2023 57th Annual Conference on Information Sciences and Systems (CISS)最新文献_第10页

Personalized Decentralized Multi-Task Learning Over Dynamic Communication Graphs 基于动态通信图的个性化分散多任务学习

2023 57th Annual Conference on Information Sciences and Systems (CISS) Pub Date : 2022-12-21 DOI: 10.1109/CISS56502.2023.10089777

Matin Mortaheb, S. Ulukus

{"title":"Personalized Decentralized Multi-Task Learning Over Dynamic Communication Graphs","authors":"Matin Mortaheb, S. Ulukus","doi":"10.1109/CISS56502.2023.10089777","DOIUrl":"https://doi.org/10.1109/CISS56502.2023.10089777","url":null,"abstract":"Decentralized and federated learning algorithms face data heterogeneity as one of the biggest challenges, especially when users want to learn a specific task. Even when personalized headers are used concatenated to a shared network (PF-MTL), aggregating all the networks with a decentralized algorithm can result in performance degradation as a result of heterogeneity in the data. Our algorithm uses exchanged gradients to calculate the correlations among tasks automatically, and dynamically adjusts the communication graph to connect mutually beneficial tasks and isolate those that may negatively impact each other. This algorithm improves the learning performance and leads to faster convergence compared to the case where all clients are connected to each other regardless of their correlations. We conduct experiments on a synthetic Gaussian dataset and a large-scale celebrity attributes (CelebA) dataset. The experiment with the synthetic data illustrates that our proposed method is capable of detecting tasks that are positively and negatively correlated. Moreover, the results of the experiments with CelebA demonstrate that the proposed method may produce significantly faster training results than fully-connected networks.","PeriodicalId":243775,"journal":{"name":"2023 57th Annual Conference on Information Sciences and Systems (CISS)","volume":"46 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129400595","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 1

Performance Analysis of LOS THz Systems Under Misalignment and Deterministic Fading 失对准和确定性衰落下LOS THz系统性能分析

2023 57th Annual Conference on Information Sciences and Systems (CISS) Pub Date : 2022-12-21 DOI: 10.13140/RG.2.2.10972.67202/1

Rayyan Abdalla, A. Cooper

引用次数: 0

Vulnerabilities of Deep Learning-Driven Semantic Communications to Backdoor (Trojan) Attacks 深度学习驱动语义通信对后门(木马)攻击的漏洞分析

2023 57th Annual Conference on Information Sciences and Systems (CISS) Pub Date : 2022-12-21 DOI: 10.1109/CISS56502.2023.10089692

Y. Sagduyu, T. Erpek, S. Ulukus, A. Yener

{"title":"Vulnerabilities of Deep Learning-Driven Semantic Communications to Backdoor (Trojan) Attacks","authors":"Y. Sagduyu, T. Erpek, S. Ulukus, A. Yener","doi":"10.1109/CISS56502.2023.10089692","DOIUrl":"https://doi.org/10.1109/CISS56502.2023.10089692","url":null,"abstract":"This paper highlights vulnerabilities of deep learning-driven semantic communications to backdoor (Trojan) attacks. Semantic communications aims to convey a desired meaning while transferring information from a transmitter to its receiver. The encoder-decoder pair of an autoencoder that is represented by deep neural networks (DNNs) is trained to reconstruct signals such as images at the receiver by transmitting latent features of small size over a limited number of channel uses. In the meantime, the DNN of a semantic task classifier at the receiver is jointly trained with the autoencoder to check the meaning conveyed to the receiver. The complex decision space of the DNNs makes semantic communications susceptible to adversarial manipulations. In a backdoor (Trojan) attack, the adversary adds triggers to a small portion of training samples and changes the label to a target label. When the transfer of images is considered, the triggers can be added to the images or equivalently to the corresponding transmitted or received signals. In test time, the adversary activates these triggers by providing poisoned samples as input to the encoder (or decoder) of semantic communications. The backdoor attack can effectively change the semantic information transferred for the poisoned input samples to a target meaning. As the performance of semantic communications improves with the signal-to-noise ratio and the number of channel uses, the success of the backdoor attack increases as well. Also, increasing the Trojan ratio in training data makes the attack more successful. On the other hand, the attack is selective and its effect on the unpoisoned input samples remains small. Overall, this paper shows that the backdoor attack poses a serious threat to semantic communications and presents novel design guidelines to preserve the meaning of transferred information in the presence of backdoor attacks.","PeriodicalId":243775,"journal":{"name":"2023 57th Annual Conference on Information Sciences and Systems (CISS)","volume":"60 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131712231","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 4

Strategic multi-task coordination over regular networks of robots with limited computation and communication capabilities 具有有限计算和通信能力的常规机器人网络中的战略性多任务协调

2023 57th Annual Conference on Information Sciences and Systems (CISS) Pub Date : 2022-12-21 DOI: 10.1109/CISS56502.2023.10089772

Yi Wei, M. Vasconcelos

引用次数: 2

Generalized Simultaneous Perturbation Stochastic Approximation with Reduced Estimator Bias 估计量偏差减小的广义同时摄动随机逼近

2023 57th Annual Conference on Information Sciences and Systems (CISS) Pub Date : 2022-12-20 DOI: 10.1109/CISS56502.2023.10089720

S. Bhatnagar, Prashanth L.A.

引用次数: 1

Physical layer insecurity 物理层不安全

2023 57th Annual Conference on Information Sciences and Systems (CISS) Pub Date : 2022-12-02 DOI: 10.1109/CISS56502.2023.10089749

Muriel M'edard, K. Duffy

{"title":"Physical layer insecurity","authors":"Muriel M'edard, K. Duffy","doi":"10.1109/CISS56502.2023.10089749","DOIUrl":"https://doi.org/10.1109/CISS56502.2023.10089749","url":null,"abstract":"In the classic wiretap model, Alice wishes to reliably communicate to Bob without being overheard by Eve who is eavesdropping over a degraded channel. Systems for achieving that physical layer security often rely on an error correction code whose rate is below the Shannon capacity of Alice and Bob's channel, so Bob can reliably decode, but above Alice and Eve's, so Eve cannot reliably decode. For the finite block length regime, several metrics have been proposed to characterise information leakage. Here we reassess a metric, the success exponent, and demonstrate it can be operationalized through the use of Guessing Random Additive Noise Decoding (GRAND) to compromise the physical-layer security of any moderate length code. Success exponents are the natural beyond-capacity analogue of error exponents that characterise the probability that a maximum likelihood decoding is correct when the code-rate is above Shannon capacity, which is exponentially decaying in the code-length. In the finite blocklength regime, success exponents can be used to approximately evaluate the frequency with which Eve's decoding is correct in beyond-capacity channel conditions. Through the use of GRAND, we demonstrate that Eve can constrain her decoding procedure through a query-number threshold so that when she does identify a decoding, it is correct with high likelihood, significantly compromising Alice and Bob's communication by truthfully revealing a proportion of it. We provide general mathematical expressions for the determination of success exponents in channels that can have temporally correlated noise as well as for the evaluation of Eve's query number threshold, using the binary symmetric channel as a worked example. As GRAND algorithms are code-book agnostic and can decode any code structure, we provide empirical results for Random Linear Codes as exemplars. Simulation results mimic the mathematical predictions, demonstrating the practical possibility of compromising physical layer security.","PeriodicalId":243775,"journal":{"name":"2023 57th Annual Conference on Information Sciences and Systems (CISS)","volume":"31 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122290632","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 4

Rebalancing Techniques for Asynchronously Distributed EEG Data to Improve Automatic Seizure Type Classification 异步分布脑电图数据再平衡技术改进癫痫类型自动分类

2023 57th Annual Conference on Information Sciences and Systems (CISS) Pub Date : 2022-11-21 DOI: 10.1109/CISS56502.2023.10089669

N. McCallan, S. Davidson, K. Y. Ng, P. Biglarbeigi, D. Finlay, B. Lan, J. Mclaughlin

{"title":"Rebalancing Techniques for Asynchronously Distributed EEG Data to Improve Automatic Seizure Type Classification","authors":"N. McCallan, S. Davidson, K. Y. Ng, P. Biglarbeigi, D. Finlay, B. Lan, J. Mclaughlin","doi":"10.1109/CISS56502.2023.10089669","DOIUrl":"https://doi.org/10.1109/CISS56502.2023.10089669","url":null,"abstract":"Epilepsy, a nervous system disorder, is charac-terised by unprovoked, unpredictable, and recurrent seizures. To diagnose epileptic seizures, electroencephalography (EEG) is frequently used in medical settings. Effective automated detection and classification strategies are needed because visual analysis and interpretation of EEG signals consume time and call for specialised expertise. The main objective of this paper is to examine the effectiveness of multiple rebalancing techniques to address the problem of asynchronously distributed data, specifically employing random resampling, synthetic minority oversampling technique (SMOTE), and adaptive synthetic sampling approach for imbalanced learning (ADASYN), for seizure type classification. The model utilises both frequency information using variational mode decomposition (VMD), and phase information by extracting the phase locking value (PLV) across 19 common EEG channels found in the Temple University Hospital EEG Seizure Corpus (TUSZ) v1.5.2 dataset. The random subspace k-nearest neighbour (RSkNN) ensemble classifier is used for seizure type classification of five classes - complex partial seizures (CPSZ), simple partial seizures (SPSZ), absence seizures (ABSZ), tonic clonic seizures (TCSZ), and tonic seizures (TNSZ) - to determine the performance of each rebalancing techniques, with the highest accuracy and weighted F1 score of 96.28% and 0.964, respectively using SMOTE with two nearest neighbours.","PeriodicalId":243775,"journal":{"name":"2023 57th Annual Conference on Information Sciences and Systems (CISS)","volume":"49 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-11-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114205905","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Tunable complexity benchmarks for evaluating physics-informed neural networks on coupled ordinary differential equations 在耦合常微分方程上评估物理信息神经网络的可调复杂性基准

2023 57th Annual Conference on Information Sciences and Systems (CISS) Pub Date : 2022-10-14 DOI: 10.1109/CISS56502.2023.10089728

Alexander New, B. Eng, A. Timm, A. Gearhart

引用次数: 1

Policy Gradients for Probabilistic Constrained Reinforcement Learning 概率约束强化学习的策略梯度

2023 57th Annual Conference on Information Sciences and Systems (CISS) Pub Date : 2022-10-02 DOI: 10.1109/CISS56502.2023.10089763

Weiqin Chen, D. Subramanian, Santiago Paternain

引用次数: 2

Multi-armed Bandit Learning on a Graph 图上的多臂强盗学习

2023 57th Annual Conference on Information Sciences and Systems (CISS) Pub Date : 2022-09-20 DOI: 10.1109/CISS56502.2023.10089744

Tianpeng Zhang, Kasper Johansson, Na Li

{"title":"Multi-armed Bandit Learning on a Graph","authors":"Tianpeng Zhang, Kasper Johansson, Na Li","doi":"10.1109/CISS56502.2023.10089744","DOIUrl":"https://doi.org/10.1109/CISS56502.2023.10089744","url":null,"abstract":"The multi-armed bandit(MAB) problem is a simple yet powerful framework that has been extensively studied in the context of decision-making under uncertainty. In many real-world applications, such as robotic applications, selecting an arm corresponds to a physical action that constrains the choices of the next available arms (actions). Motivated by this, we study an extension of MAB called the graph bandit, where an agent travels over a graph to maximize the reward collected from different nodes. The graph defines the agent's freedom in selecting the next available nodes at each step. We assume the graph structure is fully available, but the reward distributions are unknown. Built upon an offline graph-based planning algorithm and the principle of optimism, we design a learning algorithm, G-UCB, that balances long-term exploration-exploitation using the principle of optimism. We show that our proposed algorithm achieves ${O}(sqrt{vert Svert Tlog(T)}+Dvert Svert log T)$ learning regret, where $vert Svert$ is the number of nodes and $D$ is the diameter of the graph, which matches the theoretical lower bound $Omega(sqrt{vert Svert T})$ up to logarithmic factors. To our knowledge, this result is among the first tight regret bounds in non-episodic, un-discounted learning problems with known deterministic transitions. Numerical experiments confirm that our algorithm outperforms several benchmarks.","PeriodicalId":243775,"journal":{"name":"2023 57th Annual Conference on Information Sciences and Systems (CISS)","volume":"30 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131832268","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0