative project involving more than 20 institutions from eight European countries pooling their knowledge and experience to enhance EOSC's capabilities and capacities, aims to bring EOSC closer to the scientific communities. This article provides a summary analysis of the adaptations made in the ten thematic services of EOSC-SYNERGY to embrace this paradigm. These services are grouped into four categories: Earth Observation, Environment, Biomedicine, and Astrophysics. The analysis will lead to the identification of commonalities, best practices and common requirements, regardless of the thematic area of the service. Experience gained from the thematic services could be transferred to new services for the adoption of the EOSC ecosystem framework.
[ { "created": "Mon, 14 Nov 2022 20:33:27 GMT", "version": "v1" } ]
2022-11-16
[ [ "Calatrava", "Amanda", "" ], [ "Asorey", "Hernán", "" ], [ "Astalos", "Jan", "" ], [ "Azevedo", "Alberto", "" ], [ "Benincasa", "Francesco", "" ], [ "Blanquer", "Ignacio", "" ], [ "Bobak", "Martin", "" ], [ "Brasileiro", "Francisco", "" ], [ "Codó", "Laia", "" ], [ "del Cano", "Laura", "" ], [ "Esteban", "Borja", "" ], [ "Ferret", "Meritxell", "" ], [ "Handl", "Josef", "" ], [ "Kerzenmacher", "Tobias", "" ], [ "Kozlov", "Valentin", "" ], [ "Křenek", "Aleš", "" ], [ "Martins", "Ricardo", "" ], [ "Pavesio", "Manuel", "" ], [ "Rubio-Montero", "Antonio Juan", "" ], [ "Sánchez-Ferrero", "Juan", "" ] ]
Open Science is a paradigm in which scientific data, procedures, tools and results are shared transparently and reused by society as a whole. The initiative known as the European Open Science Cloud (EOSC) is an effort in Europe to provide an open, trusted, virtual and federated computing environment to execute scientific applications, and to store, share and re-use research data across borders and scientific disciplines. Additionally, scientific services are becoming increasingly data-intensive, not only in terms of computationally intensive tasks but also in terms of storage resources. Computing paradigms such as High Performance Computing (HPC) and Cloud Computing are applied to e-science applications to meet these demands. However, adapting applications and services to these paradigms is not a trivial task, commonly requiring a deep knowledge of the underlying technologies, which often constitutes a barrier for its uptake by scientists in general. In this context, EOSC-SYNERGY, a collaborative project involving more than 20 institutions from eight European countries pooling their knowledge and experience to enhance EOSC's capabilities and capacities, aims to bring EOSC closer to the scientific communities. This article provides a summary analysis of the adaptations made in the ten thematic services of EOSC-SYNERGY to embrace this paradigm. These services are grouped into four categories: Earth Observation, Environment, Biomedicine, and Astrophysics. The analysis will lead to the identification of commonalities, best practices and common requirements, regardless of the thematic area of the service. Experience gained from the thematic services could be transferred to new services for the adoption of the EOSC ecosystem framework.
2402.00086
Wenguan Wang
Xu Zhang and Yiming Mo and Wenguan Wang and Yi Yang
Retrosynthesis prediction enhanced by in-silico reaction data augmentation
null
null
null
null
cs.LG cs.AI
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Recent advances in machine learning (ML) have expedited retrosynthesis research by assisting chemists to design experiments more efficiently. However, all ML-based methods consume substantial amounts of paired training data (i.e., chemical reaction: product-reactant(s) pair), which is costly to obtain. Moreover, companies view reaction data as a valuable asset and restrict the accessibility to researchers. These issues prevent the creation of more powerful retrosynthesis models due to their data-driven nature. As a response, we exploit easy-to-access unpaired data (i.e., one component of product-reactant(s) pair) for generating in-silico paired data to facilitate model training. Specifically, we present RetroWISE, a self-boosting framework that employs a base model inferred from real paired data to perform in-silico reaction generation and augmentation using unpaired data, ultimately leading to a superior model. On three benchmark datasets, RetroWISE achieves the best overall performance against state-of-the-art models (e.g., +8.6% top-1 accuracy on the USPTO-50K test dataset). Moreover, it consistently improves the prediction accuracy of rare transformations. These results show that Retro- WISE overcomes the training bottleneck by in-silico reactions, thereby paving the way toward more effective ML-based retrosynthesis models.
[ { "created": "Wed, 31 Jan 2024 07:40:37 GMT", "version": "v1" } ]
2024-02-02
[ [ "Zhang", "Xu", "" ], [ "Mo", "Yiming", "" ], [ "Wang", "Wenguan", "" ], [ "Yang", "Yi", "" ] ]
Recent advances in machine learning (ML) have expedited retrosynthesis research by assisting chemists to design experiments more efficiently. However, all ML-based methods consume substantial amounts of paired training data (i.e., chemical reaction: product-reactant(s) pair), which is costly to obtain. Moreover, companies view reaction data as a valuable asset and restrict the accessibility to researchers. These issues prevent the creation of more powerful retrosynthesis models due to their data-driven nature. As a response, we exploit easy-to-access unpaired data (i.e., one component of product-reactant(s) pair) for generating in-silico paired data to facilitate model training. Specifically, we present RetroWISE, a self-boosting framework that employs a base model inferred from real paired data to perform in-silico reaction generation and augmentation using unpaired data, ultimately leading to a superior model. On three benchmark datasets, RetroWISE achieves the best overall performance against state-of-the-art models (e.g., +8.6% top-1 accuracy on the USPTO-50K test dataset). Moreover, it consistently improves the prediction accuracy of rare transformations. These results show that Retro- WISE overcomes the training bottleneck by in-silico reactions, thereby paving the way toward more effective ML-based retrosynthesis models.
1811.12065
Lile Cai
Lile Cai, Anne-Maelle Barneche, Arthur Herbout, Chuan Sheng Foo, Jie Lin, Vijay Ramaseshan Chandrasekhar and Mohamed M. Sabry
TEA-DNN: the Quest for Time-Energy-Accuracy Co-optimized Deep Neural Networks
Accepted by ISLPED2019
null
null
null
cs.NE cs.LG
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Embedded deep learning platforms have witnessed two simultaneous improvements. First, the accuracy of convolutional neural networks (CNNs) has been significantly improved through the use of automated neural-architecture search (NAS) algorithms to determine CNN structure. Second, there has been increasing interest in developing hardware accelerators for CNNs that provide improved inference performance and energy consumption compared to GPUs. Such embedded deep learning platforms differ in the amount of compute resources and memory-access bandwidth, which would affect performance and energy consumption of CNNs. It is therefore critical to consider the available hardware resources in the network architecture search. To this end, we introduce TEA-DNN, a NAS algorithm targeting multi-objective optimization of execution time, energy consumption, and classification accuracy of CNN workloads on embedded architectures. TEA-DNN leverages energy and execution time measurements on embedded hardware when exploring the Pareto-optimal curves across accuracy, execution time, and energy consumption and does not require additional effort to model the underlying hardware. We apply TEA-DNN for image classification on actual embedded platforms (NVIDIA Jetson TX2 and Intel Movidius Neural Compute Stick). We highlight the Pareto-optimal operating points that emphasize the necessity to explicitly consider hardware characteristics in the search process. To the best of our knowledge, this is the most comprehensive study of Pareto-optimal models across a range of hardware platforms using actual measurements on hardware to obtain objective values.
[ { "created": "Thu, 29 Nov 2018 11:05:28 GMT", "version": "v1" }, { "created": "Mon, 21 Oct 2019 07:39:19 GMT", "version": "v2" } ]
2019-10-22
[ [ "Cai", "Lile", "" ], [ "Barneche", "Anne-Maelle", "" ], [ "Herbout", "Arthur", "" ], [ "Foo", "Chuan Sheng", "" ], [ "Lin", "Jie", "" ], [ "Chandrasekhar", "Vijay Ramaseshan", "" ], [ "Sabry", "Mohamed M.", "" ] ]
Embedded deep learning platforms have witnessed two simultaneous improvements. First, the accuracy of convolutional neural networks (CNNs) has been significantly improved through the use of automated neural-architecture search (NAS) algorithms to determine CNN structure. Second, there has been increasing interest in developing hardware accelerators for CNNs that provide improved inference performance and energy consumption compared to GPUs. Such embedded deep learning platforms differ in the amount of compute resources and memory-access bandwidth, which would affect performance and energy consumption of CNNs. It is therefore critical to consider the available hardware resources in the network architecture search. To this end, we introduce TEA-DNN, a NAS algorithm targeting multi-objective optimization of execution time, energy consumption, and classification accuracy of CNN workloads on embedded architectures. TEA-DNN leverages energy and execution time measurements on embedded hardware when exploring the Pareto-optimal curves across accuracy, execution time, and energy consumption and does not require additional effort to model the underlying hardware. We apply TEA-DNN for image classification on actual embedded platforms (NVIDIA Jetson TX2 and Intel Movidius Neural Compute Stick). We highlight the Pareto-optimal operating points that emphasize the necessity to explicitly consider hardware characteristics in the search process. To the best of our knowledge, this is the most comprehensive study of Pareto-optimal models across a range of hardware platforms using actual measurements on hardware to obtain objective values.
1212.3540
Dima Kagan
Yehonatan Bitton, Michael Fire, Dima Kagan, Bracha Shapira, Lior Rokach, Judit Bar-Ilan
Social Network Based Search for Experts
Participated in HCIR 2012
null
null
null
cs.SI cs.HC cs.IR physics.soc-ph
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Our system illustrates how information retrieved from social networks can be used for suggesting experts for specific tasks. The system is designed to facilitate the task of finding the appropriate person(s) for a job, as a conference committee member, an advisor, etc. This short description will demonstrate how the system works in the context of the HCIR2012 published tasks.
[ { "created": "Fri, 14 Dec 2012 17:35:31 GMT", "version": "v1" } ]
2012-12-17
[ [ "Bitton", "Yehonatan", "" ], [ "Fire", "Michael", "" ], [ "Kagan", "Dima", "" ], [ "Shapira", "Bracha", "" ], [ "Rokach", "Lior", "" ], [ "Bar-Ilan", "Judit", "" ] ]
Our system illustrates how information retrieved from social networks can be used for suggesting experts for specific tasks. The system is designed to facilitate the task of finding the appropriate person(s) for a job, as a conference committee member, an advisor, etc. This short description will demonstrate how the system works in the context of the HCIR2012 published tasks.
2304.01472
Xinru Zhang
Xinru Zhang, Ni Ou, Chenghao Liu, Zhizheng Zhuo, Yaou Liu, and Chuyang Ye
Unsupervised Brain Tumor Segmentation with Image-based Prompts
Currently under review (from November 14th, 2022 until now)
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Automated brain tumor segmentation based on deep learning (DL) has achieved promising performance. However, it generally relies on annotated images for model training, which is not always feasible in clinical settings. Therefore, the development of unsupervised DL-based brain tumor segmentation approaches without expert annotations is desired. Motivated by the success of prompt learning (PL) in natural language processing, we propose an approach to unsupervised brain tumor segmentation by designing image-based prompts that allow indication of brain tumors, and this approach is dubbed as PL-based Brain Tumor Segmentation (PL-BTS). Specifically, instead of directly training a model for brain tumor segmentation with a large amount of annotated data, we seek to train a model that can answer the question: is a voxel in the input image associated with tumor-like hyper-/hypo-intensity? Such a model can be trained by artificially generating tumor-like hyper-/hypo-intensity on images without tumors with hand-crafted designs. Since the hand-crafted designs may be too simplistic to represent all kinds of real tumors, the trained model may overfit the simplistic hand-crafted task rather than actually answer the question of abnormality. To address this problem, we propose the use of a validation task, where we generate a different hand-crafted task to monitor overfitting. In addition, we propose PL-BTS+ that further improves PL-BTS by exploiting unannotated images with brain tumors. Compared with competing unsupervised methods, the proposed method has achieved marked improvements on both public and in-house datasets, and we have also demonstrated its possible extension to other brain lesion segmentation tasks.
[ { "created": "Tue, 4 Apr 2023 02:28:25 GMT", "version": "v1" } ]
2023-04-05
[ [ "Zhang", "Xinru", "" ], [ "Ou", "Ni", "" ], [ "Liu", "Chenghao", "" ], [ "Zhuo", "Zhizheng", "" ], [ "Liu", "Yaou", "" ], [ "Ye", "Chuyang", "" ] ]
Automated brain tumor segmentation based on deep learning (DL) has achieved promising performance. However, it generally relies on annotated images for model training, which is not always feasible in clinical settings. Therefore, the development of unsupervised DL-based brain tumor segmentation approaches without expert annotations is desired. Motivated by the success of prompt learning (PL) in natural language processing, we propose an approach to unsupervised brain tumor segmentation by designing image-based prompts that allow indication of brain tumors, and this approach is dubbed as PL-based Brain Tumor Segmentation (PL-BTS). Specifically, instead of directly training a model for brain tumor segmentation with a large amount of annotated data, we seek to train a model that can answer the question: is a voxel in the input image associated with tumor-like hyper-/hypo-intensity? Such a model can be trained by artificially generating tumor-like hyper-/hypo-intensity on images without tumors with hand-crafted designs. Since the hand-crafted designs may be too simplistic to represent all kinds of real tumors, the trained model may overfit the simplistic hand-crafted task rather than actually answer the question of abnormality. To address this problem, we propose the use of a validation task, where we generate a different hand-crafted task to monitor overfitting. In addition, we propose PL-BTS+ that further improves PL-BTS by exploiting unannotated images with brain tumors. Compared with competing unsupervised methods, the proposed method has achieved marked improvements on both public and in-house datasets, and we have also demonstrated its possible extension to other brain lesion segmentation tasks.
1306.4755
Chen Gong
Shuying Li, Chen Gong, Xiaodong Wang
Hybrid Group Decoding for Scalable Video over MIMO-OFDM Downlink Systems
null
null
null
null
cs.IT math.IT
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
We propose a scalable video broadcasting scheme over MIMO-OFDM systems. The scalable video source layers are channel encoded and modulated into independent signal streams, which are then transmitted from the allocated antennas in certain time-frequency blocks. Each receiver employs the successive group decoder to decode the signal streams of interest by treating other signal streams as interference. The transmitter performs adaptive coding and modulation, and transmission antenna and subcarrier allocation, based on the rate feedback from the receivers. We also propose a hybrid receiver that switches between the successive group decoder and the MMSE decoder depending on the rate. Extensive simulations are provided to demonstrate the performance gain of the proposed group-decoding-based scalable video broadcasting scheme over the one based on the conventional MMSE decoding.
[ { "created": "Thu, 20 Jun 2013 05:12:41 GMT", "version": "v1" } ]
2013-06-21
[ [ "Li", "Shuying", "" ], [ "Gong", "Chen", "" ], [ "Wang", "Xiaodong", "" ] ]
We propose a scalable video broadcasting scheme over MIMO-OFDM systems. The scalable video source layers are channel encoded and modulated into independent signal streams, which are then transmitted from the allocated antennas in certain time-frequency blocks. Each receiver employs the successive group decoder to decode the signal streams of interest by treating other signal streams as interference. The transmitter performs adaptive coding and modulation, and transmission antenna and subcarrier allocation, based on the rate feedback from the receivers. We also propose a hybrid receiver that switches between the successive group decoder and the MMSE decoder depending on the rate. Extensive simulations are provided to demonstrate the performance gain of the proposed group-decoding-based scalable video broadcasting scheme over the one based on the conventional MMSE decoding.
1712.01126
Huimiao Chen
Yinghao Jia, Yide Zhao, Ziyang Guo, Yu Xin, Huimiao Chen
Optimizing Electric Taxi Charging System: A Data-Driven Approach from Transport Energy Supply Chain Perspective
null
null
null
null
cs.SY math.OC
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
In the last decade, the development of electric taxis has motivated rapidly growing research interest in efficiently allocating electric charging stations in the academic literature. To address the driving pattern of electric taxis, we introduce the perspective of transport energy supply chain to capture the charging demand and to transform the charging station allocation problem to a location problem. Based on the P-median and the Min-max models, we developed a data-driven method to evaluate the system efficiency and service quality. We also conduct a case study using GPS trajectory data in Beijing, where various location strategies are evaluated from perspectives of system efficiency and service quality. Also, situations with and without congestion are comparatively evaluated.
[ { "created": "Mon, 4 Dec 2017 14:56:32 GMT", "version": "v1" } ]
2017-12-05
[ [ "Jia", "Yinghao", "" ], [ "Zhao", "Yide", "" ], [ "Guo", "Ziyang", "" ], [ "Xin", "Yu", "" ], [ "Chen", "Huimiao", "" ] ]
In the last decade, the development of electric taxis has motivated rapidly growing research interest in efficiently allocating electric charging stations in the academic literature. To address the driving pattern of electric taxis, we introduce the perspective of transport energy supply chain to capture the charging demand and to transform the charging station allocation problem to a location problem. Based on the P-median and the Min-max models, we developed a data-driven method to evaluate the system efficiency and service quality. We also conduct a case study using GPS trajectory data in Beijing, where various location strategies are evaluated from perspectives of system efficiency and service quality. Also, situations with and without congestion are comparatively evaluated.
1507.07815
Svebor Karaman
Giuseppe Lisanti and Svebor Karaman and Daniele Pezzatini and Alberto Del Bimbo
A Multi-Camera Image Processing and Visualization System for Train Safety Assessment
11 pages
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
In this paper we present a machine vision system to efficiently monitor, analyze and present visual data acquired with a railway overhead gantry equipped with multiple cameras. This solution aims to improve the safety of daily life railway transportation in a two- fold manner: (1) by providing automatic algorithms that can process large imagery of trains (2) by helping train operators to keep attention on any possible malfunction. The system is designed with the latest cutting edge, high-rate visible and thermal cameras that ob- serve a train passing under an railway overhead gantry. The machine vision system is composed of three principal modules: (1) an automatic wagon identification system, recognizing the wagon ID according to the UIC classification of railway coaches; (2) a temperature monitoring system; (3) a system for the detection, localization and visualization of the pantograph of the train. These three machine vision modules process batch trains sequences and their resulting analysis are presented to an operator using a multitouch user interface. We detail all technical aspects of our multi-camera portal: the hardware requirements, the software developed to deal with the high-frame rate cameras and ensure reliable acquisition, the algorithms proposed to solve each computer vision task, and the multitouch interaction and visualization interface. We evaluate each component of our system on a dataset recorded in an ad-hoc railway test-bed, showing the potential of our proposed portal for train safety assessment.
[ { "created": "Tue, 28 Jul 2015 15:36:24 GMT", "version": "v1" } ]
2015-07-29
[ [ "Lisanti", "Giuseppe", "" ], [ "Karaman", "Svebor", "" ], [ "Pezzatini", "Daniele", "" ], [ "Del Bimbo", "Alberto", "" ] ]
In this paper we present a machine vision system to efficiently monitor, analyze and present visual data acquired with a railway overhead gantry equipped with multiple cameras. This solution aims to improve the safety of daily life railway transportation in a two- fold manner: (1) by providing automatic algorithms that can process large imagery of trains (2) by helping train operators to keep attention on any possible malfunction. The system is designed with the latest cutting edge, high-rate visible and thermal cameras that ob- serve a train passing under an railway overhead gantry. The machine vision system is composed of three principal modules: (1) an automatic wagon identification system, recognizing the wagon ID according to the UIC classification of railway coaches; (2) a temperature monitoring system; (3) a system for the detection, localization and visualization of the pantograph of the train. These three machine vision modules process batch trains sequences and their resulting analysis are presented to an operator using a multitouch user interface. We detail all technical aspects of our multi-camera portal: the hardware requirements, the software developed to deal with the high-frame rate cameras and ensure reliable acquisition, the algorithms proposed to solve each computer vision task, and the multitouch interaction and visualization interface. We evaluate each component of our system on a dataset recorded in an ad-hoc railway test-bed, showing the potential of our proposed portal for train safety assessment.
2201.05297
Hanting Li
Hanting Li, Mingzhe Sui, Zhaoqing Zhu, Feng Zhao
MMNet: Muscle motion-guided network for micro-expression recognition
8 pages, 4 figures
Proc. 31st Int'l Joint Conf. Artificial Intelligence (IJCAI), 2022
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Facial micro-expressions (MEs) are involuntary facial motions revealing peoples real feelings and play an important role in the early intervention of mental illness, the national security, and many human-computer interaction systems. However, existing micro-expression datasets are limited and usually pose some challenges for training good classifiers. To model the subtle facial muscle motions, we propose a robust micro-expression recognition (MER) framework, namely muscle motion-guided network (MMNet). Specifically, a continuous attention (CA) block is introduced to focus on modeling local subtle muscle motion patterns with little identity information, which is different from most previous methods that directly extract features from complete video frames with much identity information. Besides, we design a position calibration (PC) module based on the vision transformer. By adding the position embeddings of the face generated by PC module at the end of the two branches, the PC module can help to add position information to facial muscle motion pattern features for the MER. Extensive experiments on three public micro-expression datasets demonstrate that our approach outperforms state-of-the-art methods by a large margin.
[ { "created": "Fri, 14 Jan 2022 04:05:49 GMT", "version": "v1" }, { "created": "Fri, 19 Aug 2022 11:24:19 GMT", "version": "v2" } ]
2022-08-22
[ [ "Li", "Hanting", "" ], [ "Sui", "Mingzhe", "" ], [ "Zhu", "Zhaoqing", "" ], [ "Zhao", "Feng", "" ] ]
Facial micro-expressions (MEs) are involuntary facial motions revealing peoples real feelings and play an important role in the early intervention of mental illness, the national security, and many human-computer interaction systems. However, existing micro-expression datasets are limited and usually pose some challenges for training good classifiers. To model the subtle facial muscle motions, we propose a robust micro-expression recognition (MER) framework, namely muscle motion-guided network (MMNet). Specifically, a continuous attention (CA) block is introduced to focus on modeling local subtle muscle motion patterns with little identity information, which is different from most previous methods that directly extract features from complete video frames with much identity information. Besides, we design a position calibration (PC) module based on the vision transformer. By adding the position embeddings of the face generated by PC module at the end of the two branches, the PC module can help to add position information to facial muscle motion pattern features for the MER. Extensive experiments on three public micro-expression datasets demonstrate that our approach outperforms state-of-the-art methods by a large margin.
2211.10202
Zhengmao He
Zhengmao He, Bin Zhao
Some problems about co-consonance of topological spaces
null
null
null
null
cs.LO math.GN
http://creativecommons.org/licenses/by/4.0/
In this paper, we first prove that the retract of a consonant space (or co-consonant space) is consonant (co-consonant). Using this result, some related results have obtained. Simultaneously, we proved that (1) the co-consonance of the Smyth powerspace implies the co-consonance of a topological space under a necessary condition; (2) the co-consonance of a topological implies the co-consonance of the smyth powerspace under some conditions; (3) if the lower powerspace is co-consonant, then the topological space is co-consonant; (4) the co-consonance of implies the co-consonance of the lower powerspace with some sufficient conditions.
[ { "created": "Fri, 18 Nov 2022 12:47:18 GMT", "version": "v1" }, { "created": "Wed, 30 Nov 2022 09:19:22 GMT", "version": "v2" } ]
2022-12-01
[ [ "He", "Zhengmao", "" ], [ "Zhao", "Bin", "" ] ]
In this paper, we first prove that the retract of a consonant space (or co-consonant space) is consonant (co-consonant). Using this result, some related results have obtained. Simultaneously, we proved that (1) the co-consonance of the Smyth powerspace implies the co-consonance of a topological space under a necessary condition; (2) the co-consonance of a topological implies the co-consonance of the smyth powerspace under some conditions; (3) if the lower powerspace is co-consonant, then the topological space is co-consonant; (4) the co-consonance of implies the co-consonance of the lower powerspace with some sufficient conditions.
2201.00434
Hanyuan Wang
Hanyuan Wang, Dima Damen, Majid Mirmehdi and Toby Perrett
TVNet: Temporal Voting Network for Action Localization
9 pages, 7 figures, 11 tables
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
We propose a Temporal Voting Network (TVNet) for action localization in untrimmed videos. This incorporates a novel Voting Evidence Module to locate temporal boundaries, more accurately, where temporal contextual evidence is accumulated to predict frame-level probabilities of start and end action boundaries. Our action-independent evidence module is incorporated within a pipeline to calculate confidence scores and action classes. We achieve an average mAP of 34.6% on ActivityNet-1.3, particularly outperforming previous methods with the highest IoU of 0.95. TVNet also achieves mAP of 56.0% when combined with PGCN and 59.1% with MUSES at 0.5 IoU on THUMOS14 and outperforms prior work at all thresholds. Our code is available at https://github.com/hanielwang/TVNet.
[ { "created": "Sun, 2 Jan 2022 23:46:18 GMT", "version": "v1" } ]
2022-01-04
[ [ "Wang", "Hanyuan", "" ], [ "Damen", "Dima", "" ], [ "Mirmehdi", "Majid", "" ], [ "Perrett", "Toby", "" ] ]
We propose a Temporal Voting Network (TVNet) for action localization in untrimmed videos. This incorporates a novel Voting Evidence Module to locate temporal boundaries, more accurately, where temporal contextual evidence is accumulated to predict frame-level probabilities of start and end action boundaries. Our action-independent evidence module is incorporated within a pipeline to calculate confidence scores and action classes. We achieve an average mAP of 34.6% on ActivityNet-1.3, particularly outperforming previous methods with the highest IoU of 0.95. TVNet also achieves mAP of 56.0% when combined with PGCN and 59.1% with MUSES at 0.5 IoU on THUMOS14 and outperforms prior work at all thresholds. Our code is available at https://github.com/hanielwang/TVNet.
2005.05743
Ertan Kaz{\i}kl{\i}
Ertan Kaz{\i}kl{\i}, Sinan Gezici and Serdar Y\"uksel
Quadratic Privacy-Signaling Games and the MMSE Information Bottleneck Problem for Gaussian Sources
16 pages, 6 figures
null
null
null
cs.IT math.IT math.OC
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
We investigate a privacy-signaling game problem in which a sender with privacy concerns observes a pair of correlated random vectors which are modeled as jointly Gaussian. The sender aims to hide one of these random vectors and convey the other one whereas the objective of the receiver is to accurately estimate both of the random vectors. We analyze these conflicting objectives in a game theoretic framework with quadratic costs where depending on the commitment conditions (of the sender), we consider Nash or Stackelberg (Bayesian persuasion) equilibria. We show that a payoff dominant Nash equilibrium among all admissible policies is attained by a set of explicitly characterized linear policies. We also show that a payoff dominant Nash equilibrium coincides with a Stackelberg equilibrium. We formulate the information bottleneck problem within our Stackelberg framework under the mean squared error distortion criterion where the information bottleneck setup has a further restriction that only one of the random variables is observed at the sender. We show that this MMSE Gaussian Information Bottleneck Problem admits a linear solution which is explicitly characterized in the paper. We provide explicit conditions on when the optimal solutions, or equilibrium solutions in the Nash setup, are informative or noninformative.
[ { "created": "Tue, 12 May 2020 13:16:05 GMT", "version": "v1" }, { "created": "Fri, 10 Jul 2020 20:44:16 GMT", "version": "v2" }, { "created": "Fri, 4 Mar 2022 22:14:52 GMT", "version": "v3" } ]
2022-03-08
[ [ "Kazıklı", "Ertan", "" ], [ "Gezici", "Sinan", "" ], [ "Yüksel", "Serdar", "" ] ]
We investigate a privacy-signaling game problem in which a sender with privacy concerns observes a pair of correlated random vectors which are modeled as jointly Gaussian. The sender aims to hide one of these random vectors and convey the other one whereas the objective of the receiver is to accurately estimate both of the random vectors. We analyze these conflicting objectives in a game theoretic framework with quadratic costs where depending on the commitment conditions (of the sender), we consider Nash or Stackelberg (Bayesian persuasion) equilibria. We show that a payoff dominant Nash equilibrium among all admissible policies is attained by a set of explicitly characterized linear policies. We also show that a payoff dominant Nash equilibrium coincides with a Stackelberg equilibrium. We formulate the information bottleneck problem within our Stackelberg framework under the mean squared error distortion criterion where the information bottleneck setup has a further restriction that only one of the random variables is observed at the sender. We show that this MMSE Gaussian Information Bottleneck Problem admits a linear solution which is explicitly characterized in the paper. We provide explicit conditions on when the optimal solutions, or equilibrium solutions in the Nash setup, are informative or noninformative.
2206.06117
Steve Mathew
Steve Mathew D A
Optimizing musical chord inversions using the cartesian coordinate system
9 pages, 5 tables
null
null
null
cs.SD eess.AS
http://creativecommons.org/licenses/by/4.0/
In classical music and in any genre of contemporary music, the tonal elements or notes used for playing are the same. The numerous possibilities of chords for a given instance in a piece make the playing, in general, very intricate, and advanced. The theory sounds quite trivial, yet the application has vast options, each leading to inarguably different outcomes, characterized by scientific and musical principles. Chords and their importance are self-explanatory. A chord is a bunch of notes played together. As far as scientists are concerned, it is a set of tonal frequencies ringing together resulting in a consonant/dissonant sound. It is well-known that the notes of a chord can be rearranged to come up with various voicings (1) of the same chord which enables a composer/player to choose the most optimal one to convey the emotion they wish to convey. Though there are numerous possibilities, it is scientific to think that there is just one appropriate voicing for a particular situation of tonal movements. In this study, we attempt to find the optimal voicings by considering chords to be points in a 3-dimensional cartesian coordinate system and further the fundamental understanding of mathematics in music theory.
[ { "created": "Fri, 10 Jun 2022 14:48:30 GMT", "version": "v1" } ]
2022-06-14
[ [ "A", "Steve Mathew D", "" ] ]
In classical music and in any genre of contemporary music, the tonal elements or notes used for playing are the same. The numerous possibilities of chords for a given instance in a piece make the playing, in general, very intricate, and advanced. The theory sounds quite trivial, yet the application has vast options, each leading to inarguably different outcomes, characterized by scientific and musical principles. Chords and their importance are self-explanatory. A chord is a bunch of notes played together. As far as scientists are concerned, it is a set of tonal frequencies ringing together resulting in a consonant/dissonant sound. It is well-known that the notes of a chord can be rearranged to come up with various voicings (1) of the same chord which enables a composer/player to choose the most optimal one to convey the emotion they wish to convey. Though there are numerous possibilities, it is scientific to think that there is just one appropriate voicing for a particular situation of tonal movements. In this study, we attempt to find the optimal voicings by considering chords to be points in a 3-dimensional cartesian coordinate system and further the fundamental understanding of mathematics in music theory.
0810.0558
Ashish Goel
Ashish Goel, Sanjeev Khanna, Brad Null
The Ratio Index for Budgeted Learning, with Applications
This paper has a substantial bug that we are trying to fix. Many thanks to Joe Halpern for pointing this bug out. Please do not cite in the meantime. Please let us know if you would like to understand and/or try to fix the bug
null
null
null
cs.DS
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
In the budgeted learning problem, we are allowed to experiment on a set of alternatives (given a fixed experimentation budget) with the goal of picking a single alternative with the largest possible expected payoff. Approximation algorithms for this problem were developed by Guha and Munagala by rounding a linear program that couples the various alternatives together. In this paper we present an index for this problem, which we call the ratio index, which also guarantees a constant factor approximation. Index-based policies have the advantage that a single number (i.e. the index) can be computed for each alternative irrespective of all other alternatives, and the alternative with the highest index is experimented upon. This is analogous to the famous Gittins index for the discounted multi-armed bandit problem. The ratio index has several interesting structural properties. First, we show that it can be computed in strongly polynomial time. Second, we show that with the appropriate discount factor, the Gittins index and our ratio index are constant factor approximations of each other, and hence the Gittins index also gives a constant factor approximation to the budgeted learning problem. Finally, we show that the ratio index can be used to create an index-based policy that achieves an O(1)-approximation for the finite horizon version of the multi-armed bandit problem. Moreover, the policy does not require any knowledge of the horizon (whereas we compare its performance against an optimal strategy that is aware of the horizon). This yields the following surprising result: there is an index-based policy that achieves an O(1)-approximation for the multi-armed bandit problem, oblivious to the underlying discount factor.
[ { "created": "Fri, 3 Oct 2008 01:37:45 GMT", "version": "v1" }, { "created": "Mon, 11 Apr 2016 18:47:16 GMT", "version": "v2" } ]
2016-04-12
[ [ "Goel", "Ashish", "" ], [ "Khanna", "Sanjeev", "" ], [ "Null", "Brad", "" ] ]
In the budgeted learning problem, we are allowed to experiment on a set of alternatives (given a fixed experimentation budget) with the goal of picking a single alternative with the largest possible expected payoff. Approximation algorithms for this problem were developed by Guha and Munagala by rounding a linear program that couples the various alternatives together. In this paper we present an index for this problem, which we call the ratio index, which also guarantees a constant factor approximation. Index-based policies have the advantage that a single number (i.e. the index) can be computed for each alternative irrespective of all other alternatives, and the alternative with the highest index is experimented upon. This is analogous to the famous Gittins index for the discounted multi-armed bandit problem. The ratio index has several interesting structural properties. First, we show that it can be computed in strongly polynomial time. Second, we show that with the appropriate discount factor, the Gittins index and our ratio index are constant factor approximations of each other, and hence the Gittins index also gives a constant factor approximation to the budgeted learning problem. Finally, we show that the ratio index can be used to create an index-based policy that achieves an O(1)-approximation for the finite horizon version of the multi-armed bandit problem. Moreover, the policy does not require any knowledge of the horizon (whereas we compare its performance against an optimal strategy that is aware of the horizon). This yields the following surprising result: there is an index-based policy that achieves an O(1)-approximation for the multi-armed bandit problem, oblivious to the underlying discount factor.
1103.1001
Xinhua Wang
Xinhua Wang, Hai Lin
Two-step differentiator for delayed signal
12 pages, 10 figures
null
null
null
cs.SY math.DS math.OC
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
This paper presents a high-order differentiator for delayed measurement signal. The proposed differentiator not only can correct the delay in signal, but aslo can estimate the undelayed derivatives. The differentiator consists of two-step algorithms with the delayed time instant. Conditions are given ensuring convergence of the estimation error for the given delay in the signals. The merits of method include its simple implementation and interesting application. Numerical simulations illustrate the effectiveness of the proposed differentiator.
[ { "created": "Sat, 5 Mar 2011 03:19:52 GMT", "version": "v1" } ]
2011-03-08
[ [ "Wang", "Xinhua", "" ], [ "Lin", "Hai", "" ] ]
This paper presents a high-order differentiator for delayed measurement signal. The proposed differentiator not only can correct the delay in signal, but aslo can estimate the undelayed derivatives. The differentiator consists of two-step algorithms with the delayed time instant. Conditions are given ensuring convergence of the estimation error for the given delay in the signals. The merits of method include its simple implementation and interesting application. Numerical simulations illustrate the effectiveness of the proposed differentiator.
1408.2293
Hugh Kennedy Dr.
Hugh L. Kennedy
Direct Digital Design of Loop-Shaping Filters for Sampled Control Systems
In addition to the brief journal paper (see v3 comments), this paper was split into 2 conference papers: "Numerical Derivation of Fading-Memory Polynomial and Sinusoidal Filters for Discrete-Time Control Systems" and "Application of Fading-Memory Polynomial Filters to the Control of an Electric Motor", to appear in Proc. 2015 IEEE Multi-Conference on Systems and Control
null
null
null
cs.SY
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
A controller design technique for shaping the frequency response of a process is described. A general linear model (GLM) is used to define the form of a lag or lead compensator in discrete time using a prescribed set of basis functions. The model is then transformed via the complex z-domain into a difference equation for a recursive digital filter with an infinite impulse response (IIR). A polynomial basis set is better for shaping the frequency response in the near-zero region; whereas a sinusoidal basis set is better for defining the response at arbitrary frequencies. The proposed compensator design method is more flexible than existing low-order approaches and more suitable than other general-purpose high-order methods. Performance of the resulting controller is compared with digital proportional-integral-differential (PID) and linear-state-space (LSS) algorithms in a real motor-control application.
[ { "created": "Mon, 11 Aug 2014 01:45:41 GMT", "version": "v1" }, { "created": "Thu, 20 Nov 2014 02:38:26 GMT", "version": "v2" }, { "created": "Mon, 19 Jan 2015 22:42:53 GMT", "version": "v3" }, { "created": "Sat, 18 Jul 2015 01:22:01 GMT", "version": "v4" } ]
2015-07-21
[ [ "Kennedy", "Hugh L.", "" ] ]
A controller design technique for shaping the frequency response of a process is described. A general linear model (GLM) is used to define the form of a lag or lead compensator in discrete time using a prescribed set of basis functions. The model is then transformed via the complex z-domain into a difference equation for a recursive digital filter with an infinite impulse response (IIR). A polynomial basis set is better for shaping the frequency response in the near-zero region; whereas a sinusoidal basis set is better for defining the response at arbitrary frequencies. The proposed compensator design method is more flexible than existing low-order approaches and more suitable than other general-purpose high-order methods. Performance of the resulting controller is compared with digital proportional-integral-differential (PID) and linear-state-space (LSS) algorithms in a real motor-control application.
2307.14199
Masoume Kazemi
Masoume Kazemi, Davood Moradkhani, Alireza Abbas Alipour
Application of Random Forest and Support Vector Machine for Investigation of Pressure Filtration Performance, a Zinc Plant Filter Cake Modeling
null
null
null
null
cs.LG
http://creativecommons.org/publicdomain/zero/1.0/
The hydrometallurgical method of zinc production involves leaching zinc from ore and then separating the solid residue from the liquid solution by pressure filtration. This separation process is very important since the solid residue contains some moisture that can reduce the amount of zinc recovered. This study modeled the pressure filtration process through Random Forest (RF) and Support Vector Machine (SVM). The models take continuous variables (extracted features) from the lab samples as inputs. Thus, regression models namely Random Forest Regression (RFR) and Support Vector Regression (SVR) were chosen. A total dataset was obtained during the pressure filtration process in two conditions: 1) Polypropylene (S1) and 2) Polyester fabrics (S2). To predict the cake moisture, solids concentration (0.2 and 0.38), temperature (35 and 65 centigrade), pH (2, 3.5, and 5), pressure, cake thickness (14, 20, 26, and 34 mm), air-blow time (2, 10 and 15 min) and filtration time were applied as input variables. The models' predictive accuracy was evaluated by the coefficient of determination (R2) parameter. The results revealed that the RFR model is superior to the SVR model for cake moisture prediction.
[ { "created": "Wed, 26 Jul 2023 13:52:53 GMT", "version": "v1" } ]
2023-07-27
[ [ "Kazemi", "Masoume", "" ], [ "Moradkhani", "Davood", "" ], [ "Alipour", "Alireza Abbas", "" ] ]
The hydrometallurgical method of zinc production involves leaching zinc from ore and then separating the solid residue from the liquid solution by pressure filtration. This separation process is very important since the solid residue contains some moisture that can reduce the amount of zinc recovered. This study modeled the pressure filtration process through Random Forest (RF) and Support Vector Machine (SVM). The models take continuous variables (extracted features) from the lab samples as inputs. Thus, regression models namely Random Forest Regression (RFR) and Support Vector Regression (SVR) were chosen. A total dataset was obtained during the pressure filtration process in two conditions: 1) Polypropylene (S1) and 2) Polyester fabrics (S2). To predict the cake moisture, solids concentration (0.2 and 0.38), temperature (35 and 65 centigrade), pH (2, 3.5, and 5), pressure, cake thickness (14, 20, 26, and 34 mm), air-blow time (2, 10 and 15 min) and filtration time were applied as input variables. The models' predictive accuracy was evaluated by the coefficient of determination (R2) parameter. The results revealed that the RFR model is superior to the SVR model for cake moisture prediction.
2310.02583
Seyed Mirvakili
Seyed Mo Mirvakili, Ehsan Haghighat, Douglas Sim
Machine Learning-Enabled Precision Position Control and Thermal Regulation in Advanced Thermal Actuators
null
null
null
null
cs.RO cs.LG
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
With their unique combination of characteristics - an energy density almost 100 times that of human muscle, and a power density of 5.3 kW/kg, similar to a jet engine's output - Nylon artificial muscles stand out as particularly apt for robotics applications. However, the necessity of integrating sensors and controllers poses a limitation to their practical usage. Here we report a constant power open-loop controller based on machine learning. We show that we can control the position of a nylon artificial muscle without external sensors. To this end, we construct a mapping from a desired displacement trajectory to a required power using an ensemble encoder-style feed-forward neural network. The neural controller is carefully trained on a physics-based denoised dataset and can be fine-tuned to accommodate various types of thermal artificial muscles, irrespective of the presence or absence of hysteresis.
[ { "created": "Wed, 4 Oct 2023 05:01:47 GMT", "version": "v1" } ]
2023-10-05
[ [ "Mirvakili", "Seyed Mo", "" ], [ "Haghighat", "Ehsan", "" ], [ "Sim", "Douglas", "" ] ]
With their unique combination of characteristics - an energy density almost 100 times that of human muscle, and a power density of 5.3 kW/kg, similar to a jet engine's output - Nylon artificial muscles stand out as particularly apt for robotics applications. However, the necessity of integrating sensors and controllers poses a limitation to their practical usage. Here we report a constant power open-loop controller based on machine learning. We show that we can control the position of a nylon artificial muscle without external sensors. To this end, we construct a mapping from a desired displacement trajectory to a required power using an ensemble encoder-style feed-forward neural network. The neural controller is carefully trained on a physics-based denoised dataset and can be fine-tuned to accommodate various types of thermal artificial muscles, irrespective of the presence or absence of hysteresis.
2002.05538
Jakkepalli Pavan Kumar
Jakkepalli Pavan Kumar and P. Venkata Subba Reddy
Algorithmic Complexity of Isolate Secure Domination in Graphs
arXiv admin note: substantial text overlap with arXiv:2002.00002; text overlap with arXiv:2001.11250
null
null
null
cs.DM cs.CC
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
A dominating set $S$ is an Isolate Dominating Set (IDS) if the induced subgraph $G[S]$ has at least one isolated vertex. In this paper, we initiate the study of new domination parameter called, isolate secure domination. An isolate dominating set $S\subseteq V$ is an isolate secure dominating set (ISDS), if for each vertex $u \in V \setminus S$, there exists a neighboring vertex $v$ of $u$ in $S$ such that $(S \setminus \{v\}) \cup \{u\}$ is an IDS of $G$. The minimum cardinality of an ISDS of $G$ is called as an isolate secure domination number, and is denoted by $\gamma_{0s}(G)$. Given a graph $ G=(V,E)$ and a positive integer $ k,$ the ISDM problem is to check whether $ G $ has an isolate secure dominating set of size at most $ k.$ We prove that ISDM is NP-complete even when restricted to bipartite graphs and split graphs. We also show that ISDM can be solved in linear time for graphs of bounded tree-width.
[ { "created": "Wed, 12 Feb 2020 07:42:51 GMT", "version": "v1" } ]
2020-02-14
[ [ "Kumar", "Jakkepalli Pavan", "" ], [ "Reddy", "P. Venkata Subba", "" ] ]
A dominating set $S$ is an Isolate Dominating Set (IDS) if the induced subgraph $G[S]$ has at least one isolated vertex. In this paper, we initiate the study of new domination parameter called, isolate secure domination. An isolate dominating set $S\subseteq V$ is an isolate secure dominating set (ISDS), if for each vertex $u \in V \setminus S$, there exists a neighboring vertex $v$ of $u$ in $S$ such that $(S \setminus \{v\}) \cup \{u\}$ is an IDS of $G$. The minimum cardinality of an ISDS of $G$ is called as an isolate secure domination number, and is denoted by $\gamma_{0s}(G)$. Given a graph $ G=(V,E)$ and a positive integer $ k,$ the ISDM problem is to check whether $ G $ has an isolate secure dominating set of size at most $ k.$ We prove that ISDM is NP-complete even when restricted to bipartite graphs and split graphs. We also show that ISDM can be solved in linear time for graphs of bounded tree-width.
2206.12343
Robin Haunschild
Lutz Bornmann and Robin Haunschild
Identification of young talented individuals in the natural and life sciences using bibliometric data
7 pages, 3 tables, to be presented at STI 2022
null
10.1016/j.joi.2023.101394
null
cs.DL
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Identification of young talented individuals is not an easy task. Citation-based data usually need too long to accrue. In this study, we proposed a method based on bibliometric data for the identification of young talented individuals. Three different indicators and their combinations were used. An older cohort with their first publication between 1999 and 2003 was used to find the most suitable indicator combination. For the validation step, citation impact on the level of individual papers was used. The best performing indicator combination was applied to the time period 2007-2011 for identifying young talented individuals who published their first paper within this time period. We produced a set of 46,200 potential talented individuals.
[ { "created": "Fri, 24 Jun 2022 15:28:03 GMT", "version": "v1" } ]
2024-08-12
[ [ "Bornmann", "Lutz", "" ], [ "Haunschild", "Robin", "" ] ]
Identification of young talented individuals is not an easy task. Citation-based data usually need too long to accrue. In this study, we proposed a method based on bibliometric data for the identification of young talented individuals. Three different indicators and their combinations were used. An older cohort with their first publication between 1999 and 2003 was used to find the most suitable indicator combination. For the validation step, citation impact on the level of individual papers was used. The best performing indicator combination was applied to the time period 2007-2011 for identifying young talented individuals who published their first paper within this time period. We produced a set of 46,200 potential talented individuals.
1911.08339
Jonathan Ullman
Alexander Edmonds and Aleksandar Nikolov and Jonathan Ullman
The Power of Factorization Mechanisms in Local and Central Differential Privacy
null
null
null
null
cs.DS cs.CR cs.LG
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
We give new characterizations of the sample complexity of answering linear queries (statistical queries) in the local and central models of differential privacy: *In the non-interactive local model, we give the first approximate characterization of the sample complexity. Informally our bounds are tight to within polylogarithmic factors in the number of queries and desired accuracy. Our characterization extends to agnostic learning in the local model. *In the central model, we give a characterization of the sample complexity in the high-accuracy regime that is analogous to that of Nikolov, Talwar, and Zhang (STOC 2013), but is both quantitatively tighter and has a dramatically simpler proof. Our lower bounds apply equally to the empirical and population estimation problems. In both cases, our characterizations show that a particular factorization mechanism is approximately optimal, and the optimal sample complexity is bounded from above and below by well studied factorization norms of a matrix associated with the queries.
[ { "created": "Tue, 19 Nov 2019 15:17:18 GMT", "version": "v1" } ]
2019-11-20
[ [ "Edmonds", "Alexander", "" ], [ "Nikolov", "Aleksandar", "" ], [ "Ullman", "Jonathan", "" ] ]
We give new characterizations of the sample complexity of answering linear queries (statistical queries) in the local and central models of differential privacy: *In the non-interactive local model, we give the first approximate characterization of the sample complexity. Informally our bounds are tight to within polylogarithmic factors in the number of queries and desired accuracy. Our characterization extends to agnostic learning in the local model. *In the central model, we give a characterization of the sample complexity in the high-accuracy regime that is analogous to that of Nikolov, Talwar, and Zhang (STOC 2013), but is both quantitatively tighter and has a dramatically simpler proof. Our lower bounds apply equally to the empirical and population estimation problems. In both cases, our characterizations show that a particular factorization mechanism is approximately optimal, and the optimal sample complexity is bounded from above and below by well studied factorization norms of a matrix associated with the queries.
1705.06086
Sebastian Werner
Sebastian Werner, Zdravko Velinov, Wenzel Jakob, Matthias B. Hullin
Scratch iridescence: Wave-optical rendering of diffractive surface structure
null
null
null
null
cs.GR
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
The surface of metal, glass and plastic objects is often characterized by microscopic scratches caused by manufacturing and/or wear. A closer look onto such scratches reveals iridescent colors with a complex dependency on viewing and lighting conditions. The physics behind this phenomenon is well understood; it is caused by diffraction of the incident light by surface features on the order of the optical wavelength. Existing analytic models are able to reproduce spatially unresolved microstructure such as the iridescent appearance of compact disks and similar materials. Spatially resolved scratches, on the other hand, have proven elusive due to the highly complex wave-optical light transport simulations needed to account for their appearance. In this paper, we propose a wave-optical shading model based on non-paraxial scalar diffraction theory to render this class of effects. Our model expresses surface roughness as a collection of line segments. To shade a point on the surface, the individual diffraction patterns for contributing scratch segments are computed analytically and superimposed coherently. This provides natural transitions from localized glint-like iridescence to smooth BRDFs representing the superposition of many reflections at large viewing distances. We demonstrate that our model is capable of recreating the overall appearance as well as characteristic detail effects observed on real-world examples.
[ { "created": "Wed, 17 May 2017 10:59:29 GMT", "version": "v1" } ]
2017-05-18
[ [ "Werner", "Sebastian", "" ], [ "Velinov", "Zdravko", "" ], [ "Jakob", "Wenzel", "" ], [ "Hullin", "Matthias B.", "" ] ]
The surface of metal, glass and plastic objects is often characterized by microscopic scratches caused by manufacturing and/or wear. A closer look onto such scratches reveals iridescent colors with a complex dependency on viewing and lighting conditions. The physics behind this phenomenon is well understood; it is caused by diffraction of the incident light by surface features on the order of the optical wavelength. Existing analytic models are able to reproduce spatially unresolved microstructure such as the iridescent appearance of compact disks and similar materials. Spatially resolved scratches, on the other hand, have proven elusive due to the highly complex wave-optical light transport simulations needed to account for their appearance. In this paper, we propose a wave-optical shading model based on non-paraxial scalar diffraction theory to render this class of effects. Our model expresses surface roughness as a collection of line segments. To shade a point on the surface, the individual diffraction patterns for contributing scratch segments are computed analytically and superimposed coherently. This provides natural transitions from localized glint-like iridescence to smooth BRDFs representing the superposition of many reflections at large viewing distances. We demonstrate that our model is capable of recreating the overall appearance as well as characteristic detail effects observed on real-world examples.
1204.2101
Sumit Katiyar
Sumit Katiyar, R. K. Jain, N. K. Agrawal
R.F. Pollution Reduction in Cellular Communication
6 pages, 7 figures, international journal, International Journal of Scientific & Engineering Research, Volume 3, Issue 3, March -2012
null
null
null
cs.CY cs.NI
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
R. F. pollution has been recognized as health hazard in India in the prevailing circumstances. There is lot of hue and cry against cellular towers installed in residential area. Recently high court in India has issued an order not to install towers in residential areas. For meeting the exponential demand of cellular communication in India this will be a set back for future growth. An appropriate solution has to be developed for meeting demand as well as RF pollution concern of the society. This paper deals with the installation of low power base stations in residential areas instead of high power macro cell base stations. Macro stations are proposed to be used for fast traffic, low power micro cell for a slow traffic / pedestrian and pico cell / femto cell for indoor use. These cells will be in hierarchical structure along with adaptive frequency allocation techniques and A-SDMA approach.
[ { "created": "Tue, 10 Apr 2012 10:45:12 GMT", "version": "v1" } ]
2012-04-11
[ [ "Katiyar", "Sumit", "" ], [ "Jain", "R. K.", "" ], [ "Agrawal", "N. K.", "" ] ]
R. F. pollution has been recognized as health hazard in India in the prevailing circumstances. There is lot of hue and cry against cellular towers installed in residential area. Recently high court in India has issued an order not to install towers in residential areas. For meeting the exponential demand of cellular communication in India this will be a set back for future growth. An appropriate solution has to be developed for meeting demand as well as RF pollution concern of the society. This paper deals with the installation of low power base stations in residential areas instead of high power macro cell base stations. Macro stations are proposed to be used for fast traffic, low power micro cell for a slow traffic / pedestrian and pico cell / femto cell for indoor use. These cells will be in hierarchical structure along with adaptive frequency allocation techniques and A-SDMA approach.
1211.6918
Mathis Seidl
Mathis Seidl, Andreas Schenk, Clemens Stierstorfer, and Johannes B. Huber
Aspects of Polar-Coded Modulation
Accepted for presentation at International ITG Conference on Systems, Communications and Coding, Munich, Germany, January 2013
null
null
null
cs.IT math.IT
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
We consider the joint design of polar coding and higher-order modulation schemes for ever increased spectral efficiency. The close connection between the polar code construction and the multi-level coding approach is described in detail. Relations between different modulation schemes such as bit-interleaved coded modulation (BICM) and multi-level coding (MLC) in case of polar-coded modulation as well as the influence of the applied labeling rule and the selection of frozen channels are demonstrated.
[ { "created": "Thu, 29 Nov 2012 14:01:23 GMT", "version": "v1" } ]
2012-11-30
[ [ "Seidl", "Mathis", "" ], [ "Schenk", "Andreas", "" ], [ "Stierstorfer", "Clemens", "" ], [ "Huber", "Johannes B.", "" ] ]
We consider the joint design of polar coding and higher-order modulation schemes for ever increased spectral efficiency. The close connection between the polar code construction and the multi-level coding approach is described in detail. Relations between different modulation schemes such as bit-interleaved coded modulation (BICM) and multi-level coding (MLC) in case of polar-coded modulation as well as the influence of the applied labeling rule and the selection of frozen channels are demonstrated.
2010.04554
Huiting Hong
Yucheng Lin, Huiting Hong, Xiaoqing Yang, Xiaodi Yang, Pinghua Gong, Jieping Ye
Meta Graph Attention on Heterogeneous Graph with Node-Edge Co-evolution
11pages, 4figures
null
null
null
cs.LG cs.SI stat.ML
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Graph neural networks have become an important tool for modeling structured data. In many real-world systems, intricate hidden information may exist, e.g., heterogeneity in nodes/edges, static node/edge attributes, and spatiotemporal node/edge features. However, most existing methods only take part of the information into consideration. In this paper, we present the Co-evolved Meta Graph Neural Network (CoMGNN), which applies meta graph attention to heterogeneous graphs with co-evolution of node and edge states. We further propose a spatiotemporal adaption of CoMGNN (ST-CoMGNN) for modeling spatiotemporal patterns on nodes and edges. We conduct experiments on two large-scale real-world datasets. Experimental results show that our models significantly outperform the state-of-the-art methods, demonstrating the effectiveness of encoding diverse information from different aspects.
[ { "created": "Fri, 9 Oct 2020 13:19:39 GMT", "version": "v1" } ]
2020-10-12
[ [ "Lin", "Yucheng", "" ], [ "Hong", "Huiting", "" ], [ "Yang", "Xiaoqing", "" ], [ "Yang", "Xiaodi", "" ], [ "Gong", "Pinghua", "" ], [ "Ye", "Jieping", "" ] ]
Graph neural networks have become an important tool for modeling structured data. In many real-world systems, intricate hidden information may exist, e.g., heterogeneity in nodes/edges, static node/edge attributes, and spatiotemporal node/edge features. However, most existing methods only take part of the information into consideration. In this paper, we present the Co-evolved Meta Graph Neural Network (CoMGNN), which applies meta graph attention to heterogeneous graphs with co-evolution of node and edge states. We further propose a spatiotemporal adaption of CoMGNN (ST-CoMGNN) for modeling spatiotemporal patterns on nodes and edges. We conduct experiments on two large-scale real-world datasets. Experimental results show that our models significantly outperform the state-of-the-art methods, demonstrating the effectiveness of encoding diverse information from different aspects.
1802.07647
Christian Konrad
Christian Konrad
MIS in the Congested Clique Model in $O(\log \log \Delta)$ Rounds
null
null
null
null
cs.DC cs.DS
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
We give a maximal independent set (MIS) algorithm that runs in $O(\log \log \Delta)$ rounds in the congested clique model, where $\Delta$ is the maximum degree of the input graph. This improves upon the $O(\frac{\log(\Delta) \cdot \log \log \Delta}{\sqrt{\log n}} + \log \log \Delta )$ rounds algorithm of [Ghaffari, PODC '17], where $n$ is the number of vertices of the input graph. In the first stage of our algorithm, we simulate the first $O(\frac{n}{\text{poly} \log n})$ iterations of the sequential random order Greedy algorithm for MIS in the congested clique model in $O(\log \log \Delta)$ rounds. This thins out the input graph relatively quickly: After this stage, the maximum degree of the residual graph is poly-logarithmic. In the second stage, we run the MIS algorithm of [Ghaffari, PODC '17] on the residual graph, which completes in $O(\log \log \Delta)$ rounds on graphs of poly-logarithmic degree.
[ { "created": "Wed, 21 Feb 2018 16:21:34 GMT", "version": "v1" } ]
2018-02-22
[ [ "Konrad", "Christian", "" ] ]
We give a maximal independent set (MIS) algorithm that runs in $O(\log \log \Delta)$ rounds in the congested clique model, where $\Delta$ is the maximum degree of the input graph. This improves upon the $O(\frac{\log(\Delta) \cdot \log \log \Delta}{\sqrt{\log n}} + \log \log \Delta )$ rounds algorithm of [Ghaffari, PODC '17], where $n$ is the number of vertices of the input graph. In the first stage of our algorithm, we simulate the first $O(\frac{n}{\text{poly} \log n})$ iterations of the sequential random order Greedy algorithm for MIS in the congested clique model in $O(\log \log \Delta)$ rounds. This thins out the input graph relatively quickly: After this stage, the maximum degree of the residual graph is poly-logarithmic. In the second stage, we run the MIS algorithm of [Ghaffari, PODC '17] on the residual graph, which completes in $O(\log \log \Delta)$ rounds on graphs of poly-logarithmic degree.
0910.2632
Laurent Romary
Laurent Romary (INRIA Saclay - Ile de France, IDSL)
Communication scientifique : Pour le meilleur et pour le PEER
null
Hermes (2009)
null
null
cs.DL
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
This paper provides an overview (in French) of the European PEER project, focusing on its origins, the actual objectives and the technical deployment.
[ { "created": "Wed, 14 Oct 2009 14:28:28 GMT", "version": "v1" } ]
2009-10-15
[ [ "Romary", "Laurent", "", "INRIA Saclay - Ile de France, IDSL" ] ]
This paper provides an overview (in French) of the European PEER project, focusing on its origins, the actual objectives and the technical deployment.
2003.05420
Peng Jiang Dr.
Guangnan Wu and Zhiyi Pan and Peng Jiang and Changhe Tu
Bi-Directional Attention for Joint Instance and Semantic Segmentation in Point Clouds
null
null
null
null
cs.CV
http://creativecommons.org/licenses/by-nc-sa/4.0/
Instance segmentation in point clouds is one of the most fine-grained ways to understand the 3D scene. Due to its close relationship to semantic segmentation, many works approach these two tasks simultaneously and leverage the benefits of multi-task learning. However, most of them only considered simple strategies such as element-wise feature fusion, which may not lead to mutual promotion. In this work, we build a Bi-Directional Attention module on backbone neural networks for 3D point cloud perception, which uses similarity matrix measured from features for one task to help aggregate non-local information for the other task, avoiding the potential feature exclusion and task conflict. From comprehensive experiments and ablation studies on the S3DIS dataset and the PartNet dataset, the superiority of our method is verified. Moreover, the mechanism of how bi-directional attention module helps joint instance and semantic segmentation is also analyzed.
[ { "created": "Wed, 11 Mar 2020 17:16:07 GMT", "version": "v1" } ]
2020-03-12
[ [ "Wu", "Guangnan", "" ], [ "Pan", "Zhiyi", "" ], [ "Jiang", "Peng", "" ], [ "Tu", "Changhe", "" ] ]
Instance segmentation in point clouds is one of the most fine-grained ways to understand the 3D scene. Due to its close relationship to semantic segmentation, many works approach these two tasks simultaneously and leverage the benefits of multi-task learning. However, most of them only considered simple strategies such as element-wise feature fusion, which may not lead to mutual promotion. In this work, we build a Bi-Directional Attention module on backbone neural networks for 3D point cloud perception, which uses similarity matrix measured from features for one task to help aggregate non-local information for the other task, avoiding the potential feature exclusion and task conflict. From comprehensive experiments and ablation studies on the S3DIS dataset and the PartNet dataset, the superiority of our method is verified. Moreover, the mechanism of how bi-directional attention module helps joint instance and semantic segmentation is also analyzed.
2211.13557
Fernando Alonso-Fernandez
Hartwig Fronthaler, Klaus Kollreider, Josef Bigun, Julian Fierrez, Fernando Alonso-Fernandez, Javier Ortega-Garcia, Joaquin Gonzalez-Rodriguez
Fingerprint Image-Quality Estimation and its Application to Multialgorithm Verification
Published at IEEE Transactions on Information Forensics and Security
null
10.1109/TIFS.2008.920725
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Signal-quality awareness has been found to increase recognition rates and to support decisions in multisensor environments significantly. Nevertheless, automatic quality assessment is still an open issue. Here, we study the orientation tensor of fingerprint images to quantify signal impairments, such as noise, lack of structure, blur, with the help of symmetry descriptors. A strongly reduced reference is especially favorable in biometrics, but less information is not sufficient for the approach. This is also supported by numerous experiments involving a simpler quality estimator, a trained method (NFIQ), as well as the human perception of fingerprint quality on several public databases. Furthermore, quality measurements are extensively reused to adapt fusion parameters in a monomodal multialgorithm fingerprint recognition environment. In this study, several trained and nontrained score-level fusion schemes are investigated. A Bayes-based strategy for incorporating experts past performances and current quality conditions, a novel cascaded scheme for computational efficiency, besides simple fusion rules, is presented. The quantitative results favor quality awareness under all aspects, boosting recognition rates and fusing differently skilled experts efficiently as well as effectively (by training).
[ { "created": "Thu, 24 Nov 2022 12:17:49 GMT", "version": "v1" } ]
2022-11-28
[ [ "Fronthaler", "Hartwig", "" ], [ "Kollreider", "Klaus", "" ], [ "Bigun", "Josef", "" ], [ "Fierrez", "Julian", "" ], [ "Alonso-Fernandez", "Fernando", "" ], [ "Ortega-Garcia", "Javier", "" ], [ "Gonzalez-Rodriguez", "Joaquin", "" ] ]
Signal-quality awareness has been found to increase recognition rates and to support decisions in multisensor environments significantly. Nevertheless, automatic quality assessment is still an open issue. Here, we study the orientation tensor of fingerprint images to quantify signal impairments, such as noise, lack of structure, blur, with the help of symmetry descriptors. A strongly reduced reference is especially favorable in biometrics, but less information is not sufficient for the approach. This is also supported by numerous experiments involving a simpler quality estimator, a trained method (NFIQ), as well as the human perception of fingerprint quality on several public databases. Furthermore, quality measurements are extensively reused to adapt fusion parameters in a monomodal multialgorithm fingerprint recognition environment. In this study, several trained and nontrained score-level fusion schemes are investigated. A Bayes-based strategy for incorporating experts past performances and current quality conditions, a novel cascaded scheme for computational efficiency, besides simple fusion rules, is presented. The quantitative results favor quality awareness under all aspects, boosting recognition rates and fusing differently skilled experts efficiently as well as effectively (by training).
2108.09402
Bonaventure Molokwu Ph.D.
Bonaventure Chidube Molokwu, Shaon Bhatta Shuvo, Ziad Kobti, Anne Snowdon
A Multi-Task Learning Framework for COVID-19 Monitoring and Prediction of PPE Demand in Community Health Centres
6-page article/manuscript
null
null
null
cs.LG cs.AI cs.SI
http://creativecommons.org/licenses/by/4.0/
Currently, the world seeks to find appropriate mitigation techniques to control and prevent the spread of the new SARS-CoV-2. In our paper herein, we present a peculiar Multi-Task Learning framework that jointly predicts the effect of SARS-CoV-2 as well as Personal-Protective-Equipment consumption in Community Health Centres for a given populace. Predicting the effect of the virus (SARS-CoV-2), via studies and analyses, enables us to understand the nature of SARS-CoV- 2 with reference to factors that promote its growth and spread. Therefore, these foster widespread awareness; and the populace can become more proactive and cautious so as to mitigate the spread of Corona Virus Disease 2019 (COVID- 19). Furthermore, understanding and predicting the demand for Personal Protective Equipment promotes the efficiency and safety of healthcare workers in Community Health Centres. Owing to the novel nature and strains of SARS-CoV-2, relatively few literature and research exist in this regard. These existing literature have attempted to solve the problem statement(s) using either Agent-based Models, Machine Learning Models, or Mathematical Models. In view of this, our work herein adds to existing literature via modeling our problem statements as Multi- Task Learning problems. Results from our research indicate that government actions and human factors are the most significant determinants that influence the spread of SARS-CoV-2.
[ { "created": "Fri, 20 Aug 2021 23:32:41 GMT", "version": "v1" } ]
2021-08-24
[ [ "Molokwu", "Bonaventure Chidube", "" ], [ "Shuvo", "Shaon Bhatta", "" ], [ "Kobti", "Ziad", "" ], [ "Snowdon", "Anne", "" ] ]
Currently, the world seeks to find appropriate mitigation techniques to control and prevent the spread of the new SARS-CoV-2. In our paper herein, we present a peculiar Multi-Task Learning framework that jointly predicts the effect of SARS-CoV-2 as well as Personal-Protective-Equipment consumption in Community Health Centres for a given populace. Predicting the effect of the virus (SARS-CoV-2), via studies and analyses, enables us to understand the nature of SARS-CoV- 2 with reference to factors that promote its growth and spread. Therefore, these foster widespread awareness; and the populace can become more proactive and cautious so as to mitigate the spread of Corona Virus Disease 2019 (COVID- 19). Furthermore, understanding and predicting the demand for Personal Protective Equipment promotes the efficiency and safety of healthcare workers in Community Health Centres. Owing to the novel nature and strains of SARS-CoV-2, relatively few literature and research exist in this regard. These existing literature have attempted to solve the problem statement(s) using either Agent-based Models, Machine Learning Models, or Mathematical Models. In view of this, our work herein adds to existing literature via modeling our problem statements as Multi- Task Learning problems. Results from our research indicate that government actions and human factors are the most significant determinants that influence the spread of SARS-CoV-2.
2101.11744
Matthew Smart
Matthew Smart, Anton Zilman
On the mapping between Hopfield networks and Restricted Boltzmann Machines
ICLR 2021 oral paper
The 9th International Conference on Learning Representations (ICLR 2021)
null
null
cs.LG cond-mat.dis-nn cond-mat.stat-mech
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Hopfield networks (HNs) and Restricted Boltzmann Machines (RBMs) are two important models at the interface of statistical physics, machine learning, and neuroscience. Recently, there has been interest in the relationship between HNs and RBMs, due to their similarity under the statistical mechanics formalism. An exact mapping between HNs and RBMs has been previously noted for the special case of orthogonal (uncorrelated) encoded patterns. We present here an exact mapping in the case of correlated pattern HNs, which are more broadly applicable to existing datasets. Specifically, we show that any HN with $N$ binary variables and $p<N$ arbitrary binary patterns can be transformed into an RBM with $N$ binary visible variables and $p$ gaussian hidden variables. We outline the conditions under which the reverse mapping exists, and conduct experiments on the MNIST dataset which suggest the mapping provides a useful initialization to the RBM weights. We discuss extensions, the potential importance of this correspondence for the training of RBMs, and for understanding the performance of deep architectures which utilize RBMs.
[ { "created": "Wed, 27 Jan 2021 23:49:48 GMT", "version": "v1" }, { "created": "Sat, 6 Mar 2021 02:08:12 GMT", "version": "v2" } ]
2021-03-09
[ [ "Smart", "Matthew", "" ], [ "Zilman", "Anton", "" ] ]
Hopfield networks (HNs) and Restricted Boltzmann Machines (RBMs) are two important models at the interface of statistical physics, machine learning, and neuroscience. Recently, there has been interest in the relationship between HNs and RBMs, due to their similarity under the statistical mechanics formalism. An exact mapping between HNs and RBMs has been previously noted for the special case of orthogonal (uncorrelated) encoded patterns. We present here an exact mapping in the case of correlated pattern HNs, which are more broadly applicable to existing datasets. Specifically, we show that any HN with $N$ binary variables and $p<N$ arbitrary binary patterns can be transformed into an RBM with $N$ binary visible variables and $p$ gaussian hidden variables. We outline the conditions under which the reverse mapping exists, and conduct experiments on the MNIST dataset which suggest the mapping provides a useful initialization to the RBM weights. We discuss extensions, the potential importance of this correspondence for the training of RBMs, and for understanding the performance of deep architectures which utilize RBMs.
1506.07257
Jingyu Gao
Jingyu Gao, Jinfu Yang, Guanghui Wang and Mingai Li
A Novel Feature Extraction Method for Scene Recognition Based on Centered Convolutional Restricted Boltzmann Machines
22 pages, 11 figures
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Scene recognition is an important research topic in computer vision, while feature extraction is a key step of object recognition. Although classical Restricted Boltzmann machines (RBM) can efficiently represent complicated data, it is hard to handle large images due to its complexity in computation. In this paper, a novel feature extraction method, named Centered Convolutional Restricted Boltzmann Machines (CCRBM), is proposed for scene recognition. The proposed model is an improved Convolutional Restricted Boltzmann Machines (CRBM) by introducing centered factors in its learning strategy to reduce the source of instabilities. First, the visible units of the network are redefined using centered factors. Then, the hidden units are learned with a modified energy function by utilizing a distribution function, and the visible units are reconstructed using the learned hidden units. In order to achieve better generative ability, the Centered Convolutional Deep Belief Networks (CCDBN) is trained in a greedy layer-wise way. Finally, a softmax regression is incorporated for scene recognition. Extensive experimental evaluations using natural scenes, MIT-indoor scenes, and Caltech 101 datasets show that the proposed approach performs better than other counterparts in terms of stability, generalization, and discrimination. The CCDBN model is more suitable for natural scene image recognition by virtue of convolutional property.
[ { "created": "Wed, 24 Jun 2015 06:42:42 GMT", "version": "v1" } ]
2015-06-25
[ [ "Gao", "Jingyu", "" ], [ "Yang", "Jinfu", "" ], [ "Wang", "Guanghui", "" ], [ "Li", "Mingai", "" ] ]
Scene recognition is an important research topic in computer vision, while feature extraction is a key step of object recognition. Although classical Restricted Boltzmann machines (RBM) can efficiently represent complicated data, it is hard to handle large images due to its complexity in computation. In this paper, a novel feature extraction method, named Centered Convolutional Restricted Boltzmann Machines (CCRBM), is proposed for scene recognition. The proposed model is an improved Convolutional Restricted Boltzmann Machines (CRBM) by introducing centered factors in its learning strategy to reduce the source of instabilities. First, the visible units of the network are redefined using centered factors. Then, the hidden units are learned with a modified energy function by utilizing a distribution function, and the visible units are reconstructed using the learned hidden units. In order to achieve better generative ability, the Centered Convolutional Deep Belief Networks (CCDBN) is trained in a greedy layer-wise way. Finally, a softmax regression is incorporated for scene recognition. Extensive experimental evaluations using natural scenes, MIT-indoor scenes, and Caltech 101 datasets show that the proposed approach performs better than other counterparts in terms of stability, generalization, and discrimination. The CCDBN model is more suitable for natural scene image recognition by virtue of convolutional property.
2212.12603
Yao Yao
Yao Yao, Qihang Lin, Tianbao Yang
Stochastic Methods for AUC Optimization subject to AUC-based Fairness Constraints
Published in AISTATS 2023
null
null
null
cs.LG math.OC stat.ML
http://creativecommons.org/licenses/by/4.0/
As machine learning being used increasingly in making high-stakes decisions, an arising challenge is to avoid unfair AI systems that lead to discriminatory decisions for protected population. A direct approach for obtaining a fair predictive model is to train the model through optimizing its prediction performance subject to fairness constraints, which achieves Pareto efficiency when trading off performance against fairness. Among various fairness metrics, the ones based on the area under the ROC curve (AUC) are emerging recently because they are threshold-agnostic and effective for unbalanced data. In this work, we formulate the training problem of a fairness-aware machine learning model as an AUC optimization problem subject to a class of AUC-based fairness constraints. This problem can be reformulated as a min-max optimization problem with min-max constraints, which we solve by stochastic first-order methods based on a new Bregman divergence designed for the special structure of the problem. We numerically demonstrate the effectiveness of our approach on real-world data under different fairness metrics.
[ { "created": "Fri, 23 Dec 2022 22:29:08 GMT", "version": "v1" }, { "created": "Tue, 27 Dec 2022 02:01:30 GMT", "version": "v2" }, { "created": "Wed, 22 Feb 2023 21:26:56 GMT", "version": "v3" } ]
2023-02-24
[ [ "Yao", "Yao", "" ], [ "Lin", "Qihang", "" ], [ "Yang", "Tianbao", "" ] ]
As machine learning being used increasingly in making high-stakes decisions, an arising challenge is to avoid unfair AI systems that lead to discriminatory decisions for protected population. A direct approach for obtaining a fair predictive model is to train the model through optimizing its prediction performance subject to fairness constraints, which achieves Pareto efficiency when trading off performance against fairness. Among various fairness metrics, the ones based on the area under the ROC curve (AUC) are emerging recently because they are threshold-agnostic and effective for unbalanced data. In this work, we formulate the training problem of a fairness-aware machine learning model as an AUC optimization problem subject to a class of AUC-based fairness constraints. This problem can be reformulated as a min-max optimization problem with min-max constraints, which we solve by stochastic first-order methods based on a new Bregman divergence designed for the special structure of the problem. We numerically demonstrate the effectiveness of our approach on real-world data under different fairness metrics.
2302.03596
Jaehyeong Jo
Jaehyeong Jo, Dongki Kim, Sung Ju Hwang
Graph Generation with Diffusion Mixture
ICML 2024
null
null
null
cs.LG
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Generation of graphs is a major challenge for real-world tasks that require understanding the complex nature of their non-Euclidean structures. Although diffusion models have achieved notable success in graph generation recently, they are ill-suited for modeling the topological properties of graphs since learning to denoise the noisy samples does not explicitly learn the graph structures to be generated. To tackle this limitation, we propose a generative framework that models the topology of graphs by explicitly learning the final graph structures of the diffusion process. Specifically, we design the generative process as a mixture of endpoint-conditioned diffusion processes which is driven toward the predicted graph that results in rapid convergence. We further introduce a simple parameterization of the mixture process and develop an objective for learning the final graph structure, which enables maximum likelihood training. Through extensive experimental validation on general graph and 2D/3D molecule generation tasks, we show that our method outperforms previous generative models, generating graphs with correct topology with both continuous (e.g. 3D coordinates) and discrete (e.g. atom types) features. Our code is available at https://github.com/harryjo97/GruM.
[ { "created": "Tue, 7 Feb 2023 17:07:46 GMT", "version": "v1" }, { "created": "Wed, 24 May 2023 06:09:45 GMT", "version": "v2" }, { "created": "Mon, 5 Feb 2024 02:22:58 GMT", "version": "v3" }, { "created": "Sun, 2 Jun 2024 20:00:20 GMT", "version": "v4" } ]
2024-06-04
[ [ "Jo", "Jaehyeong", "" ], [ "Kim", "Dongki", "" ], [ "Hwang", "Sung Ju", "" ] ]
Generation of graphs is a major challenge for real-world tasks that require understanding the complex nature of their non-Euclidean structures. Although diffusion models have achieved notable success in graph generation recently, they are ill-suited for modeling the topological properties of graphs since learning to denoise the noisy samples does not explicitly learn the graph structures to be generated. To tackle this limitation, we propose a generative framework that models the topology of graphs by explicitly learning the final graph structures of the diffusion process. Specifically, we design the generative process as a mixture of endpoint-conditioned diffusion processes which is driven toward the predicted graph that results in rapid convergence. We further introduce a simple parameterization of the mixture process and develop an objective for learning the final graph structure, which enables maximum likelihood training. Through extensive experimental validation on general graph and 2D/3D molecule generation tasks, we show that our method outperforms previous generative models, generating graphs with correct topology with both continuous (e.g. 3D coordinates) and discrete (e.g. atom types) features. Our code is available at https://github.com/harryjo97/GruM.
2003.06945
Cho-Ying Wu
Cho-Ying Wu, Ulrich Neumann
Scene Completeness-Aware Lidar Depth Completion for Driving Scenario
Present at ICASSP 2021; fix typos
null
null
null
cs.CV cs.RO
http://creativecommons.org/licenses/by/4.0/
This paper introduces Scene Completeness-Aware Depth Completion (SCADC) to complete raw lidar scans into dense depth maps with fine and complete scene structures. Recent sparse depth completion for lidars only focuses on the lower scenes and produces irregular estimations on the upper because existing datasets, such as KITTI, do not provide groundtruth for upper areas. These areas are considered less important since they are usually sky or trees of less scene understanding interest. However, we argue that in several driving scenarios such as large trucks or cars with loads, objects could extend to the upper parts of scenes. Thus depth maps with structured upper scene estimation are important for RGBD algorithms. SCADC adopts stereo images that produce disparities with better scene completeness but are generally less precise than lidars, to help sparse lidar depth completion. To our knowledge, we are the first to focus on scene completeness of sparse depth completion. We validate our SCADC on both depth estimate precision and scene-completeness on KITTI. Moreover, we experiment on less-explored outdoor RGBD semantic segmentation with scene completeness-aware D-input to validate our method.
[ { "created": "Sun, 15 Mar 2020 23:23:26 GMT", "version": "v1" }, { "created": "Wed, 18 Mar 2020 03:46:18 GMT", "version": "v2" }, { "created": "Sat, 20 Feb 2021 23:50:16 GMT", "version": "v3" }, { "created": "Wed, 17 Jan 2024 05:29:28 GMT", "version": "v4" } ]
2024-01-18
[ [ "Wu", "Cho-Ying", "" ], [ "Neumann", "Ulrich", "" ] ]
This paper introduces Scene Completeness-Aware Depth Completion (SCADC) to complete raw lidar scans into dense depth maps with fine and complete scene structures. Recent sparse depth completion for lidars only focuses on the lower scenes and produces irregular estimations on the upper because existing datasets, such as KITTI, do not provide groundtruth for upper areas. These areas are considered less important since they are usually sky or trees of less scene understanding interest. However, we argue that in several driving scenarios such as large trucks or cars with loads, objects could extend to the upper parts of scenes. Thus depth maps with structured upper scene estimation are important for RGBD algorithms. SCADC adopts stereo images that produce disparities with better scene completeness but are generally less precise than lidars, to help sparse lidar depth completion. To our knowledge, we are the first to focus on scene completeness of sparse depth completion. We validate our SCADC on both depth estimate precision and scene-completeness on KITTI. Moreover, we experiment on less-explored outdoor RGBD semantic segmentation with scene completeness-aware D-input to validate our method.
2401.12624
Yongjun Kim
Yongjun Kim, Sejin Seo, Jihong Park, Mehdi Bennis, Seong-Lyun Kim, Junil Choi
Knowledge Distillation from Language-Oriented to Emergent Communication for Multi-Agent Remote Control
null
null
null
null
cs.AI cs.IT cs.LG cs.NI math.IT
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
In this work, we compare emergent communication (EC) built upon multi-agent deep reinforcement learning (MADRL) and language-oriented semantic communication (LSC) empowered by a pre-trained large language model (LLM) using human language. In a multi-agent remote navigation task, with multimodal input data comprising location and channel maps, it is shown that EC incurs high training cost and struggles when using multimodal data, whereas LSC yields high inference computing cost due to the LLM's large size. To address their respective bottlenecks, we propose a novel framework of language-guided EC (LEC) by guiding the EC training using LSC via knowledge distillation (KD). Simulations corroborate that LEC achieves faster travel time while avoiding areas with poor channel conditions, as well as speeding up the MADRL training convergence by up to 61.8% compared to EC.
[ { "created": "Tue, 23 Jan 2024 10:23:13 GMT", "version": "v1" }, { "created": "Sun, 3 Mar 2024 14:15:52 GMT", "version": "v2" } ]
2024-03-05
[ [ "Kim", "Yongjun", "" ], [ "Seo", "Sejin", "" ], [ "Park", "Jihong", "" ], [ "Bennis", "Mehdi", "" ], [ "Kim", "Seong-Lyun", "" ], [ "Choi", "Junil", "" ] ]
In this work, we compare emergent communication (EC) built upon multi-agent deep reinforcement learning (MADRL) and language-oriented semantic communication (LSC) empowered by a pre-trained large language model (LLM) using human language. In a multi-agent remote navigation task, with multimodal input data comprising location and channel maps, it is shown that EC incurs high training cost and struggles when using multimodal data, whereas LSC yields high inference computing cost due to the LLM's large size. To address their respective bottlenecks, we propose a novel framework of language-guided EC (LEC) by guiding the EC training using LSC via knowledge distillation (KD). Simulations corroborate that LEC achieves faster travel time while avoiding areas with poor channel conditions, as well as speeding up the MADRL training convergence by up to 61.8% compared to EC.
2402.08245
Manh Duong Phung
Duy Nam Bui, Manh Duong Phung, Hung Pham Duy
Self-Reconfigurable V-shape Formation of Multiple UAVs in Narrow Space Environments
Published in: 2024 IEEE/SICE International Symposium on System Integration (SII)
null
10.1109/SII58957.2024.10417519
null
cs.RO
http://creativecommons.org/licenses/by/4.0/
This paper presents the design and implementation of a self-reconfigurable V-shape formation controller for multiple unmanned aerial vehicles (UAVs) navigating through narrow spaces in a dense obstacle environment. The selection of the V-shape formation is motivated by its maneuverability and visibility advantages. The main objective is to develop an effective formation control strategy that allows UAVs to autonomously adjust their positions to form the desired formation while navigating through obstacles. To achieve this, we propose a distributed behavior-based control algorithm that combines the behaviors designed for individual UAVs so that they together navigate the UAVs to their desired positions. The reconfiguration process is automatic, utilizing individual UAV sensing within the formation, allowing for dynamic adaptations such as opening/closing wings or merging into a straight line. Simulation results show that the self-reconfigurable V-shape formation offers adaptability and effectiveness for UAV formations in complex operational scenarios.
[ { "created": "Tue, 13 Feb 2024 06:19:11 GMT", "version": "v1" } ]
2024-02-14
[ [ "Bui", "Duy Nam", "" ], [ "Phung", "Manh Duong", "" ], [ "Duy", "Hung Pham", "" ] ]
This paper presents the design and implementation of a self-reconfigurable V-shape formation controller for multiple unmanned aerial vehicles (UAVs) navigating through narrow spaces in a dense obstacle environment. The selection of the V-shape formation is motivated by its maneuverability and visibility advantages. The main objective is to develop an effective formation control strategy that allows UAVs to autonomously adjust their positions to form the desired formation while navigating through obstacles. To achieve this, we propose a distributed behavior-based control algorithm that combines the behaviors designed for individual UAVs so that they together navigate the UAVs to their desired positions. The reconfiguration process is automatic, utilizing individual UAV sensing within the formation, allowing for dynamic adaptations such as opening/closing wings or merging into a straight line. Simulation results show that the self-reconfigurable V-shape formation offers adaptability and effectiveness for UAV formations in complex operational scenarios.
2005.10266
Liang-Chieh Chen
Liang-Chieh Chen, Raphael Gontijo Lopes, Bowen Cheng, Maxwell D. Collins, Ekin D. Cubuk, Barret Zoph, Hartwig Adam, Jonathon Shlens
Naive-Student: Leveraging Semi-Supervised Learning in Video Sequences for Urban Scene Segmentation
Accepted to ECCV 2020
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Supervised learning in large discriminative models is a mainstay for modern computer vision. Such an approach necessitates investing in large-scale human-annotated datasets for achieving state-of-the-art results. In turn, the efficacy of supervised learning may be limited by the size of the human annotated dataset. This limitation is particularly notable for image segmentation tasks, where the expense of human annotation is especially large, yet large amounts of unlabeled data may exist. In this work, we ask if we may leverage semi-supervised learning in unlabeled video sequences and extra images to improve the performance on urban scene segmentation, simultaneously tackling semantic, instance, and panoptic segmentation. The goal of this work is to avoid the construction of sophisticated, learned architectures specific to label propagation (e.g., patch matching and optical flow). Instead, we simply predict pseudo-labels for the unlabeled data and train subsequent models with both human-annotated and pseudo-labeled data. The procedure is iterated for several times. As a result, our Naive-Student model, trained with such simple yet effective iterative semi-supervised learning, attains state-of-the-art results at all three Cityscapes benchmarks, reaching the performance of 67.8% PQ, 42.6% AP, and 85.2% mIOU on the test set. We view this work as a notable step towards building a simple procedure to harness unlabeled video sequences and extra images to surpass state-of-the-art performance on core computer vision tasks.
[ { "created": "Wed, 20 May 2020 18:00:05 GMT", "version": "v1" }, { "created": "Fri, 22 May 2020 04:38:50 GMT", "version": "v2" }, { "created": "Wed, 8 Jul 2020 16:29:10 GMT", "version": "v3" }, { "created": "Mon, 20 Jul 2020 03:40:38 GMT", "version": "v4" } ]
2020-07-21
[ [ "Chen", "Liang-Chieh", "" ], [ "Lopes", "Raphael Gontijo", "" ], [ "Cheng", "Bowen", "" ], [ "Collins", "Maxwell D.", "" ], [ "Cubuk", "Ekin D.", "" ], [ "Zoph", "Barret", "" ], [ "Adam", "Hartwig", "" ], [ "Shlens", "Jonathon", "" ] ]
Supervised learning in large discriminative models is a mainstay for modern computer vision. Such an approach necessitates investing in large-scale human-annotated datasets for achieving state-of-the-art results. In turn, the efficacy of supervised learning may be limited by the size of the human annotated dataset. This limitation is particularly notable for image segmentation tasks, where the expense of human annotation is especially large, yet large amounts of unlabeled data may exist. In this work, we ask if we may leverage semi-supervised learning in unlabeled video sequences and extra images to improve the performance on urban scene segmentation, simultaneously tackling semantic, instance, and panoptic segmentation. The goal of this work is to avoid the construction of sophisticated, learned architectures specific to label propagation (e.g., patch matching and optical flow). Instead, we simply predict pseudo-labels for the unlabeled data and train subsequent models with both human-annotated and pseudo-labeled data. The procedure is iterated for several times. As a result, our Naive-Student model, trained with such simple yet effective iterative semi-supervised learning, attains state-of-the-art results at all three Cityscapes benchmarks, reaching the performance of 67.8% PQ, 42.6% AP, and 85.2% mIOU on the test set. We view this work as a notable step towards building a simple procedure to harness unlabeled video sequences and extra images to surpass state-of-the-art performance on core computer vision tasks.
1806.02325
Ay\c{c}a \"Oz\c{c}elikkale
Ayca Ozcelikkale, Mehmet Koseoglu and Mani Srivastava
Optimization vs. Reinforcement Learning for Wirelessly Powered Sensor Networks
null
null
null
null
cs.IT math.IT
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
We consider a sensing application where the sensor nodes are wirelessly powered by an energy beacon. We focus on the problem of jointly optimizing the energy allocation of the energy beacon to different sensors and the data transmission powers of the sensors in order to minimize the field reconstruction error at the sink. In contrast to the standard ideal linear energy harvesting (EH) model, we consider practical non-linear EH models. We investigate this problem under two different frameworks: i) an optimization approach where the energy beacon knows the utility function of the nodes, channel state information and the energy harvesting characteristics of the devices; hence optimal power allocation strategies can be designed using an optimization problem and ii) a learning approach where the energy beacon decides on its strategies adaptively with battery level information and feedback on the utility function. Our results illustrate that deep reinforcement learning approach can obtain the same error levels with the optimization approach and provides a promising alternative to the optimization framework.
[ { "created": "Wed, 6 Jun 2018 17:43:31 GMT", "version": "v1" } ]
2018-06-07
[ [ "Ozcelikkale", "Ayca", "" ], [ "Koseoglu", "Mehmet", "" ], [ "Srivastava", "Mani", "" ] ]
We consider a sensing application where the sensor nodes are wirelessly powered by an energy beacon. We focus on the problem of jointly optimizing the energy allocation of the energy beacon to different sensors and the data transmission powers of the sensors in order to minimize the field reconstruction error at the sink. In contrast to the standard ideal linear energy harvesting (EH) model, we consider practical non-linear EH models. We investigate this problem under two different frameworks: i) an optimization approach where the energy beacon knows the utility function of the nodes, channel state information and the energy harvesting characteristics of the devices; hence optimal power allocation strategies can be designed using an optimization problem and ii) a learning approach where the energy beacon decides on its strategies adaptively with battery level information and feedback on the utility function. Our results illustrate that deep reinforcement learning approach can obtain the same error levels with the optimization approach and provides a promising alternative to the optimization framework.
2401.15545
Simin Chen
Simin Chen, Xiaoning Feng, Xiaohong Han, Cong Liu, Wei Yang
PPM: Automated Generation of Diverse Programming Problems for Benchmarking Code Generation Models
This paper has been accepted to The ACM International Conference on the Foundations of Software Engineering FSE 2024
null
null
null
cs.SE cs.AI cs.CL cs.PL
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
In recent times, a plethora of Large Code Generation Models (LCGMs) have been proposed, showcasing significant potential in assisting developers with complex programming tasks. Benchmarking LCGMs necessitates the creation of a set of diverse programming problems, and each problem comprises the prompt (including the task description), canonical solution, and test inputs. The existing methods for constructing such a problem set can be categorized into two main types: manual methods and perturbation-based methods. However, manual methods demand high effort and lack scalability, while also risking data integrity due to LCGMs' potentially contaminated data collection, and perturbation-based approaches mainly generate semantically homogeneous problems with the same canonical solutions and introduce typos that can be easily auto-corrected by IDE, making them ineffective and unrealistic. In this work, we propose the idea of programming problem merging (PPM) and provide two implementation of this idea, we utilize our tool on two widely-used datasets and compare it against nine baseline methods using eight code generation models. The results demonstrate the effectiveness of our tool in generating more challenging, diverse, and natural programming problems, comparing to the baselines.
[ { "created": "Sun, 28 Jan 2024 02:27:38 GMT", "version": "v1" } ]
2024-01-30
[ [ "Chen", "Simin", "" ], [ "Feng", "Xiaoning", "" ], [ "Han", "Xiaohong", "" ], [ "Liu", "Cong", "" ], [ "Yang", "Wei", "" ] ]
In recent times, a plethora of Large Code Generation Models (LCGMs) have been proposed, showcasing significant potential in assisting developers with complex programming tasks. Benchmarking LCGMs necessitates the creation of a set of diverse programming problems, and each problem comprises the prompt (including the task description), canonical solution, and test inputs. The existing methods for constructing such a problem set can be categorized into two main types: manual methods and perturbation-based methods. However, manual methods demand high effort and lack scalability, while also risking data integrity due to LCGMs' potentially contaminated data collection, and perturbation-based approaches mainly generate semantically homogeneous problems with the same canonical solutions and introduce typos that can be easily auto-corrected by IDE, making them ineffective and unrealistic. In this work, we propose the idea of programming problem merging (PPM) and provide two implementation of this idea, we utilize our tool on two widely-used datasets and compare it against nine baseline methods using eight code generation models. The results demonstrate the effectiveness of our tool in generating more challenging, diverse, and natural programming problems, comparing to the baselines.
2305.06236
Sara Ahmadi Majd
Mohammad Mashayekhi, Sara Ahmadi Majd, Arian Amiramjadi, Babak Mashayekhi
Radious: Unveiling the Enigma of Dental Radiology with BEIT Adaptor and Mask2Former in Semantic Segmentation
null
null
null
null
cs.CV cs.AI
http://creativecommons.org/publicdomain/zero/1.0/
X-ray images are the first steps for diagnosing and further treating dental problems. So, early diagnosis prevents the development and increase of oral and dental diseases. In this paper, we developed a semantic segmentation algorithm based on BEIT adaptor and Mask2Former to detect and identify teeth, roots, and multiple dental diseases and abnormalities such as pulp chamber, restoration, endodontics, crown, decay, pin, composite, bridge, pulpitis, orthodontics, radicular cyst, periapical cyst, cyst, implant, and bone graft material in panoramic, periapical, and bitewing X-ray images. We compared the result of our algorithm to two state-of-the-art algorithms in image segmentation named: Deeplabv3 and Segformer on our own data set. We discovered that Radious outperformed those algorithms by increasing the mIoU scores by 9% and 33% in Deeplabv3+ and Segformer, respectively.
[ { "created": "Wed, 10 May 2023 15:15:09 GMT", "version": "v1" } ]
2023-05-11
[ [ "Mashayekhi", "Mohammad", "" ], [ "Majd", "Sara Ahmadi", "" ], [ "Amiramjadi", "Arian", "" ], [ "Mashayekhi", "Babak", "" ] ]
X-ray images are the first steps for diagnosing and further treating dental problems. So, early diagnosis prevents the development and increase of oral and dental diseases. In this paper, we developed a semantic segmentation algorithm based on BEIT adaptor and Mask2Former to detect and identify teeth, roots, and multiple dental diseases and abnormalities such as pulp chamber, restoration, endodontics, crown, decay, pin, composite, bridge, pulpitis, orthodontics, radicular cyst, periapical cyst, cyst, implant, and bone graft material in panoramic, periapical, and bitewing X-ray images. We compared the result of our algorithm to two state-of-the-art algorithms in image segmentation named: Deeplabv3 and Segformer on our own data set. We discovered that Radious outperformed those algorithms by increasing the mIoU scores by 9% and 33% in Deeplabv3+ and Segformer, respectively.
1811.08366
Stephen Bonner
Stephen Bonner, John Brennan, Ibad Kureshi, Georgios Theodoropoulos, Andrew Stephen McGough, Boguslaw Obara
Temporal Graph Offset Reconstruction: Towards Temporally Robust Graph Representation Learning
Accepted as a workshop paper at IEEE Big Data 2018
null
null
null
cs.SI cs.LG
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Graphs are a commonly used construct for representing relationships between elements in complex high dimensional datasets. Many real-world phenomenon are dynamic in nature, meaning that any graph used to represent them is inherently temporal. However, many of the machine learning models designed to capture knowledge about the structure of these graphs ignore this rich temporal information when creating representations of the graph. This results in models which do not perform well when used to make predictions about the future state of the graph -- especially when the delta between time stamps is not small. In this work, we explore a novel training procedure and an associated unsupervised model which creates graph representations optimised to predict the future state of the graph. We make use of graph convolutional neural networks to encode the graph into a latent representation, which we then use to train our temporal offset reconstruction method, inspired by auto-encoders, to predict a later time point -- multiple time steps into the future. Using our method, we demonstrate superior performance for the task of future link prediction compared with none-temporal state-of-the-art baselines. We show our approach to be capable of outperforming non-temporal baselines by 38% on a real world dataset.
[ { "created": "Tue, 20 Nov 2018 17:01:16 GMT", "version": "v1" } ]
2018-11-21
[ [ "Bonner", "Stephen", "" ], [ "Brennan", "John", "" ], [ "Kureshi", "Ibad", "" ], [ "Theodoropoulos", "Georgios", "" ], [ "McGough", "Andrew Stephen", "" ], [ "Obara", "Boguslaw", "" ] ]
Graphs are a commonly used construct for representing relationships between elements in complex high dimensional datasets. Many real-world phenomenon are dynamic in nature, meaning that any graph used to represent them is inherently temporal. However, many of the machine learning models designed to capture knowledge about the structure of these graphs ignore this rich temporal information when creating representations of the graph. This results in models which do not perform well when used to make predictions about the future state of the graph -- especially when the delta between time stamps is not small. In this work, we explore a novel training procedure and an associated unsupervised model which creates graph representations optimised to predict the future state of the graph. We make use of graph convolutional neural networks to encode the graph into a latent representation, which we then use to train our temporal offset reconstruction method, inspired by auto-encoders, to predict a later time point -- multiple time steps into the future. Using our method, we demonstrate superior performance for the task of future link prediction compared with none-temporal state-of-the-art baselines. We show our approach to be capable of outperforming non-temporal baselines by 38% on a real world dataset.
2308.13094
Takamichi Miyata Ph.D.
Takamichi Miyata
Interpretable Image Quality Assessment via CLIP with Multiple Antonym-Prompt Pairs
2pages, 1 figure
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
No reference image quality assessment (NR-IQA) is a task to estimate the perceptual quality of an image without its corresponding original image. It is even more difficult to perform this task in a zero-shot manner, i.e., without task-specific training. In this paper, we propose a new zero-shot and interpretable NRIQA method that exploits the ability of a pre-trained visionlanguage model to estimate the correlation between an image and a textual prompt. The proposed method employs a prompt pairing strategy and multiple antonym-prompt pairs corresponding to carefully selected descriptive features corresponding to the perceptual image quality. Thus, the proposed method is able to identify not only the perceptual quality evaluation of the image, but also the cause on which the quality evaluation is based. Experimental results show that the proposed method outperforms existing zero-shot NR-IQA methods in terms of accuracy and can evaluate the causes of perceptual quality degradation.
[ { "created": "Thu, 24 Aug 2023 21:37:00 GMT", "version": "v1" } ]
2023-08-28
[ [ "Miyata", "Takamichi", "" ] ]
No reference image quality assessment (NR-IQA) is a task to estimate the perceptual quality of an image without its corresponding original image. It is even more difficult to perform this task in a zero-shot manner, i.e., without task-specific training. In this paper, we propose a new zero-shot and interpretable NRIQA method that exploits the ability of a pre-trained visionlanguage model to estimate the correlation between an image and a textual prompt. The proposed method employs a prompt pairing strategy and multiple antonym-prompt pairs corresponding to carefully selected descriptive features corresponding to the perceptual image quality. Thus, the proposed method is able to identify not only the perceptual quality evaluation of the image, but also the cause on which the quality evaluation is based. Experimental results show that the proposed method outperforms existing zero-shot NR-IQA methods in terms of accuracy and can evaluate the causes of perceptual quality degradation.
2305.17773
Madhav Desai
Madhav P. Desai
An evaluation of a microprocessor with two independent hardware execution threads coupled through a shared cache
null
null
null
null
cs.AR
http://creativecommons.org/licenses/by/4.0/
We investigate the utility of augmenting a microprocessor with a single execution pipeline by adding a second copy of the execution pipeline in parallel with the existing one. The resulting dual-hardware-threaded microprocessor has two identical, independent, single-issue in-order execution pipelines (hardware threads) which share a common memory sub-system (consisting of instruction and data caches together with a memory management unit). From a design perspective, the assembly and verification of the dual threaded processor is simplified by the use of existing verified implementations of the execution pipeline and a memory unit. Because the memory unit is shared by the two hardware threads, the relative area overhead of adding the second hardware thread is 25\% of the area of the existing single threaded processor. Using an FPGA implementation we evaluate the performance of the dual threaded processor relative to the single threaded one. On applications which can be parallelized, we observe speedups of 1.6X to 1.88X. For applications that are not parallelizable, the speedup is more modest. We also observe that the dual threaded processor performance is degraded on applications which generate large numbers of cache misses.
[ { "created": "Sun, 28 May 2023 16:47:56 GMT", "version": "v1" } ]
2023-05-30
[ [ "Desai", "Madhav P.", "" ] ]
We investigate the utility of augmenting a microprocessor with a single execution pipeline by adding a second copy of the execution pipeline in parallel with the existing one. The resulting dual-hardware-threaded microprocessor has two identical, independent, single-issue in-order execution pipelines (hardware threads) which share a common memory sub-system (consisting of instruction and data caches together with a memory management unit). From a design perspective, the assembly and verification of the dual threaded processor is simplified by the use of existing verified implementations of the execution pipeline and a memory unit. Because the memory unit is shared by the two hardware threads, the relative area overhead of adding the second hardware thread is 25\% of the area of the existing single threaded processor. Using an FPGA implementation we evaluate the performance of the dual threaded processor relative to the single threaded one. On applications which can be parallelized, we observe speedups of 1.6X to 1.88X. For applications that are not parallelizable, the speedup is more modest. We also observe that the dual threaded processor performance is degraded on applications which generate large numbers of cache misses.
2209.00841
Zeyong Wei
Honghua Chen, Mingqiang Wei, Jun Wang
Geometric and Learning-based Mesh Denoising: A Comprehensive Survey
null
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Mesh denoising is a fundamental problem in digital geometry processing. It seeks to remove surface noise, while preserving surface intrinsic signals as accurately as possible. While the traditional wisdom has been built upon specialized priors to smooth surfaces, learning-based approaches are making their debut with great success in generalization and automation. In this work, we provide a comprehensive review of the advances in mesh denoising, containing both traditional geometric approaches and recent learning-based methods. First, to familiarize readers with the denoising tasks, we summarize four common issues in mesh denoising. We then provide two categorizations of the existing denoising methods. Furthermore, three important categories, including optimization-, filter-, and data-driven-based techniques, are introduced and analyzed in detail, respectively. Both qualitative and quantitative comparisons are illustrated, to demonstrate the effectiveness of the state-of-the-art denoising methods. Finally, potential directions of future work are pointed out to solve the common problems of these approaches. A mesh denoising benchmark is also built in this work, and future researchers will easily and conveniently evaluate their methods with the state-of-the-art approaches.
[ { "created": "Fri, 2 Sep 2022 06:54:32 GMT", "version": "v1" } ]
2022-09-05
[ [ "Chen", "Honghua", "" ], [ "Wei", "Mingqiang", "" ], [ "Wang", "Jun", "" ] ]
Mesh denoising is a fundamental problem in digital geometry processing. It seeks to remove surface noise, while preserving surface intrinsic signals as accurately as possible. While the traditional wisdom has been built upon specialized priors to smooth surfaces, learning-based approaches are making their debut with great success in generalization and automation. In this work, we provide a comprehensive review of the advances in mesh denoising, containing both traditional geometric approaches and recent learning-based methods. First, to familiarize readers with the denoising tasks, we summarize four common issues in mesh denoising. We then provide two categorizations of the existing denoising methods. Furthermore, three important categories, including optimization-, filter-, and data-driven-based techniques, are introduced and analyzed in detail, respectively. Both qualitative and quantitative comparisons are illustrated, to demonstrate the effectiveness of the state-of-the-art denoising methods. Finally, potential directions of future work are pointed out to solve the common problems of these approaches. A mesh denoising benchmark is also built in this work, and future researchers will easily and conveniently evaluate their methods with the state-of-the-art approaches.
2011.07932
Siyeong Lee
Kwanghee Choi and Siyeong Lee
Combating the Instability of Mutual Information-based Losses via Regularization
Kwanghee Choi and Siyeong Lee contributed equally to this paper. Accepted for the 38th Conference on Uncertainty in Artificial Intelligence (UAI 2022)
null
null
null
cs.LG cs.IT math.IT stat.ML
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Notable progress has been made in numerous fields of machine learning based on neural network-driven mutual information (MI) bounds. However, utilizing the conventional MI-based losses is often challenging due to their practical and mathematical limitations. In this work, we first identify the symptoms behind their instability: (1) the neural network not converging even after the loss seemed to converge, and (2) saturating neural network outputs causing the loss to diverge. We mitigate both issues by adding a novel regularization term to the existing losses. We theoretically and experimentally demonstrate that added regularization stabilizes training. Finally, we present a novel benchmark that evaluates MI-based losses on both the MI estimation power and its capability on the downstream tasks, closely following the pre-existing supervised and contrastive learning settings. We evaluate six different MI-based losses and their regularized counterparts on multiple benchmarks to show that our approach is simple yet effective.
[ { "created": "Mon, 16 Nov 2020 13:29:15 GMT", "version": "v1" }, { "created": "Mon, 28 Feb 2022 14:19:26 GMT", "version": "v2" }, { "created": "Fri, 4 Mar 2022 01:10:24 GMT", "version": "v3" }, { "created": "Sat, 18 Jun 2022 04:01:51 GMT", "version": "v4" } ]
2022-06-22
[ [ "Choi", "Kwanghee", "" ], [ "Lee", "Siyeong", "" ] ]
Notable progress has been made in numerous fields of machine learning based on neural network-driven mutual information (MI) bounds. However, utilizing the conventional MI-based losses is often challenging due to their practical and mathematical limitations. In this work, we first identify the symptoms behind their instability: (1) the neural network not converging even after the loss seemed to converge, and (2) saturating neural network outputs causing the loss to diverge. We mitigate both issues by adding a novel regularization term to the existing losses. We theoretically and experimentally demonstrate that added regularization stabilizes training. Finally, we present a novel benchmark that evaluates MI-based losses on both the MI estimation power and its capability on the downstream tasks, closely following the pre-existing supervised and contrastive learning settings. We evaluate six different MI-based losses and their regularized counterparts on multiple benchmarks to show that our approach is simple yet effective.
1706.06457
Mohsen Imani
Mohsen Imani, Mehrdad Amirabadi, and Matthew Wright
The Evaluation of Circuit Selection Methods on Tor
arXiv admin note: substantial text overlap with arXiv:1608.07343
null
null
null
cs.CR
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Tor provides anonymity online by routing traffic through encrypted tunnels, called circuits, over paths of anonymizing relays. To enable users to connect to their selected destination servers without waiting for the circuit to be build, the Tor client maintains a few circuits at all times. Nevertheless, Tor is slower to use than directly connecting to the destination server. In this paper, we propose to have the Tor client measure the performance of the pre-built circuits and select the fastest circuits for users to send their traffic over. To this end, we define and evaluate nine metrics for selecting which pre-built circuit to use based on different combinations of circuit length, Round Trip Time (RTT), and congestion. We also explore the effect on performance of the number of pre-built circuits at the time of the selection. Through whole-network experiments in Shadow, we show that using circuit RTT with at least three pre-built circuits allows the Tor client to identify fast circuits and improves median time to first byte (TTFB) by 22% over Tor and 15% over congestion-aware routing, the state-of-the-art in Tor circuit selection. We evaluate the security of the proposed circuit selection mechanism against both a relay-level and a network-level adversary and find no loss of security compared with Tor.
[ { "created": "Sat, 17 Jun 2017 23:39:19 GMT", "version": "v1" } ]
2017-06-21
[ [ "Imani", "Mohsen", "" ], [ "Amirabadi", "Mehrdad", "" ], [ "Wright", "Matthew", "" ] ]
Tor provides anonymity online by routing traffic through encrypted tunnels, called circuits, over paths of anonymizing relays. To enable users to connect to their selected destination servers without waiting for the circuit to be build, the Tor client maintains a few circuits at all times. Nevertheless, Tor is slower to use than directly connecting to the destination server. In this paper, we propose to have the Tor client measure the performance of the pre-built circuits and select the fastest circuits for users to send their traffic over. To this end, we define and evaluate nine metrics for selecting which pre-built circuit to use based on different combinations of circuit length, Round Trip Time (RTT), and congestion. We also explore the effect on performance of the number of pre-built circuits at the time of the selection. Through whole-network experiments in Shadow, we show that using circuit RTT with at least three pre-built circuits allows the Tor client to identify fast circuits and improves median time to first byte (TTFB) by 22% over Tor and 15% over congestion-aware routing, the state-of-the-art in Tor circuit selection. We evaluate the security of the proposed circuit selection mechanism against both a relay-level and a network-level adversary and find no loss of security compared with Tor.
1103.5320
Francesco De Pellegrini Dr.
Alberto Montresor, Francesco De Pellegrini and Daniele Miorandi
Distributed k-Core Decomposition
null
null
null
null
cs.OH
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Among the novel metrics used to study the relative importance of nodes in complex networks, k-core decomposition has found a number of applications in areas as diverse as sociology, proteinomics, graph visualization, and distributed system analysis and design. This paper proposes new distributed algorithms for the computation of the k-core decomposition of a network, with the purpose of (i) enabling the run-time computation of k-cores in "live" distributed systems and (ii) allowing the decomposition, over a set of connected machines, of very large graphs, that cannot be hosted in a single machine. Lower bounds on the algorithms complexity are given, and an exhaustive experimental analysis on real-world graphs is provided.
[ { "created": "Mon, 28 Mar 2011 10:27:48 GMT", "version": "v1" }, { "created": "Tue, 29 Mar 2011 10:12:14 GMT", "version": "v2" } ]
2011-03-30
[ [ "Montresor", "Alberto", "" ], [ "De Pellegrini", "Francesco", "" ], [ "Miorandi", "Daniele", "" ] ]
Among the novel metrics used to study the relative importance of nodes in complex networks, k-core decomposition has found a number of applications in areas as diverse as sociology, proteinomics, graph visualization, and distributed system analysis and design. This paper proposes new distributed algorithms for the computation of the k-core decomposition of a network, with the purpose of (i) enabling the run-time computation of k-cores in "live" distributed systems and (ii) allowing the decomposition, over a set of connected machines, of very large graphs, that cannot be hosted in a single machine. Lower bounds on the algorithms complexity are given, and an exhaustive experimental analysis on real-world graphs is provided.
2305.07244
Prasad Talasila
Prasad Talasila, Cl\'audio Gomes, Peter H{\o}gh Mikkelsen, Santiago Gil Arboleda, Eduard Kamburjan, Peter Gorm Larsen
Digital Twin as a Service (DTaaS): A Platform for Digital Twin Developers and Users
8 pages, 6 figures. Accepted at Digital Twin 2023
null
10.1109/SWC57546.2023.10448890
null
cs.SE
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Establishing digital twins is a non-trivial endeavour especially when users face significant challenges in creating them from scratch. Ready availability of reusable models, data and tool assets, can help with creation and use of digital twins. A number of digital twin frameworks exist to facilitate creation and use of digital twins. In this paper we propose a digital twin framework to author digital twin assets, create digital twins from reusable assets and make the digital twins available as a service to other users. The proposed framework automates the management of reusable assets, storage, provision of compute infrastructure, communication and monitoring tasks. The users operate at the level of digital twins and delegate rest of the work to the digital twin as a service framework.
[ { "created": "Fri, 12 May 2023 04:34:30 GMT", "version": "v1" }, { "created": "Tue, 13 Jun 2023 08:59:12 GMT", "version": "v2" } ]
2024-03-11
[ [ "Talasila", "Prasad", "" ], [ "Gomes", "Cláudio", "" ], [ "Mikkelsen", "Peter Høgh", "" ], [ "Arboleda", "Santiago Gil", "" ], [ "Kamburjan", "Eduard", "" ], [ "Larsen", "Peter Gorm", "" ] ]
Establishing digital twins is a non-trivial endeavour especially when users face significant challenges in creating them from scratch. Ready availability of reusable models, data and tool assets, can help with creation and use of digital twins. A number of digital twin frameworks exist to facilitate creation and use of digital twins. In this paper we propose a digital twin framework to author digital twin assets, create digital twins from reusable assets and make the digital twins available as a service to other users. The proposed framework automates the management of reusable assets, storage, provision of compute infrastructure, communication and monitoring tasks. The users operate at the level of digital twins and delegate rest of the work to the digital twin as a service framework.
2303.00109
Felix Schr\"oder
Stefan Felsner and Hendrik Schrezenmaier and Felix Schr\"oder and Raphael Steiner
Linear Size Universal Point Sets for Classes of Planar Graphs
null
null
null
null
cs.CG cs.DM math.CO
http://creativecommons.org/licenses/by-sa/4.0/
A finite set $P$ of points in the plane is $n$-universal with respect to a class $\mathcal{C}$ of planar graphs if every $n$-vertex graph in $\mathcal{C}$ admits a crossing-free straight-line drawing with vertices at points of $P$. For the class of all planar graphs the best known upper bound on the size of a universal point set is quadratic and the best known lower bound is linear in $n$. Some classes of planar graphs are known to admit universal point sets of near linear size, however, there are no truly linear bounds for interesting classes beyond outerplanar graphs. In this paper, we show that there is a universal point set of size $2n-2$ for the class of bipartite planar graphs with $n$ vertices. The same point set is also universal for the class of $n$-vertex planar graphs of maximum degree $3$. The point set used for the results is what we call an exploding double chain, and we prove that this point set allows planar straight-line embeddings of many more planar graphs, namely of all subgraphs of planar graphs admitting a one-sided Hamiltonian cycle. The result for bipartite graphs also implies that every $n$-vertex plane graph has a $1$-bend drawing all whose bends and vertices are contained in a specific point set of size $4n-6$, this improves a bound of $6n-10$ for the same problem by L\"offler and T\'oth.
[ { "created": "Tue, 28 Feb 2023 22:15:38 GMT", "version": "v1" } ]
2023-03-02
[ [ "Felsner", "Stefan", "" ], [ "Schrezenmaier", "Hendrik", "" ], [ "Schröder", "Felix", "" ], [ "Steiner", "Raphael", "" ] ]
A finite set $P$ of points in the plane is $n$-universal with respect to a class $\mathcal{C}$ of planar graphs if every $n$-vertex graph in $\mathcal{C}$ admits a crossing-free straight-line drawing with vertices at points of $P$. For the class of all planar graphs the best known upper bound on the size of a universal point set is quadratic and the best known lower bound is linear in $n$. Some classes of planar graphs are known to admit universal point sets of near linear size, however, there are no truly linear bounds for interesting classes beyond outerplanar graphs. In this paper, we show that there is a universal point set of size $2n-2$ for the class of bipartite planar graphs with $n$ vertices. The same point set is also universal for the class of $n$-vertex planar graphs of maximum degree $3$. The point set used for the results is what we call an exploding double chain, and we prove that this point set allows planar straight-line embeddings of many more planar graphs, namely of all subgraphs of planar graphs admitting a one-sided Hamiltonian cycle. The result for bipartite graphs also implies that every $n$-vertex plane graph has a $1$-bend drawing all whose bends and vertices are contained in a specific point set of size $4n-6$, this improves a bound of $6n-10$ for the same problem by L\"offler and T\'oth.
1608.03580
Erik Waingarten
Alexandr Andoni and Thijs Laarhoven and Ilya Razenshteyn and Erik Waingarten
Optimal Hashing-based Time-Space Trade-offs for Approximate Near Neighbors
62 pages, 5 figures; a merger of arXiv:1511.07527 [cs.DS] and arXiv:1605.02701 [cs.DS], which subsumes both of the preprints. New version contains more elaborated proofs and fixed some typos
null
10.1137/1.9781611974782.4
null
cs.DS cs.CC cs.CG cs.IR
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
[See the paper for the full abstract.] We show tight upper and lower bounds for time-space trade-offs for the $c$-Approximate Near Neighbor Search problem. For the $d$-dimensional Euclidean space and $n$-point datasets, we develop a data structure with space $n^{1 + \rho_u + o(1)} + O(dn)$ and query time $n^{\rho_q + o(1)} + d n^{o(1)}$ for every $\rho_u, \rho_q \geq 0$ such that: \begin{equation} c^2 \sqrt{\rho_q} + (c^2 - 1) \sqrt{\rho_u} = \sqrt{2c^2 - 1}. \end{equation} This is the first data structure that achieves sublinear query time and near-linear space for every approximation factor $c > 1$, improving upon [Kapralov, PODS 2015]. The data structure is a culmination of a long line of work on the problem for all space regimes; it builds on Spherical Locality-Sensitive Filtering [Becker, Ducas, Gama, Laarhoven, SODA 2016] and data-dependent hashing [Andoni, Indyk, Nguyen, Razenshteyn, SODA 2014] [Andoni, Razenshteyn, STOC 2015]. Our matching lower bounds are of two types: conditional and unconditional. First, we prove tightness of the whole above trade-off in a restricted model of computation, which captures all known hashing-based approaches. We then show unconditional cell-probe lower bounds for one and two probes that match the above trade-off for $\rho_q = 0$, improving upon the best known lower bounds from [Panigrahy, Talwar, Wieder, FOCS 2010]. In particular, this is the first space lower bound (for any static data structure) for two probes which is not polynomially smaller than the one-probe bound. To show the result for two probes, we establish and exploit a connection to locally-decodable codes.
[ { "created": "Thu, 11 Aug 2016 19:50:00 GMT", "version": "v1" }, { "created": "Sun, 21 May 2017 16:57:47 GMT", "version": "v2" } ]
2019-10-04
[ [ "Andoni", "Alexandr", "" ], [ "Laarhoven", "Thijs", "" ], [ "Razenshteyn", "Ilya", "" ], [ "Waingarten", "Erik", "" ] ]
[See the paper for the full abstract.] We show tight upper and lower bounds for time-space trade-offs for the $c$-Approximate Near Neighbor Search problem. For the $d$-dimensional Euclidean space and $n$-point datasets, we develop a data structure with space $n^{1 + \rho_u + o(1)} + O(dn)$ and query time $n^{\rho_q + o(1)} + d n^{o(1)}$ for every $\rho_u, \rho_q \geq 0$ such that: \begin{equation} c^2 \sqrt{\rho_q} + (c^2 - 1) \sqrt{\rho_u} = \sqrt{2c^2 - 1}. \end{equation} This is the first data structure that achieves sublinear query time and near-linear space for every approximation factor $c > 1$, improving upon [Kapralov, PODS 2015]. The data structure is a culmination of a long line of work on the problem for all space regimes; it builds on Spherical Locality-Sensitive Filtering [Becker, Ducas, Gama, Laarhoven, SODA 2016] and data-dependent hashing [Andoni, Indyk, Nguyen, Razenshteyn, SODA 2014] [Andoni, Razenshteyn, STOC 2015]. Our matching lower bounds are of two types: conditional and unconditional. First, we prove tightness of the whole above trade-off in a restricted model of computation, which captures all known hashing-based approaches. We then show unconditional cell-probe lower bounds for one and two probes that match the above trade-off for $\rho_q = 0$, improving upon the best known lower bounds from [Panigrahy, Talwar, Wieder, FOCS 2010]. In particular, this is the first space lower bound (for any static data structure) for two probes which is not polynomially smaller than the one-probe bound. To show the result for two probes, we establish and exploit a connection to locally-decodable codes.
2103.15339
Haw-Shiuan Chang
Rohan Paul, Haw-Shiuan Chang, Andrew McCallum
Multi-facet Universal Schema
EACL 2021
null
null
null
cs.CL
http://creativecommons.org/licenses/by/4.0/
Universal schema (USchema) assumes that two sentence patterns that share the same entity pairs are similar to each other. This assumption is widely adopted for solving various types of relation extraction (RE) tasks. Nevertheless, each sentence pattern could contain multiple facets, and not every facet is similar to all the facets of another sentence pattern co-occurring with the same entity pair. To address the violation of the USchema assumption, we propose multi-facet universal schema that uses a neural model to represent each sentence pattern as multiple facet embeddings and encourage one of these facet embeddings to be close to that of another sentence pattern if they co-occur with the same entity pair. In our experiments, we demonstrate that multi-facet embeddings significantly outperform their single-facet embedding counterpart, compositional universal schema (CUSchema) (Verga et al., 2016), in distantly supervised relation extraction tasks. Moreover, we can also use multiple embeddings to detect the entailment relation between two sentence patterns when no manual label is available.
[ { "created": "Mon, 29 Mar 2021 05:10:10 GMT", "version": "v1" } ]
2021-03-30
[ [ "Paul", "Rohan", "" ], [ "Chang", "Haw-Shiuan", "" ], [ "McCallum", "Andrew", "" ] ]
Universal schema (USchema) assumes that two sentence patterns that share the same entity pairs are similar to each other. This assumption is widely adopted for solving various types of relation extraction (RE) tasks. Nevertheless, each sentence pattern could contain multiple facets, and not every facet is similar to all the facets of another sentence pattern co-occurring with the same entity pair. To address the violation of the USchema assumption, we propose multi-facet universal schema that uses a neural model to represent each sentence pattern as multiple facet embeddings and encourage one of these facet embeddings to be close to that of another sentence pattern if they co-occur with the same entity pair. In our experiments, we demonstrate that multi-facet embeddings significantly outperform their single-facet embedding counterpart, compositional universal schema (CUSchema) (Verga et al., 2016), in distantly supervised relation extraction tasks. Moreover, we can also use multiple embeddings to detect the entailment relation between two sentence patterns when no manual label is available.
2404.05044
Morteza Maleki
Morteza Maleki, SeyedAli Ghahari
Clinical Trials Protocol Authoring using LLMs
29 pages, under review by IEEE Journal
null
null
null
cs.CE
http://creativecommons.org/licenses/by/4.0/
This report embarks on a mission to revolutionize clinical trial protocol development through the integration of advanced AI technologies. With a focus on leveraging the capabilities of generative AI, specifically GPT-4, this initiative aimed to streamline and enhance the efficiency and accuracy of clinical trial protocols. The methodology encompassed a detailed analysis and preparation of comprehensive drug and study level metadata, followed by the deployment of GPT-4 for automated protocol section generation. Results demonstrated a significant improvement in protocol authoring, highlighted by increases in efficiency, accuracy, and the customization of protocols to specific trial requirements. Challenges encountered during model selection and prompt engineering were systematically addressed, leading to refined methodologies that capitalized on the advanced text generation capabilities of GPT-4. This project not only showcases the practical applications and benefits of generative AI in clinical trial design but also sets a foundation for future innovations in the field.
[ { "created": "Sun, 7 Apr 2024 18:59:03 GMT", "version": "v1" }, { "created": "Sun, 4 Aug 2024 20:31:35 GMT", "version": "v2" } ]
2024-08-06
[ [ "Maleki", "Morteza", "" ], [ "Ghahari", "SeyedAli", "" ] ]
This report embarks on a mission to revolutionize clinical trial protocol development through the integration of advanced AI technologies. With a focus on leveraging the capabilities of generative AI, specifically GPT-4, this initiative aimed to streamline and enhance the efficiency and accuracy of clinical trial protocols. The methodology encompassed a detailed analysis and preparation of comprehensive drug and study level metadata, followed by the deployment of GPT-4 for automated protocol section generation. Results demonstrated a significant improvement in protocol authoring, highlighted by increases in efficiency, accuracy, and the customization of protocols to specific trial requirements. Challenges encountered during model selection and prompt engineering were systematically addressed, leading to refined methodologies that capitalized on the advanced text generation capabilities of GPT-4. This project not only showcases the practical applications and benefits of generative AI in clinical trial design but also sets a foundation for future innovations in the field.
2204.11842
Michael Beukman
Michael Beukman and Michael Mitchley and Dean Wookey and Steven James and George Konidaris
Adaptive Online Value Function Approximation with Wavelets
Accepted to RLDM 2022. Code is located at https://github.com/Michael-Beukman/WaveletRL
null
null
null
cs.LG cs.AI
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Using function approximation to represent a value function is necessary for continuous and high-dimensional state spaces. Linear function approximation has desirable theoretical guarantees and often requires less compute and samples than neural networks, but most approaches suffer from an exponential growth in the number of functions as the dimensionality of the state space increases. In this work, we introduce the wavelet basis for reinforcement learning. Wavelets can effectively be used as a fixed basis and additionally provide the ability to adaptively refine the basis set as learning progresses, making it feasible to start with a minimal basis set. This adaptive method can either increase the granularity of the approximation at a point in state space, or add in interactions between different dimensions as necessary. We prove that wavelets are both necessary and sufficient if we wish to construct a function approximator that can be adaptively refined without loss of precision. We further demonstrate that a fixed wavelet basis set performs comparably against the high-performing Fourier basis on Mountain Car and Acrobot, and that the adaptive methods provide a convenient approach to addressing an oversized initial basis set, while demonstrating performance comparable to, or greater than, the fixed wavelet basis.
[ { "created": "Fri, 22 Apr 2022 11:35:57 GMT", "version": "v1" } ]
2022-04-27
[ [ "Beukman", "Michael", "" ], [ "Mitchley", "Michael", "" ], [ "Wookey", "Dean", "" ], [ "James", "Steven", "" ], [ "Konidaris", "George", "" ] ]
Using function approximation to represent a value function is necessary for continuous and high-dimensional state spaces. Linear function approximation has desirable theoretical guarantees and often requires less compute and samples than neural networks, but most approaches suffer from an exponential growth in the number of functions as the dimensionality of the state space increases. In this work, we introduce the wavelet basis for reinforcement learning. Wavelets can effectively be used as a fixed basis and additionally provide the ability to adaptively refine the basis set as learning progresses, making it feasible to start with a minimal basis set. This adaptive method can either increase the granularity of the approximation at a point in state space, or add in interactions between different dimensions as necessary. We prove that wavelets are both necessary and sufficient if we wish to construct a function approximator that can be adaptively refined without loss of precision. We further demonstrate that a fixed wavelet basis set performs comparably against the high-performing Fourier basis on Mountain Car and Acrobot, and that the adaptive methods provide a convenient approach to addressing an oversized initial basis set, while demonstrating performance comparable to, or greater than, the fixed wavelet basis.
1904.06268
Zuxuan Wu
Zuxuan Wu, Xin Wang, Joseph E. Gonzalez, Tom Goldstein, Larry S. Davis
ACE: Adapting to Changing Environments for Semantic Segmentation
null
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Deep neural networks exhibit exceptional accuracy when they are trained and tested on the same data distributions. However, neural classifiers are often extremely brittle when confronted with domain shift---changes in the input distribution that occur over time. We present ACE, a framework for semantic segmentation that dynamically adapts to changing environments over the time. By aligning the distribution of labeled training data from the original source domain with the distribution of incoming data in a shifted domain, ACE synthesizes labeled training data for environments as it sees them. This stylized data is then used to update a segmentation model so that it performs well in new environments. To avoid forgetting knowledge from past environments, we introduce a memory that stores feature statistics from previously seen domains. These statistics can be used to replay images in any of the previously observed domains, thus preventing catastrophic forgetting. In addition to standard batch training using stochastic gradient decent (SGD), we also experiment with fast adaptation methods based on adaptive meta-learning. Extensive experiments are conducted on two datasets from SYNTHIA, the results demonstrate the effectiveness of the proposed approach when adapting to a number of tasks.
[ { "created": "Fri, 12 Apr 2019 15:15:15 GMT", "version": "v1" } ]
2019-04-15
[ [ "Wu", "Zuxuan", "" ], [ "Wang", "Xin", "" ], [ "Gonzalez", "Joseph E.", "" ], [ "Goldstein", "Tom", "" ], [ "Davis", "Larry S.", "" ] ]
Deep neural networks exhibit exceptional accuracy when they are trained and tested on the same data distributions. However, neural classifiers are often extremely brittle when confronted with domain shift---changes in the input distribution that occur over time. We present ACE, a framework for semantic segmentation that dynamically adapts to changing environments over the time. By aligning the distribution of labeled training data from the original source domain with the distribution of incoming data in a shifted domain, ACE synthesizes labeled training data for environments as it sees them. This stylized data is then used to update a segmentation model so that it performs well in new environments. To avoid forgetting knowledge from past environments, we introduce a memory that stores feature statistics from previously seen domains. These statistics can be used to replay images in any of the previously observed domains, thus preventing catastrophic forgetting. In addition to standard batch training using stochastic gradient decent (SGD), we also experiment with fast adaptation methods based on adaptive meta-learning. Extensive experiments are conducted on two datasets from SYNTHIA, the results demonstrate the effectiveness of the proposed approach when adapting to a number of tasks.
2111.04798
Cristina Menghini
Wasu Piriyakulkij and Cristina Menghini and Ross Briden and Nihal V. Nayak and Jeffrey Zhu and Elaheh Raisi and Stephen H. Bach
TAGLETS: A System for Automatic Semi-Supervised Learning with Auxiliary Data
Paper published at MLSys 2022. It passed the artifact evaluation earning two ACM badges: (1) Artifacts Evaluated Functional v1.1 and (2) Artifacts Available v1.1
null
null
null
cs.LG cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Machine learning practitioners often have access to a spectrum of data: labeled data for the target task (which is often limited), unlabeled data, and auxiliary data, the many available labeled datasets for other tasks. We describe TAGLETS, a system built to study techniques for automatically exploiting all three types of data and creating high-quality, servable classifiers. The key components of TAGLETS are: (1) auxiliary data organized according to a knowledge graph, (2) modules encapsulating different methods for exploiting auxiliary and unlabeled data, and (3) a distillation stage in which the ensembled modules are combined into a servable model. We compare TAGLETS with state-of-the-art transfer learning and semi-supervised learning methods on four image classification tasks. Our study covers a range of settings, varying the amount of labeled data and the semantic relatedness of the auxiliary data to the target task. We find that the intelligent incorporation of auxiliary and unlabeled data into multiple learning techniques enables TAGLETS to match-and most often significantly surpass-these alternatives. TAGLETS is available as an open-source system at github.com/BatsResearch/taglets.
[ { "created": "Mon, 8 Nov 2021 20:08:45 GMT", "version": "v1" }, { "created": "Wed, 10 Nov 2021 15:33:24 GMT", "version": "v2" }, { "created": "Thu, 5 May 2022 23:49:23 GMT", "version": "v3" } ]
2022-05-09
[ [ "Piriyakulkij", "Wasu", "" ], [ "Menghini", "Cristina", "" ], [ "Briden", "Ross", "" ], [ "Nayak", "Nihal V.", "" ], [ "Zhu", "Jeffrey", "" ], [ "Raisi", "Elaheh", "" ], [ "Bach", "Stephen H.", "" ] ]
Machine learning practitioners often have access to a spectrum of data: labeled data for the target task (which is often limited), unlabeled data, and auxiliary data, the many available labeled datasets for other tasks. We describe TAGLETS, a system built to study techniques for automatically exploiting all three types of data and creating high-quality, servable classifiers. The key components of TAGLETS are: (1) auxiliary data organized according to a knowledge graph, (2) modules encapsulating different methods for exploiting auxiliary and unlabeled data, and (3) a distillation stage in which the ensembled modules are combined into a servable model. We compare TAGLETS with state-of-the-art transfer learning and semi-supervised learning methods on four image classification tasks. Our study covers a range of settings, varying the amount of labeled data and the semantic relatedness of the auxiliary data to the target task. We find that the intelligent incorporation of auxiliary and unlabeled data into multiple learning techniques enables TAGLETS to match-and most often significantly surpass-these alternatives. TAGLETS is available as an open-source system at github.com/BatsResearch/taglets.
2105.04632
Hunter Priniski
J. Hunter Priniski, Mason McClay, Keith J. Holyoak
Rise of QAnon: A Mental Model of Good and Evil Stews in an Echochamber
2 figures, 7 pages
null
null
null
cs.SI cs.CY
http://creativecommons.org/licenses/by/4.0/
The QAnon conspiracy posits that Satan-worshiping Democrats operate a covert child sex-trafficking operation, which Donald Trump is destined to expose and annihilate. Emblematic of the ease with which political misconceptions can spread through social media, QAnon originated in late 2017 and rapidly grew to shape the political beliefs of millions. To illuminate the process by which a conspiracy theory spreads, we report two computational studies examining the social network structure and semantic content of tweets produced by users central to the early QAnon network on Twitter. Using data mined in the summer of 2018, we examined over 800,000 tweets about QAnon made by about 100,000 users. The majority of users disseminated rather than produced information, serving to create an online echochamber. Users appeared to hold a simplistic mental model in which political events are viewed as a struggle between antithetical forces-both observed and unobserved-of Good and Evil.
[ { "created": "Mon, 10 May 2021 19:34:35 GMT", "version": "v1" } ]
2021-05-12
[ [ "Priniski", "J. Hunter", "" ], [ "McClay", "Mason", "" ], [ "Holyoak", "Keith J.", "" ] ]
The QAnon conspiracy posits that Satan-worshiping Democrats operate a covert child sex-trafficking operation, which Donald Trump is destined to expose and annihilate. Emblematic of the ease with which political misconceptions can spread through social media, QAnon originated in late 2017 and rapidly grew to shape the political beliefs of millions. To illuminate the process by which a conspiracy theory spreads, we report two computational studies examining the social network structure and semantic content of tweets produced by users central to the early QAnon network on Twitter. Using data mined in the summer of 2018, we examined over 800,000 tweets about QAnon made by about 100,000 users. The majority of users disseminated rather than produced information, serving to create an online echochamber. Users appeared to hold a simplistic mental model in which political events are viewed as a struggle between antithetical forces-both observed and unobserved-of Good and Evil.
2404.13370
Zihao Yue
Zihao Yue, Yepeng Zhang, Ziheng Wang, Qin Jin
Movie101v2: Improved Movie Narration Benchmark
null
null
null
null
cs.CV cs.CL cs.MM
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Automatic movie narration targets at creating video-aligned plot descriptions to assist visually impaired audiences. It differs from standard video captioning in that it requires not only describing key visual details but also inferring the plots developed across multiple movie shots, thus posing unique and ongoing challenges. To advance the development of automatic movie narrating systems, we first revisit the limitations of existing datasets and develop a large-scale, bilingual movie narration dataset, Movie101v2. Second, taking into account the essential difficulties in achieving applicable movie narration, we break the long-term goal into three progressive stages and tentatively focus on the initial stages featuring understanding within individual clips. We also introduce a new narration assessment to align with our staged task goals. Third, using our new dataset, we baseline several leading large vision-language models, including GPT-4V, and conduct in-depth investigations into the challenges current models face for movie narration generation. Our findings reveal that achieving applicable movie narration generation is a fascinating goal that requires thorough research.
[ { "created": "Sat, 20 Apr 2024 13:15:27 GMT", "version": "v1" } ]
2024-04-23
[ [ "Yue", "Zihao", "" ], [ "Zhang", "Yepeng", "" ], [ "Wang", "Ziheng", "" ], [ "Jin", "Qin", "" ] ]
Automatic movie narration targets at creating video-aligned plot descriptions to assist visually impaired audiences. It differs from standard video captioning in that it requires not only describing key visual details but also inferring the plots developed across multiple movie shots, thus posing unique and ongoing challenges. To advance the development of automatic movie narrating systems, we first revisit the limitations of existing datasets and develop a large-scale, bilingual movie narration dataset, Movie101v2. Second, taking into account the essential difficulties in achieving applicable movie narration, we break the long-term goal into three progressive stages and tentatively focus on the initial stages featuring understanding within individual clips. We also introduce a new narration assessment to align with our staged task goals. Third, using our new dataset, we baseline several leading large vision-language models, including GPT-4V, and conduct in-depth investigations into the challenges current models face for movie narration generation. Our findings reveal that achieving applicable movie narration generation is a fascinating goal that requires thorough research.
2404.13749
Xinyu Huang
Xinyu Huang and Shisheng Hu and Mushu Li and Cheng Huang and Xuemin Shen
Efficient Digital Twin Data Processing for Low-Latency Multicast Short Video Streaming
6 pages, 6 figures, submitted to ICCC 2024
null
null
null
cs.NI
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
In this paper, we propose a novel efficient digital twin (DT) data processing scheme to reduce service latency for multicast short video streaming. Particularly, DT is constructed to emulate and analyze user status for multicast group update and swipe feature abstraction. Then, a precise measurement model of DT data processing is developed to characterize the relationship among DT model size, user dynamics, and user clustering accuracy. A service latency model, consisting of DT data processing delay, video transcoding delay, and multicast transmission delay, is constructed by incorporating the impact of user clustering accuracy. Finally, a joint optimization problem of DT model size selection and bandwidth allocation is formulated to minimize the service latency. To efficiently solve this problem, a diffusion-based resource management algorithm is proposed, which utilizes the denoising technique to improve the action-generation process in the deep reinforcement learning algorithm. Simulation results based on the real-world dataset demonstrate that the proposed DT data processing scheme outperforms benchmark schemes in terms of service latency.
[ { "created": "Sun, 21 Apr 2024 19:12:22 GMT", "version": "v1" } ]
2024-04-23
[ [ "Huang", "Xinyu", "" ], [ "Hu", "Shisheng", "" ], [ "Li", "Mushu", "" ], [ "Huang", "Cheng", "" ], [ "Shen", "Xuemin", "" ] ]
In this paper, we propose a novel efficient digital twin (DT) data processing scheme to reduce service latency for multicast short video streaming. Particularly, DT is constructed to emulate and analyze user status for multicast group update and swipe feature abstraction. Then, a precise measurement model of DT data processing is developed to characterize the relationship among DT model size, user dynamics, and user clustering accuracy. A service latency model, consisting of DT data processing delay, video transcoding delay, and multicast transmission delay, is constructed by incorporating the impact of user clustering accuracy. Finally, a joint optimization problem of DT model size selection and bandwidth allocation is formulated to minimize the service latency. To efficiently solve this problem, a diffusion-based resource management algorithm is proposed, which utilizes the denoising technique to improve the action-generation process in the deep reinforcement learning algorithm. Simulation results based on the real-world dataset demonstrate that the proposed DT data processing scheme outperforms benchmark schemes in terms of service latency.
2309.11248
Xuyang Chen
Xuyang Chen, Dong Wang, Konrad Schindler, Mingwei Sun, Yongliang Wang, Nicolo Savioli, Liqiu Meng
Box2Poly: Memory-Efficient Polygon Prediction of Arbitrarily Shaped and Rotated Text
null
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Recently, Transformer-based text detection techniques have sought to predict polygons by encoding the coordinates of individual boundary vertices using distinct query features. However, this approach incurs a significant memory overhead and struggles to effectively capture the intricate relationships between vertices belonging to the same instance. Consequently, irregular text layouts often lead to the prediction of outlined vertices, diminishing the quality of results. To address these challenges, we present an innovative approach rooted in Sparse R-CNN: a cascade decoding pipeline for polygon prediction. Our method ensures precision by iteratively refining polygon predictions, considering both the scale and location of preceding results. Leveraging this stabilized regression pipeline, even employing just a single feature vector to guide polygon instance regression yields promising detection results. Simultaneously, the leverage of instance-level feature proposal substantially enhances memory efficiency (>50% less vs. the state-of-the-art method DPText-DETR) and reduces inference speed (>40% less vs. DPText-DETR) with minor performance drop on benchmarks.
[ { "created": "Wed, 20 Sep 2023 12:19:07 GMT", "version": "v1" } ]
2023-09-21
[ [ "Chen", "Xuyang", "" ], [ "Wang", "Dong", "" ], [ "Schindler", "Konrad", "" ], [ "Sun", "Mingwei", "" ], [ "Wang", "Yongliang", "" ], [ "Savioli", "Nicolo", "" ], [ "Meng", "Liqiu", "" ] ]
Recently, Transformer-based text detection techniques have sought to predict polygons by encoding the coordinates of individual boundary vertices using distinct query features. However, this approach incurs a significant memory overhead and struggles to effectively capture the intricate relationships between vertices belonging to the same instance. Consequently, irregular text layouts often lead to the prediction of outlined vertices, diminishing the quality of results. To address these challenges, we present an innovative approach rooted in Sparse R-CNN: a cascade decoding pipeline for polygon prediction. Our method ensures precision by iteratively refining polygon predictions, considering both the scale and location of preceding results. Leveraging this stabilized regression pipeline, even employing just a single feature vector to guide polygon instance regression yields promising detection results. Simultaneously, the leverage of instance-level feature proposal substantially enhances memory efficiency (>50% less vs. the state-of-the-art method DPText-DETR) and reduces inference speed (>40% less vs. DPText-DETR) with minor performance drop on benchmarks.
2303.15754
Jianping Zhang
Jianping Zhang, Yizhan Huang, Weibin Wu, Michael R. Lyu
Transferable Adversarial Attacks on Vision Transformers with Token Gradient Regularization
CVPR 2023, Code is available at https://github.com/jpzhang1810/TGR
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Vision transformers (ViTs) have been successfully deployed in a variety of computer vision tasks, but they are still vulnerable to adversarial samples. Transfer-based attacks use a local model to generate adversarial samples and directly transfer them to attack a target black-box model. The high efficiency of transfer-based attacks makes it a severe security threat to ViT-based applications. Therefore, it is vital to design effective transfer-based attacks to identify the deficiencies of ViTs beforehand in security-sensitive scenarios. Existing efforts generally focus on regularizing the input gradients to stabilize the updated direction of adversarial samples. However, the variance of the back-propagated gradients in intermediate blocks of ViTs may still be large, which may make the generated adversarial samples focus on some model-specific features and get stuck in poor local optima. To overcome the shortcomings of existing approaches, we propose the Token Gradient Regularization (TGR) method. According to the structural characteristics of ViTs, TGR reduces the variance of the back-propagated gradient in each internal block of ViTs in a token-wise manner and utilizes the regularized gradient to generate adversarial samples. Extensive experiments on attacking both ViTs and CNNs confirm the superiority of our approach. Notably, compared to the state-of-the-art transfer-based attacks, our TGR offers a performance improvement of 8.8% on average.
[ { "created": "Tue, 28 Mar 2023 06:23:17 GMT", "version": "v1" }, { "created": "Mon, 5 Jun 2023 07:25:12 GMT", "version": "v2" } ]
2023-06-06
[ [ "Zhang", "Jianping", "" ], [ "Huang", "Yizhan", "" ], [ "Wu", "Weibin", "" ], [ "Lyu", "Michael R.", "" ] ]
Vision transformers (ViTs) have been successfully deployed in a variety of computer vision tasks, but they are still vulnerable to adversarial samples. Transfer-based attacks use a local model to generate adversarial samples and directly transfer them to attack a target black-box model. The high efficiency of transfer-based attacks makes it a severe security threat to ViT-based applications. Therefore, it is vital to design effective transfer-based attacks to identify the deficiencies of ViTs beforehand in security-sensitive scenarios. Existing efforts generally focus on regularizing the input gradients to stabilize the updated direction of adversarial samples. However, the variance of the back-propagated gradients in intermediate blocks of ViTs may still be large, which may make the generated adversarial samples focus on some model-specific features and get stuck in poor local optima. To overcome the shortcomings of existing approaches, we propose the Token Gradient Regularization (TGR) method. According to the structural characteristics of ViTs, TGR reduces the variance of the back-propagated gradient in each internal block of ViTs in a token-wise manner and utilizes the regularized gradient to generate adversarial samples. Extensive experiments on attacking both ViTs and CNNs confirm the superiority of our approach. Notably, compared to the state-of-the-art transfer-based attacks, our TGR offers a performance improvement of 8.8% on average.
2404.11045
James Y. Huang
James Y. Huang, Wenxuan Zhou, Fei Wang, Fred Morstatter, Sheng Zhang, Hoifung Poon, Muhao Chen
Offset Unlearning for Large Language Models
null
null
null
null
cs.CL
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Despite the strong capabilities of Large Language Models (LLMs) to acquire knowledge from their training corpora, the memorization of sensitive information in the corpora such as copyrighted, harmful, and private content has led to ethical and legal concerns. In response to these challenges, unlearning has emerged as a potential remedy for LLMs affected by problematic training data. However, previous unlearning techniques are either not applicable to black-box LLMs due to required access to model internal weights, or violate data protection principles by retaining sensitive data for inference-time correction. We propose $\delta$-unlearning, an offset unlearning framework for black-box LLMs. Instead of tuning the black-box LLM itself, $\delta$-unlearning learns the logit offset needed for unlearning by contrasting the logits from a pair of smaller models. Experiments demonstrate that $\delta$-unlearning can effectively unlearn target data while maintaining similar or even stronger performance on general out-of-forget-scope tasks. $\delta$-unlearning also effectively incorporates different unlearning algorithms, making our approach a versatile solution to adapting various existing unlearning algorithms to black-box LLMs.
[ { "created": "Wed, 17 Apr 2024 03:39:51 GMT", "version": "v1" } ]
2024-04-18
[ [ "Huang", "James Y.", "" ], [ "Zhou", "Wenxuan", "" ], [ "Wang", "Fei", "" ], [ "Morstatter", "Fred", "" ], [ "Zhang", "Sheng", "" ], [ "Poon", "Hoifung", "" ], [ "Chen", "Muhao", "" ] ]
Despite the strong capabilities of Large Language Models (LLMs) to acquire knowledge from their training corpora, the memorization of sensitive information in the corpora such as copyrighted, harmful, and private content has led to ethical and legal concerns. In response to these challenges, unlearning has emerged as a potential remedy for LLMs affected by problematic training data. However, previous unlearning techniques are either not applicable to black-box LLMs due to required access to model internal weights, or violate data protection principles by retaining sensitive data for inference-time correction. We propose $\delta$-unlearning, an offset unlearning framework for black-box LLMs. Instead of tuning the black-box LLM itself, $\delta$-unlearning learns the logit offset needed for unlearning by contrasting the logits from a pair of smaller models. Experiments demonstrate that $\delta$-unlearning can effectively unlearn target data while maintaining similar or even stronger performance on general out-of-forget-scope tasks. $\delta$-unlearning also effectively incorporates different unlearning algorithms, making our approach a versatile solution to adapting various existing unlearning algorithms to black-box LLMs.
2006.15374
Debashmita Poddar
Gianlorenzo D'Angelo, Debashmita Poddar, Cosimo Vinci
Better Bounds on the Adaptivity Gap of Influence Maximization under Full-adoption Feedback
18 pages
The 35th AAAI Conference on Artificial Intelligence (AAAI 2021)
null
null
cs.SI cs.LG
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
In the influence maximization (IM) problem, we are given a social network and a budget $k$, and we look for a set of $k$ nodes in the network, called seeds, that maximize the expected number of nodes that are reached by an influence cascade generated by the seeds, according to some stochastic model for influence diffusion. In this paper, we study the adaptive IM, where the nodes are selected sequentially one by one, and the decision on the $i$th seed can be based on the observed cascade produced by the first $i-1$ seeds. We focus on the full-adoption feedback in which we can observe the entire cascade of each previously selected seed and on the independent cascade model where each edge is associated with an independent probability of diffusing influence. Our main result is the first sub-linear upper bound that holds for any graph. Specifically, we show that the adaptivity gap is upper-bounded by $\lceil n^{1/3}\rceil $, where $n$ is the number of nodes in the graph. Moreover, we improve over the known upper bound for in-arborescences from $\frac{2e}{e-1}\approx 3.16$ to $\frac{2e^2}{e^2-1}\approx 2.31$. Finally, we study $\alpha$-bounded graphs, a class of undirected graphs in which the sum of node degrees higher than two is at most $\alpha$, and show that the adaptivity gap is upper-bounded by $\sqrt{\alpha}+O(1)$. Moreover, we show that in 0-bounded graphs, i.e. undirected graphs in which each connected component is a path or a cycle, the adaptivity gap is at most $\frac{3e^3}{e^3-1}\approx 3.16$. To prove our bounds, we introduce new techniques to relate adaptive policies with non-adaptive ones that might be of their own interest.
[ { "created": "Sat, 27 Jun 2020 14:43:34 GMT", "version": "v1" } ]
2021-05-11
[ [ "D'Angelo", "Gianlorenzo", "" ], [ "Poddar", "Debashmita", "" ], [ "Vinci", "Cosimo", "" ] ]
In the influence maximization (IM) problem, we are given a social network and a budget $k$, and we look for a set of $k$ nodes in the network, called seeds, that maximize the expected number of nodes that are reached by an influence cascade generated by the seeds, according to some stochastic model for influence diffusion. In this paper, we study the adaptive IM, where the nodes are selected sequentially one by one, and the decision on the $i$th seed can be based on the observed cascade produced by the first $i-1$ seeds. We focus on the full-adoption feedback in which we can observe the entire cascade of each previously selected seed and on the independent cascade model where each edge is associated with an independent probability of diffusing influence. Our main result is the first sub-linear upper bound that holds for any graph. Specifically, we show that the adaptivity gap is upper-bounded by $\lceil n^{1/3}\rceil $, where $n$ is the number of nodes in the graph. Moreover, we improve over the known upper bound for in-arborescences from $\frac{2e}{e-1}\approx 3.16$ to $\frac{2e^2}{e^2-1}\approx 2.31$. Finally, we study $\alpha$-bounded graphs, a class of undirected graphs in which the sum of node degrees higher than two is at most $\alpha$, and show that the adaptivity gap is upper-bounded by $\sqrt{\alpha}+O(1)$. Moreover, we show that in 0-bounded graphs, i.e. undirected graphs in which each connected component is a path or a cycle, the adaptivity gap is at most $\frac{3e^3}{e^3-1}\approx 3.16$. To prove our bounds, we introduce new techniques to relate adaptive policies with non-adaptive ones that might be of their own interest.
2406.15565
Paridhi Singh
Paridhi Singh, Arun Kumar
Unseen Object Reasoning with Shared Appearance Cues
null
null
null
null
cs.CV cs.LG
http://creativecommons.org/licenses/by-sa/4.0/
This paper introduces an innovative approach to open world recognition (OWR), where we leverage knowledge acquired from known objects to address the recognition of previously unseen objects. The traditional method of object modeling relies on supervised learning with strict closed-set assumptions, presupposing that objects encountered during inference are already known at the training phase. However, this assumption proves inadequate for real-world scenarios due to the impracticality of accounting for the immense diversity of objects. Our hypothesis posits that object appearances can be represented as collections of "shareable" mid-level features, arranged in constellations to form object instances. By adopting this framework, we can efficiently dissect and represent both known and unknown objects in terms of their appearance cues. Our paper introduces a straightforward yet elegant method for modeling novel or unseen objects, utilizing established appearance cues and accounting for inherent uncertainties. This representation not only enables the detection of out-of-distribution objects or novel categories among unseen objects but also facilitates a deeper level of reasoning, empowering the identification of the superclass to which an unknown instance belongs. This novel approach holds promise for advancing open world recognition in diverse applications.
[ { "created": "Fri, 21 Jun 2024 18:04:13 GMT", "version": "v1" } ]
2024-06-25
[ [ "Singh", "Paridhi", "" ], [ "Kumar", "Arun", "" ] ]
This paper introduces an innovative approach to open world recognition (OWR), where we leverage knowledge acquired from known objects to address the recognition of previously unseen objects. The traditional method of object modeling relies on supervised learning with strict closed-set assumptions, presupposing that objects encountered during inference are already known at the training phase. However, this assumption proves inadequate for real-world scenarios due to the impracticality of accounting for the immense diversity of objects. Our hypothesis posits that object appearances can be represented as collections of "shareable" mid-level features, arranged in constellations to form object instances. By adopting this framework, we can efficiently dissect and represent both known and unknown objects in terms of their appearance cues. Our paper introduces a straightforward yet elegant method for modeling novel or unseen objects, utilizing established appearance cues and accounting for inherent uncertainties. This representation not only enables the detection of out-of-distribution objects or novel categories among unseen objects but also facilitates a deeper level of reasoning, empowering the identification of the superclass to which an unknown instance belongs. This novel approach holds promise for advancing open world recognition in diverse applications.
2310.09166
Seth Benson
Seth P. Benson and Iain J. Cruickshank
Developing a Natural Language Understanding Model to Characterize Cable News Bias
null
null
null
null
cs.CL
http://creativecommons.org/licenses/by/4.0/
Media bias has been extensively studied by both social and computational sciences. However, current work still has a large reliance on human input and subjective assessment to label biases. This is especially true for cable news research. To address these issues, we develop an unsupervised machine learning method to characterize the bias of cable news programs without any human input. This method relies on the analysis of what topics are mentioned through Named Entity Recognition and how those topics are discussed through Stance Analysis in order to cluster programs with similar biases together. Applying our method to 2020 cable news transcripts, we find that program clusters are consistent over time and roughly correspond to the cable news network of the program. This method reveals the potential for future tools to objectively assess media bias and characterize unfamiliar media environments.
[ { "created": "Fri, 13 Oct 2023 15:01:17 GMT", "version": "v1" }, { "created": "Tue, 17 Oct 2023 22:37:58 GMT", "version": "v2" } ]
2023-10-19
[ [ "Benson", "Seth P.", "" ], [ "Cruickshank", "Iain J.", "" ] ]
Media bias has been extensively studied by both social and computational sciences. However, current work still has a large reliance on human input and subjective assessment to label biases. This is especially true for cable news research. To address these issues, we develop an unsupervised machine learning method to characterize the bias of cable news programs without any human input. This method relies on the analysis of what topics are mentioned through Named Entity Recognition and how those topics are discussed through Stance Analysis in order to cluster programs with similar biases together. Applying our method to 2020 cable news transcripts, we find that program clusters are consistent over time and roughly correspond to the cable news network of the program. This method reveals the potential for future tools to objectively assess media bias and characterize unfamiliar media environments.
2312.03721
Simon Lermen
Simon Lermen and Ond\v{r}ej Kvapil
Exploring the Robustness of Model-Graded Evaluations and Automated Interpretability
null
null
null
null
cs.CL cs.AI
http://creativecommons.org/licenses/by/4.0/
There has been increasing interest in evaluations of language models for a variety of risks and characteristics. Evaluations relying on natural language understanding for grading can often be performed at scale by using other language models. We test the robustness of these model-graded evaluations to injections on different datasets including a new Deception Eval. These injections resemble direct communication between the testee and the evaluator to change their grading. We extrapolate that future, more intelligent models might manipulate or cooperate with their evaluation model. We find significant susceptibility to these injections in state-of-the-art commercial models on all examined evaluations. Furthermore, similar injections can be used on automated interpretability frameworks to produce misleading model-written explanations. The results inspire future work and should caution against unqualified trust in evaluations and automated interpretability.
[ { "created": "Sun, 26 Nov 2023 17:11:55 GMT", "version": "v1" }, { "created": "Fri, 8 Dec 2023 11:16:39 GMT", "version": "v2" } ]
2023-12-11
[ [ "Lermen", "Simon", "" ], [ "Kvapil", "Ondřej", "" ] ]
There has been increasing interest in evaluations of language models for a variety of risks and characteristics. Evaluations relying on natural language understanding for grading can often be performed at scale by using other language models. We test the robustness of these model-graded evaluations to injections on different datasets including a new Deception Eval. These injections resemble direct communication between the testee and the evaluator to change their grading. We extrapolate that future, more intelligent models might manipulate or cooperate with their evaluation model. We find significant susceptibility to these injections in state-of-the-art commercial models on all examined evaluations. Furthermore, similar injections can be used on automated interpretability frameworks to produce misleading model-written explanations. The results inspire future work and should caution against unqualified trust in evaluations and automated interpretability.
2306.06872
Hao Sun
Hao Sun, Yang Li, Liwei Deng, Bowen Li, Binyuan Hui, Binhua Li, Yunshi Lan, Yan Zhang, Yongbin Li
History Semantic Graph Enhanced Conversational KBQA with Temporal Information Modeling
Accepted to ACL 2023 Main Conference
null
null
null
cs.CL
http://creativecommons.org/licenses/by/4.0/
Context information modeling is an important task in conversational KBQA. However, existing methods usually assume the independence of utterances and model them in isolation. In this paper, we propose a History Semantic Graph Enhanced KBQA model (HSGE) that is able to effectively model long-range semantic dependencies in conversation history while maintaining low computational cost. The framework incorporates a context-aware encoder, which employs a dynamic memory decay mechanism and models context at different levels of granularity. We evaluate HSGE on a widely used benchmark dataset for complex sequential question answering. Experimental results demonstrate that it outperforms existing baselines averaged on all question types.
[ { "created": "Mon, 12 Jun 2023 05:10:58 GMT", "version": "v1" } ]
2023-06-13
[ [ "Sun", "Hao", "" ], [ "Li", "Yang", "" ], [ "Deng", "Liwei", "" ], [ "Li", "Bowen", "" ], [ "Hui", "Binyuan", "" ], [ "Li", "Binhua", "" ], [ "Lan", "Yunshi", "" ], [ "Zhang", "Yan", "" ], [ "Li", "Yongbin", "" ] ]
Context information modeling is an important task in conversational KBQA. However, existing methods usually assume the independence of utterances and model them in isolation. In this paper, we propose a History Semantic Graph Enhanced KBQA model (HSGE) that is able to effectively model long-range semantic dependencies in conversation history while maintaining low computational cost. The framework incorporates a context-aware encoder, which employs a dynamic memory decay mechanism and models context at different levels of granularity. We evaluate HSGE on a widely used benchmark dataset for complex sequential question answering. Experimental results demonstrate that it outperforms existing baselines averaged on all question types.
2404.03344
Juri Opitz
Juri Opitz
Schroedinger's Threshold: When the AUC doesn't predict Accuracy
LREC-COLING 2024, added more details on data setups, fixed typo
null
null
null
cs.CL
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
The Area Under Curve measure (AUC) seems apt to evaluate and compare diverse models, possibly without calibration. An important example of AUC application is the evaluation and benchmarking of models that predict faithfulness of generated text. But we show that the AUC yields an academic and optimistic notion of accuracy that can misalign with the actual accuracy observed in application, yielding significant changes in benchmark rankings. To paint a more realistic picture of downstream model performance (and prepare a model for actual application), we explore different calibration modes, testing calibration data and method.
[ { "created": "Thu, 4 Apr 2024 10:18:03 GMT", "version": "v1" }, { "created": "Mon, 27 May 2024 10:33:40 GMT", "version": "v2" } ]
2024-05-28
[ [ "Opitz", "Juri", "" ] ]
The Area Under Curve measure (AUC) seems apt to evaluate and compare diverse models, possibly without calibration. An important example of AUC application is the evaluation and benchmarking of models that predict faithfulness of generated text. But we show that the AUC yields an academic and optimistic notion of accuracy that can misalign with the actual accuracy observed in application, yielding significant changes in benchmark rankings. To paint a more realistic picture of downstream model performance (and prepare a model for actual application), we explore different calibration modes, testing calibration data and method.
1804.09160
Xin Wang
Xin Wang, Wenhu Chen, Yuan-Fang Wang, William Yang Wang
No Metrics Are Perfect: Adversarial Reward Learning for Visual Storytelling
ACL 2018. 15 pages, 10 figures, 4 tables, with supplementary material
null
null
null
cs.CL cs.AI cs.CV cs.LG
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Though impressive results have been achieved in visual captioning, the task of generating abstract stories from photo streams is still a little-tapped problem. Different from captions, stories have more expressive language styles and contain many imaginary concepts that do not appear in the images. Thus it poses challenges to behavioral cloning algorithms. Furthermore, due to the limitations of automatic metrics on evaluating story quality, reinforcement learning methods with hand-crafted rewards also face difficulties in gaining an overall performance boost. Therefore, we propose an Adversarial REward Learning (AREL) framework to learn an implicit reward function from human demonstrations, and then optimize policy search with the learned reward function. Though automatic eval- uation indicates slight performance boost over state-of-the-art (SOTA) methods in cloning expert behaviors, human evaluation shows that our approach achieves significant improvement in generating more human-like stories than SOTA systems.
[ { "created": "Tue, 24 Apr 2018 17:41:24 GMT", "version": "v1" }, { "created": "Mon, 9 Jul 2018 00:15:14 GMT", "version": "v2" } ]
2018-07-10
[ [ "Wang", "Xin", "" ], [ "Chen", "Wenhu", "" ], [ "Wang", "Yuan-Fang", "" ], [ "Wang", "William Yang", "" ] ]
Though impressive results have been achieved in visual captioning, the task of generating abstract stories from photo streams is still a little-tapped problem. Different from captions, stories have more expressive language styles and contain many imaginary concepts that do not appear in the images. Thus it poses challenges to behavioral cloning algorithms. Furthermore, due to the limitations of automatic metrics on evaluating story quality, reinforcement learning methods with hand-crafted rewards also face difficulties in gaining an overall performance boost. Therefore, we propose an Adversarial REward Learning (AREL) framework to learn an implicit reward function from human demonstrations, and then optimize policy search with the learned reward function. Though automatic eval- uation indicates slight performance boost over state-of-the-art (SOTA) methods in cloning expert behaviors, human evaluation shows that our approach achieves significant improvement in generating more human-like stories than SOTA systems.
2403.20046
Yongqi Tong
Yongqi Tong, Dawei Li, Sizhe Wang, Yujia Wang, Fei Teng, Jingbo Shang
Can LLMs Learn from Previous Mistakes? Investigating LLMs' Errors to Boost for Reasoning
The 62nd Annual Meeting of the Association for Computational Linguistics (ACL 2024) - Main Conference
null
null
null
cs.CL
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Recent works have shown the benefits to LLMs from fine-tuning golden-standard Chain-of-Thought (CoT) rationales or using them as correct examples in few-shot prompting. While humans can indeed imitate correct examples, learning from our mistakes is another vital aspect of human cognition. Hence, a question naturally arises: \textit{can LLMs learn and benefit from their mistakes, especially for their reasoning? } This study investigates this problem from both the prompting and model-tuning perspectives. We begin by introducing \textsc{CoTErrorSet}, a new benchmark with 609,432 questions, each designed with both correct and error references, and demonstrating the types and reasons for making such mistakes. To explore the effectiveness of those mistakes, we design two methods: (1) \textbf{Self-rethinking} prompting guides LLMs to rethink whether they have made similar previous mistakes; and (2) \textbf{Mistake tuning} involves finetuning models in both correct and incorrect reasoning domains, rather than only tuning models to learn ground truth in traditional methodology. We conduct a series of experiments to prove LLMs can obtain benefits from mistakes in both directions. Our two methods offer potentially cost-effective strategies by leveraging errors to enhance reasoning capabilities, which costs significantly less than creating meticulously hand-crafted golden references. We ultimately make a thorough analysis of the reasons behind LLMs' errors, which provides directions that future research needs to overcome. \textsc{CoTErrorSet} will be published soon on \texttt{\url{https://github.com/YookiTong/Learn-from-Mistakes-CotErrorSet}}.
[ { "created": "Fri, 29 Mar 2024 08:30:34 GMT", "version": "v1" }, { "created": "Fri, 7 Jun 2024 06:27:50 GMT", "version": "v2" } ]
2024-06-10
[ [ "Tong", "Yongqi", "" ], [ "Li", "Dawei", "" ], [ "Wang", "Sizhe", "" ], [ "Wang", "Yujia", "" ], [ "Teng", "Fei", "" ], [ "Shang", "Jingbo", "" ] ]
Recent works have shown the benefits to LLMs from fine-tuning golden-standard Chain-of-Thought (CoT) rationales or using them as correct examples in few-shot prompting. While humans can indeed imitate correct examples, learning from our mistakes is another vital aspect of human cognition. Hence, a question naturally arises: \textit{can LLMs learn and benefit from their mistakes, especially for their reasoning? } This study investigates this problem from both the prompting and model-tuning perspectives. We begin by introducing \textsc{CoTErrorSet}, a new benchmark with 609,432 questions, each designed with both correct and error references, and demonstrating the types and reasons for making such mistakes. To explore the effectiveness of those mistakes, we design two methods: (1) \textbf{Self-rethinking} prompting guides LLMs to rethink whether they have made similar previous mistakes; and (2) \textbf{Mistake tuning} involves finetuning models in both correct and incorrect reasoning domains, rather than only tuning models to learn ground truth in traditional methodology. We conduct a series of experiments to prove LLMs can obtain benefits from mistakes in both directions. Our two methods offer potentially cost-effective strategies by leveraging errors to enhance reasoning capabilities, which costs significantly less than creating meticulously hand-crafted golden references. We ultimately make a thorough analysis of the reasons behind LLMs' errors, which provides directions that future research needs to overcome. \textsc{CoTErrorSet} will be published soon on \texttt{\url{https://github.com/YookiTong/Learn-from-Mistakes-CotErrorSet}}.
1902.06034
Michihiro Yasunaga
Michihiro Yasunaga, John Lafferty
TopicEq: A Joint Topic and Mathematical Equation Model for Scientific Texts
AAAI 2019
null
null
null
cs.IR cs.CL cs.LG stat.ML
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Scientific documents rely on both mathematics and text to communicate ideas. Inspired by the topical correspondence between mathematical equations and word contexts observed in scientific texts, we propose a novel topic model that jointly generates mathematical equations and their surrounding text (TopicEq). Using an extension of the correlated topic model, the context is generated from a mixture of latent topics, and the equation is generated by an RNN that depends on the latent topic activations. To experiment with this model, we create a corpus of 400K equation-context pairs extracted from a range of scientific articles from arXiv, and fit the model using a variational autoencoder approach. Experimental results show that this joint model significantly outperforms existing topic models and equation models for scientific texts. Moreover, we qualitatively show that the model effectively captures the relationship between topics and mathematics, enabling novel applications such as topic-aware equation generation, equation topic inference, and topic-aware alignment of mathematical symbols and words.
[ { "created": "Sat, 16 Feb 2019 03:39:51 GMT", "version": "v1" }, { "created": "Wed, 20 Feb 2019 16:55:23 GMT", "version": "v2" }, { "created": "Thu, 25 Apr 2019 21:24:05 GMT", "version": "v3" } ]
2019-04-29
[ [ "Yasunaga", "Michihiro", "" ], [ "Lafferty", "John", "" ] ]
Scientific documents rely on both mathematics and text to communicate ideas. Inspired by the topical correspondence between mathematical equations and word contexts observed in scientific texts, we propose a novel topic model that jointly generates mathematical equations and their surrounding text (TopicEq). Using an extension of the correlated topic model, the context is generated from a mixture of latent topics, and the equation is generated by an RNN that depends on the latent topic activations. To experiment with this model, we create a corpus of 400K equation-context pairs extracted from a range of scientific articles from arXiv, and fit the model using a variational autoencoder approach. Experimental results show that this joint model significantly outperforms existing topic models and equation models for scientific texts. Moreover, we qualitatively show that the model effectively captures the relationship between topics and mathematics, enabling novel applications such as topic-aware equation generation, equation topic inference, and topic-aware alignment of mathematical symbols and words.
1302.5442
Weisheng Si
Weisheng Si
Are Yao Graph and Theta Graph Void Free?
4 pages
null
null
null
cs.NI
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Greedy Forwarding algorithm is a widely-used routing algorithm for wireless networks. However, it can fail if network topologies (usually modeled by geometric graphs) contain voids. Since Yao Graph and Theta Graph are two types of geometric graphs exploited to construct wireless network topologies, this paper studies whether these two types of graphs can contain voids. Specifically, this paper shows that when the number of cones in a Yao Graph or Theta Graph is less than 6, Yao Graph and Theta Graph can have voids, but when the number of cones equals or exceeds 6, Yao Graph and Theta Graph are free of voids.
[ { "created": "Thu, 21 Feb 2013 22:09:08 GMT", "version": "v1" }, { "created": "Tue, 2 Jul 2013 23:04:20 GMT", "version": "v2" } ]
2013-07-04
[ [ "Si", "Weisheng", "" ] ]
Greedy Forwarding algorithm is a widely-used routing algorithm for wireless networks. However, it can fail if network topologies (usually modeled by geometric graphs) contain voids. Since Yao Graph and Theta Graph are two types of geometric graphs exploited to construct wireless network topologies, this paper studies whether these two types of graphs can contain voids. Specifically, this paper shows that when the number of cones in a Yao Graph or Theta Graph is less than 6, Yao Graph and Theta Graph can have voids, but when the number of cones equals or exceeds 6, Yao Graph and Theta Graph are free of voids.
1011.1531
Jaydip Sen
Jaydip Sen
An Agent-Based Intrusion Detection System for Local Area Networks
13 pages, 5 figures, 2 tables
International Journal of Communication Networks and Information Security (IJCNIS), Vol 2, No 2, August 2010
null
null
cs.CR cs.NI
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Since it is impossible to predict and identify all the vulnerabilities of a network beforehand, and penetration into a system by malicious intruders cannot always be prevented, intrusion detection systems (IDSs) are essential entities to ensure the security of a networked system. To be effective in carrying out their functions, the IDSs need to be accurate, adaptive, and extensible. Given these stringent requirements and the high level of vulnerabilities of the current days' networks, the design of an IDS has become a very challenging task. Although, an extensive research has been done on intrusion detection in a distributed environment, distributed IDSs suffer from a number of drawbacks e.g., high rates of false positives, low detection efficiency etc. In this paper, the design of a distributed IDS is proposed that consists of a group of autonomous and cooperating agents. In addition to its ability to detect attacks, the system is capable of identifying and isolating compromised nodes in the network thereby introducing fault-tolerance in its operations. The experiments conducted on the system have shown that it has a high detection efficiency and low false positives compared to some of the currently existing systems.
[ { "created": "Sat, 6 Nov 2010 01:05:20 GMT", "version": "v1" } ]
2010-11-13
[ [ "Sen", "Jaydip", "" ] ]
Since it is impossible to predict and identify all the vulnerabilities of a network beforehand, and penetration into a system by malicious intruders cannot always be prevented, intrusion detection systems (IDSs) are essential entities to ensure the security of a networked system. To be effective in carrying out their functions, the IDSs need to be accurate, adaptive, and extensible. Given these stringent requirements and the high level of vulnerabilities of the current days' networks, the design of an IDS has become a very challenging task. Although, an extensive research has been done on intrusion detection in a distributed environment, distributed IDSs suffer from a number of drawbacks e.g., high rates of false positives, low detection efficiency etc. In this paper, the design of a distributed IDS is proposed that consists of a group of autonomous and cooperating agents. In addition to its ability to detect attacks, the system is capable of identifying and isolating compromised nodes in the network thereby introducing fault-tolerance in its operations. The experiments conducted on the system have shown that it has a high detection efficiency and low false positives compared to some of the currently existing systems.
1509.02840
Andrew Knyazev
Andrew Knyazev, Peizhen Zhu, Stefano Di Cairano
Explicit model predictive control accuracy analysis
6 pages, 7 figures. Accepted to IEEE CDC 2015
2015 54th IEEE Conference on Decision and Control (CDC), Osaka, 2015, pp. 2389-2394
10.1109/CDC.2015.7402565
MERL TR2015-149
cs.SY math.NA math.OC
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Model Predictive Control (MPC) can efficiently control constrained systems in real-time applications. MPC feedback law for a linear system with linear inequality constraints can be explicitly computed off-line, which results in an off-line partition of the state space into non-overlapped convex regions, with affine control laws associated to each region of the partition. An actual implementation of this explicit MPC in low cost micro-controllers requires the data to be "quantized", i.e. represented with a small number of memory bits. An aggressive quantization decreases the number of bits and the controller manufacturing costs, and may increase the speed of the controller, but reduces accuracy of the control input computation. We derive upper bounds for the absolute error in the control depending on the number of quantization bits and system parameters. The bounds can be used to determine how many quantization bits are needed in order to guarantee a specific level of accuracy in the control input.
[ { "created": "Wed, 9 Sep 2015 16:22:08 GMT", "version": "v1" } ]
2016-06-13
[ [ "Knyazev", "Andrew", "" ], [ "Zhu", "Peizhen", "" ], [ "Di Cairano", "Stefano", "" ] ]
Model Predictive Control (MPC) can efficiently control constrained systems in real-time applications. MPC feedback law for a linear system with linear inequality constraints can be explicitly computed off-line, which results in an off-line partition of the state space into non-overlapped convex regions, with affine control laws associated to each region of the partition. An actual implementation of this explicit MPC in low cost micro-controllers requires the data to be "quantized", i.e. represented with a small number of memory bits. An aggressive quantization decreases the number of bits and the controller manufacturing costs, and may increase the speed of the controller, but reduces accuracy of the control input computation. We derive upper bounds for the absolute error in the control depending on the number of quantization bits and system parameters. The bounds can be used to determine how many quantization bits are needed in order to guarantee a specific level of accuracy in the control input.
2407.08334
TianChen Wang
TianChen Wang
ADMM Based Semi-Structured Pattern Pruning Framework For Transformer
11 pages, 5 figures
null
null
null
cs.IR
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
NLP(natural language processsing) has achieved great success through the transformer model.However, the model has hundreds of millions or billions parameters,which is huge burden for its deployment on personal computer or small scale of server.To deal with it, we either make the model's weight matrix relatively sparser, or compress attention layer. Pattern pruning ,one of the most important pruning methods, permits selecting fixed number of parameters in each divided pattern block and prunes it. However, the effect of pattern pruning is strictly limited by the sparsity within a region of weights in each layer. In this paper,we first introduced Alternating Direction Method of Multipliers(ADMM) based pattern pruning framework to reshape the distribution of activation map. Specifically, we propose to formulate the pattern pruning on transformer as a constrained optimization and use ADMM to optimize the problem. In this way, the initial dense feature maps is transformed to rather regionally sparsified ones.Therefore, we can then achieve higher compression ratio with better performance based on pattern pruning method. Additionally, this paper provides a theoretical derivations of the ADMM with local sparsity. Finally, we also extend the proposed ADMM based framework with SR-STE to demonstrate its generalization and to avoid gradient vanishing problem. We conduct extensive experiments on classification tasks over GLUE datasets. Significantly, we achieve 50% percent compression ratio while maintaining overall score 80.1 on GLUE dataset.
[ { "created": "Thu, 11 Jul 2024 09:35:08 GMT", "version": "v1" }, { "created": "Fri, 12 Jul 2024 03:36:01 GMT", "version": "v2" }, { "created": "Sat, 20 Jul 2024 03:40:43 GMT", "version": "v3" } ]
2024-07-23
[ [ "Wang", "TianChen", "" ] ]
NLP(natural language processsing) has achieved great success through the transformer model.However, the model has hundreds of millions or billions parameters,which is huge burden for its deployment on personal computer or small scale of server.To deal with it, we either make the model's weight matrix relatively sparser, or compress attention layer. Pattern pruning ,one of the most important pruning methods, permits selecting fixed number of parameters in each divided pattern block and prunes it. However, the effect of pattern pruning is strictly limited by the sparsity within a region of weights in each layer. In this paper,we first introduced Alternating Direction Method of Multipliers(ADMM) based pattern pruning framework to reshape the distribution of activation map. Specifically, we propose to formulate the pattern pruning on transformer as a constrained optimization and use ADMM to optimize the problem. In this way, the initial dense feature maps is transformed to rather regionally sparsified ones.Therefore, we can then achieve higher compression ratio with better performance based on pattern pruning method. Additionally, this paper provides a theoretical derivations of the ADMM with local sparsity. Finally, we also extend the proposed ADMM based framework with SR-STE to demonstrate its generalization and to avoid gradient vanishing problem. We conduct extensive experiments on classification tasks over GLUE datasets. Significantly, we achieve 50% percent compression ratio while maintaining overall score 80.1 on GLUE dataset.
2209.13017
Karish Grover
Karish Grover, S.M. Phaneendra Angara, Md. Shad Akhtar, Tanmoy Chakraborty
Public Wisdom Matters! Discourse-Aware Hyperbolic Fourier Co-Attention for Social-Text Classification
NeurIPS 2022
null
null
null
cs.CL cs.LG cs.SI
http://creativecommons.org/licenses/by/4.0/
Social media has become the fulcrum of all forms of communication. Classifying social texts such as fake news, rumour, sarcasm, etc. has gained significant attention. The surface-level signals expressed by a social-text itself may not be adequate for such tasks; therefore, recent methods attempted to incorporate other intrinsic signals such as user behavior and the underlying graph structure. Oftentimes, the `public wisdom' expressed through the comments/replies to a social-text acts as a surrogate of crowd-sourced view and may provide us with complementary signals. State-of-the-art methods on social-text classification tend to ignore such a rich hierarchical signal. Here, we propose Hyphen, a discourse-aware hyperbolic spectral co-attention network. Hyphen is a fusion of hyperbolic graph representation learning with a novel Fourier co-attention mechanism in an attempt to generalise the social-text classification tasks by incorporating public discourse. We parse public discourse as an Abstract Meaning Representation (AMR) graph and use the powerful hyperbolic geometric representation to model graphs with hierarchical structure. Finally, we equip it with a novel Fourier co-attention mechanism to capture the correlation between the source post and public discourse. Extensive experiments on four different social-text classification tasks, namely detecting fake news, hate speech, rumour, and sarcasm, show that Hyphen generalises well, and achieves state-of-the-art results on ten benchmark datasets. We also employ a sentence-level fact-checked and annotated dataset to evaluate how Hyphen is capable of producing explanations as analogous evidence to the final prediction.
[ { "created": "Thu, 15 Sep 2022 16:04:32 GMT", "version": "v1" }, { "created": "Tue, 11 Oct 2022 15:57:31 GMT", "version": "v2" } ]
2022-10-12
[ [ "Grover", "Karish", "" ], [ "Angara", "S. M. Phaneendra", "" ], [ "Akhtar", "Md. Shad", "" ], [ "Chakraborty", "Tanmoy", "" ] ]
Social media has become the fulcrum of all forms of communication. Classifying social texts such as fake news, rumour, sarcasm, etc. has gained significant attention. The surface-level signals expressed by a social-text itself may not be adequate for such tasks; therefore, recent methods attempted to incorporate other intrinsic signals such as user behavior and the underlying graph structure. Oftentimes, the `public wisdom' expressed through the comments/replies to a social-text acts as a surrogate of crowd-sourced view and may provide us with complementary signals. State-of-the-art methods on social-text classification tend to ignore such a rich hierarchical signal. Here, we propose Hyphen, a discourse-aware hyperbolic spectral co-attention network. Hyphen is a fusion of hyperbolic graph representation learning with a novel Fourier co-attention mechanism in an attempt to generalise the social-text classification tasks by incorporating public discourse. We parse public discourse as an Abstract Meaning Representation (AMR) graph and use the powerful hyperbolic geometric representation to model graphs with hierarchical structure. Finally, we equip it with a novel Fourier co-attention mechanism to capture the correlation between the source post and public discourse. Extensive experiments on four different social-text classification tasks, namely detecting fake news, hate speech, rumour, and sarcasm, show that Hyphen generalises well, and achieves state-of-the-art results on ten benchmark datasets. We also employ a sentence-level fact-checked and annotated dataset to evaluate how Hyphen is capable of producing explanations as analogous evidence to the final prediction.
2407.20530
Youqiang Zheng
Youqiang Zheng, Weiping Tu, Li Xiao, Xinmeng Xu
SuperCodec: A Neural Speech Codec with Selective Back-Projection Network
Accepted by ICASSP 2024
null
null
null
cs.SD eess.AS
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Neural speech coding is a rapidly developing topic, where state-of-the-art approaches now exhibit superior compression performance than conventional methods. Despite significant progress, existing methods still have limitations in preserving and reconstructing fine details for optimal reconstruction, especially at low bitrates. In this study, we introduce SuperCodec, a neural speech codec that achieves state-of-the-art performance at low bitrates. It employs a novel back projection method with selective feature fusion for augmented representation. Specifically, we propose to use Selective Up-sampling Back Projection (SUBP) and Selective Down-sampling Back Projection (SDBP) modules to replace the standard up- and down-sampling layers at the encoder and decoder, respectively. Experimental results show that our method outperforms the existing neural speech codecs operating at various bitrates. Specifically, our proposed method can achieve higher quality reconstructed speech at 1 kbps than Lyra V2 at 3.2 kbps and Encodec at 6 kbps.
[ { "created": "Tue, 30 Jul 2024 04:12:17 GMT", "version": "v1" } ]
2024-07-31
[ [ "Zheng", "Youqiang", "" ], [ "Tu", "Weiping", "" ], [ "Xiao", "Li", "" ], [ "Xu", "Xinmeng", "" ] ]
Neural speech coding is a rapidly developing topic, where state-of-the-art approaches now exhibit superior compression performance than conventional methods. Despite significant progress, existing methods still have limitations in preserving and reconstructing fine details for optimal reconstruction, especially at low bitrates. In this study, we introduce SuperCodec, a neural speech codec that achieves state-of-the-art performance at low bitrates. It employs a novel back projection method with selective feature fusion for augmented representation. Specifically, we propose to use Selective Up-sampling Back Projection (SUBP) and Selective Down-sampling Back Projection (SDBP) modules to replace the standard up- and down-sampling layers at the encoder and decoder, respectively. Experimental results show that our method outperforms the existing neural speech codecs operating at various bitrates. Specifically, our proposed method can achieve higher quality reconstructed speech at 1 kbps than Lyra V2 at 3.2 kbps and Encodec at 6 kbps.
2010.14503
Jean De Dieu Mutangana
Jean de Dieu Mutangana, Ravi Tandon
Topological Interference Management with Confidential Messages
Accepted and published in IEEE Transactions on Information Theory
null
null
null
cs.IT cs.CR math.IT
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
The topological interference management (TIM) problem refers to the study of the K-user partially connected interference networks with no channel state information at the transmitters (CSIT), except for the knowledge of network topology. In this paper, we study the TIM problem with confidential messages (TIM-CM), where message confidentiality must be satisfied in addition to reliability constraints. In particular, each transmitted message must be decodable at its intended receiver and remain confidential at the remaining (K-1) receivers. Our main contribution is to present a comprehensive set of results for the TIM-CM problem by studying the symmetric secure degrees of freedom (SDoF). To this end, we first characterize necessary and sufficient conditions for feasibility of positive symmetric SDoF for any arbitrary topology. We next present two achievable schemes for the TIM-CM problem: For the first scheme, we use the concept of secure partition and, for the second one, we use the concept of secure independent sets. We also present outer bounds on symmetric SDoF for any arbitrary network topology. Using these bounds, we characterize the optimal symmetric SDoF of all K=2-user and K=3-user network topologies.
[ { "created": "Tue, 27 Oct 2020 17:59:07 GMT", "version": "v1" }, { "created": "Mon, 30 Jan 2023 14:26:58 GMT", "version": "v2" } ]
2023-01-31
[ [ "Mutangana", "Jean de Dieu", "" ], [ "Tandon", "Ravi", "" ] ]
The topological interference management (TIM) problem refers to the study of the K-user partially connected interference networks with no channel state information at the transmitters (CSIT), except for the knowledge of network topology. In this paper, we study the TIM problem with confidential messages (TIM-CM), where message confidentiality must be satisfied in addition to reliability constraints. In particular, each transmitted message must be decodable at its intended receiver and remain confidential at the remaining (K-1) receivers. Our main contribution is to present a comprehensive set of results for the TIM-CM problem by studying the symmetric secure degrees of freedom (SDoF). To this end, we first characterize necessary and sufficient conditions for feasibility of positive symmetric SDoF for any arbitrary topology. We next present two achievable schemes for the TIM-CM problem: For the first scheme, we use the concept of secure partition and, for the second one, we use the concept of secure independent sets. We also present outer bounds on symmetric SDoF for any arbitrary network topology. Using these bounds, we characterize the optimal symmetric SDoF of all K=2-user and K=3-user network topologies.
2305.13469
Saurabh Srivastava
Saurabh Srivastava, Gaurav Singh, Shou Matsumoto, Ali Raz, Paulo Costa, Joshua Poore, Ziyu Yao
MAILEX: Email Event and Argument Extraction
Accepted at EMNLP 2023
null
null
null
cs.CL cs.AI
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
In this work, we present the first dataset, MailEx, for performing event extraction from conversational email threads. To this end, we first proposed a new taxonomy covering 10 event types and 76 arguments in the email domain. Our final dataset includes 1.5K email threads and ~4K emails, which are annotated with totally ~8K event instances. To understand the task challenges, we conducted a series of experiments comparing three types of approaches, i.e., fine-tuned sequence labeling, fine-tuned generative extraction, and few-shot in-context learning. Our results showed that the task of email event extraction is far from being addressed, due to challenges lying in, e.g., extracting non-continuous, shared trigger spans, extracting non-named entity arguments, and modeling the email conversational history. Our work thus suggests more future investigations in this domain-specific event extraction task.
[ { "created": "Mon, 22 May 2023 20:28:23 GMT", "version": "v1" }, { "created": "Sat, 21 Oct 2023 02:15:22 GMT", "version": "v2" } ]
2023-10-24
[ [ "Srivastava", "Saurabh", "" ], [ "Singh", "Gaurav", "" ], [ "Matsumoto", "Shou", "" ], [ "Raz", "Ali", "" ], [ "Costa", "Paulo", "" ], [ "Poore", "Joshua", "" ], [ "Yao", "Ziyu", "" ] ]
In this work, we present the first dataset, MailEx, for performing event extraction from conversational email threads. To this end, we first proposed a new taxonomy covering 10 event types and 76 arguments in the email domain. Our final dataset includes 1.5K email threads and ~4K emails, which are annotated with totally ~8K event instances. To understand the task challenges, we conducted a series of experiments comparing three types of approaches, i.e., fine-tuned sequence labeling, fine-tuned generative extraction, and few-shot in-context learning. Our results showed that the task of email event extraction is far from being addressed, due to challenges lying in, e.g., extracting non-continuous, shared trigger spans, extracting non-named entity arguments, and modeling the email conversational history. Our work thus suggests more future investigations in this domain-specific event extraction task.
1711.03473
Michel Melo Silva
Michel Melo Silva, Washington Luis Souza Ramos, Felipe Cadar Chamone, Jo\~ao Pedro Klock Ferreira, Mario Fernando Montenegro Campos, Erickson Rangel Nascimento
Making a long story short: A Multi-Importance fast-forwarding egocentric videos with the emphasis on relevant objects
Accepted to publication in the Journal of Visual Communication and Image Representation (JVCI) 2018. Project website: https://www.verlab.dcc.ufmg.br/semantic-hyperlapse
null
10.1016/j.jvcir.2018.02.013
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
The emergence of low-cost high-quality personal wearable cameras combined with the increasing storage capacity of video-sharing websites have evoked a growing interest in first-person videos, since most videos are composed of long-running unedited streams which are usually tedious and unpleasant to watch. State-of-the-art semantic fast-forward methods currently face the challenge of providing an adequate balance between smoothness in visual flow and the emphasis on the relevant parts. In this work, we present the Multi-Importance Fast-Forward (MIFF), a fully automatic methodology to fast-forward egocentric videos facing these challenges. The dilemma of defining what is the semantic information of a video is addressed by a learning process based on the preferences of the user. Results show that the proposed method keeps over $3$ times more semantic content than the state-of-the-art fast-forward. Finally, we discuss the need of a particular video stabilization technique for fast-forward egocentric videos.
[ { "created": "Thu, 9 Nov 2017 17:03:29 GMT", "version": "v1" }, { "created": "Thu, 1 Mar 2018 15:56:26 GMT", "version": "v2" }, { "created": "Wed, 7 Mar 2018 17:59:11 GMT", "version": "v3" } ]
2018-03-08
[ [ "Silva", "Michel Melo", "" ], [ "Ramos", "Washington Luis Souza", "" ], [ "Chamone", "Felipe Cadar", "" ], [ "Ferreira", "João Pedro Klock", "" ], [ "Campos", "Mario Fernando Montenegro", "" ], [ "Nascimento", "Erickson Rangel", "" ] ]
The emergence of low-cost high-quality personal wearable cameras combined with the increasing storage capacity of video-sharing websites have evoked a growing interest in first-person videos, since most videos are composed of long-running unedited streams which are usually tedious and unpleasant to watch. State-of-the-art semantic fast-forward methods currently face the challenge of providing an adequate balance between smoothness in visual flow and the emphasis on the relevant parts. In this work, we present the Multi-Importance Fast-Forward (MIFF), a fully automatic methodology to fast-forward egocentric videos facing these challenges. The dilemma of defining what is the semantic information of a video is addressed by a learning process based on the preferences of the user. Results show that the proposed method keeps over $3$ times more semantic content than the state-of-the-art fast-forward. Finally, we discuss the need of a particular video stabilization technique for fast-forward egocentric videos.
2002.05973
Dongfang Zhao
Dongfang Zhao
Algebraic Structure of Blockchains: A Group-Theoretical Primer
null
null
null
null
cs.DC
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Although recent advances of blockchain systems, notably in the form of cryptocurrency, have drawn tremendous interests from both researchers and practitioners, limited studies existed toward the theoretical foundation of blockchains. This paper presents the first study on the algebraic structure of blockchains with an emphasis on the internal properties under algebraic groups. We axiomatically construct a blockchain group and derive some interesting properties that can be potentially taken into the design space and parametric analysis of real-world blockchain systems.
[ { "created": "Fri, 14 Feb 2020 11:28:59 GMT", "version": "v1" } ]
2020-02-17
[ [ "Zhao", "Dongfang", "" ] ]
Although recent advances of blockchain systems, notably in the form of cryptocurrency, have drawn tremendous interests from both researchers and practitioners, limited studies existed toward the theoretical foundation of blockchains. This paper presents the first study on the algebraic structure of blockchains with an emphasis on the internal properties under algebraic groups. We axiomatically construct a blockchain group and derive some interesting properties that can be potentially taken into the design space and parametric analysis of real-world blockchain systems.
2210.00058
Gino Chacon
Gino A. Chacon, Charles Williams, Johann Knechtel, Ozgur Sinanoglu, Paul V. Gratz
Hardware Trojan Threats to Cache Coherence in Modern 2.5D Chiplet Systems
null
null
null
null
cs.CR cs.AR
http://creativecommons.org/licenses/by/4.0/
As industry moves toward chiplet-based designs, the insertion of hardware Trojans poses a significant threat to the security of these systems. These systems rely heavily on cache coherence for coherent data communication, making coherence an attractive target. Critically, unlike prior work, which focuses only on malicious packet modifications, a Trojan attack that exploits coherence can modify data in memory that was never touched and is not owned by the chiplet which contains the Trojan. Further, the Trojan need not even be physically between the victim and the memory controller to attack the victim's memory transactions. Here, we explore the fundamental attack vectors possible in chiplet-based systems and provide an example Trojan implementation capable of directly modifying victim data in memory. This work aims to highlight the need for developing mechanisms that can protect and secure the coherence scheme from these forms of attacks.
[ { "created": "Fri, 30 Sep 2022 19:45:04 GMT", "version": "v1" } ]
2022-10-04
[ [ "Chacon", "Gino A.", "" ], [ "Williams", "Charles", "" ], [ "Knechtel", "Johann", "" ], [ "Sinanoglu", "Ozgur", "" ], [ "Gratz", "Paul V.", "" ] ]
As industry moves toward chiplet-based designs, the insertion of hardware Trojans poses a significant threat to the security of these systems. These systems rely heavily on cache coherence for coherent data communication, making coherence an attractive target. Critically, unlike prior work, which focuses only on malicious packet modifications, a Trojan attack that exploits coherence can modify data in memory that was never touched and is not owned by the chiplet which contains the Trojan. Further, the Trojan need not even be physically between the victim and the memory controller to attack the victim's memory transactions. Here, we explore the fundamental attack vectors possible in chiplet-based systems and provide an example Trojan implementation capable of directly modifying victim data in memory. This work aims to highlight the need for developing mechanisms that can protect and secure the coherence scheme from these forms of attacks.
1803.05069
Maofan Yin
Maofan Yin, Dahlia Malkhi, Michael K. Reiter, Guy Golan Gueta, Ittai Abraham
HotStuff: BFT Consensus in the Lens of Blockchain
a shorter version of this paper has been published in PODC'19, which does not include interpretation of other protocols using the framework, system evaluation or additional proofs in appendices
null
null
null
cs.DC
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
We present HotStuff, a leader-based Byzantine fault-tolerant replication protocol for the partially synchronous model. Once network communication becomes synchronous, HotStuff enables a correct leader to drive the protocol to consensus at the pace of actual (vs. maximum) network delay--a property called responsiveness--and with communication complexity that is linear in the number of replicas. To our knowledge, HotStuff is the first partially synchronous BFT replication protocol exhibiting these combined properties. HotStuff is built around a novel framework that forms a bridge between classical BFT foundations and blockchains. It allows the expression of other known protocols (DLS, PBFT, Tendermint, Casper), and ours, in a common framework. Our deployment of HotStuff over a network with over 100 replicas achieves throughput and latency comparable to that of BFT-SMaRt, while enjoying linear communication footprint during leader failover (vs. quadratic with BFT-SMaRt).
[ { "created": "Tue, 13 Mar 2018 23:01:05 GMT", "version": "v1" }, { "created": "Thu, 18 Oct 2018 15:39:12 GMT", "version": "v2" }, { "created": "Mon, 18 Mar 2019 18:21:08 GMT", "version": "v3" }, { "created": "Tue, 2 Apr 2019 00:48:38 GMT", "version": "v4" }, { "created": "Wed, 5 Jun 2019 04:26:20 GMT", "version": "v5" }, { "created": "Tue, 23 Jul 2019 05:19:36 GMT", "version": "v6" } ]
2019-07-24
[ [ "Yin", "Maofan", "" ], [ "Malkhi", "Dahlia", "" ], [ "Reiter", "Michael K.", "" ], [ "Gueta", "Guy Golan", "" ], [ "Abraham", "Ittai", "" ] ]
We present HotStuff, a leader-based Byzantine fault-tolerant replication protocol for the partially synchronous model. Once network communication becomes synchronous, HotStuff enables a correct leader to drive the protocol to consensus at the pace of actual (vs. maximum) network delay--a property called responsiveness--and with communication complexity that is linear in the number of replicas. To our knowledge, HotStuff is the first partially synchronous BFT replication protocol exhibiting these combined properties. HotStuff is built around a novel framework that forms a bridge between classical BFT foundations and blockchains. It allows the expression of other known protocols (DLS, PBFT, Tendermint, Casper), and ours, in a common framework. Our deployment of HotStuff over a network with over 100 replicas achieves throughput and latency comparable to that of BFT-SMaRt, while enjoying linear communication footprint during leader failover (vs. quadratic with BFT-SMaRt).