IEEE journal on selected areas in information theory最新文献_第2页

Foundations of Efficient Model Reconstruction Using Counterfactual Explanations 利用反事实解释进行有效模型重建的基础

IF 2.3

IEEE journal on selected areas in information theory Pub Date : 2026-01-01 Epub Date: 2026-03-22 DOI: 10.1109/JSAIT.2026.3695917

Pasan Dissanayake;Sanghamitra Dutta

{"title":"Foundations of Efficient Model Reconstruction Using Counterfactual Explanations","authors":"Pasan Dissanayake;Sanghamitra Dutta","doi":"10.1109/JSAIT.2026.3695917","DOIUrl":"https://doi.org/10.1109/JSAIT.2026.3695917","url":null,"abstract":"With the ever-growing computational demands of AI, extracting richer information from fewer data points is of utmost importance. This work proposes data-efficient model reconstruction and compression by leveraging an explainability technique called counterfactual explanations. Counterfactual explanations are traditionally used to explain black-box machine learning models by providing ways of achieving a favorable model outcome with minimum input perturbation (essentially, the closest point on the other side of the decision boundary). Interestingly, such counterfactuals can also be used to reconstruct a model by training a surrogate model on them to give similar predictions as the original model. In this work, we analyze how counterfactual explanations can improve model reconstruction by specifically utilizing the fact that the counterfactuals lie quite close to the decision boundary. Our main contribution is to derive novel theoretical relationships between the error in model reconstruction and the number of counterfactual queries required by leveraging polytope theory. We also derive additional bounds considering multiple model families such as ReLU networks and Lipschitz-continuous models. Our results establish that counterfactual explanations can act as information-rich probes for model reconstruction, which has significant implications for energy-efficiency, sustainability, and also privacy.","PeriodicalId":73295,"journal":{"name":"IEEE journal on selected areas in information theory","volume":"7 ","pages":"302-312"},"PeriodicalIF":2.3,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148481020","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

Lossless Compression via Chained Lightweight Neural Predictors With Information Inheritance 基于信息继承的链式轻量级神经预测器的无损压缩

IF 2.3

IEEE journal on selected areas in information theory Pub Date : 2026-01-01 Epub Date: 2026-08-03 DOI: 10.1109/JSAIT.2026.3719186

Yuriy Kim;Evgeny Belyaev

{"title":"Lossless Compression via Chained Lightweight Neural Predictors With Information Inheritance","authors":"Yuriy Kim;Evgeny Belyaev","doi":"10.1109/JSAIT.2026.3719186","DOIUrl":"https://doi.org/10.1109/JSAIT.2026.3719186","url":null,"abstract":"This paper is dedicated to lossless data compression with probability estimation using neural networks. First, we propose a probability estimation architecture based on a chain of neural predictors, so that each unit of the chain is defined as a neural network with the minimum number of weights, which is sufficient for efficient compression of data generated by Markov sources of a given order. We show that this architecture allows us to minimize the overall number of weights participating in the probability estimation process depending on the statistical properties of the input data. Second, in order to improve compression efficiency, we introduce an information inheritance mechanism, where the probability estimate obtained by a low-order unit is used at the next higher-order unit. Experimental results show that the proposed lossless data compressor, equipped with the chained probability estimation architecture, provides compression ratios close to the state-of-the-art EDPC compressor. At the same time, it outperforms EDPC by a factor of 1.2 to 5.2 in encoding throughput and by a factor of 2.8 to 11.6 in decoding throughput on a consumer GPU.","PeriodicalId":73295,"journal":{"name":"IEEE journal on selected areas in information theory","volume":"7 ","pages":"384-395"},"PeriodicalIF":2.3,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148746946","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

Secure Semantic Communication Over Wiretap Channels: Rate-Distortion-Equivocation Tradeoff 窃听信道上的安全语义通信：速率失真-歧义权衡

IF 2.2

IEEE journal on selected areas in information theory Pub Date : 2026-01-01 Epub Date: 2026-03-23 DOI: 10.1109/JSAIT.2026.3676713

Denis Kozlov;Mahtab Mirmohseni;Rahim Tafazolli

{"title":"Secure Semantic Communication Over Wiretap Channels: Rate-Distortion-Equivocation Tradeoff","authors":"Denis Kozlov;Mahtab Mirmohseni;Rahim Tafazolli","doi":"10.1109/JSAIT.2026.3676713","DOIUrl":"https://doi.org/10.1109/JSAIT.2026.3676713","url":null,"abstract":"This paper investigates an information-theoretic model of secure semantic-aware communication. For this purpose, we consider the lossy joint source-channel coding (JSCC) of a memoryless semantic source transmitted over a memoryless wiretap channel. The source consists of two correlated parts that represent semantic and observed aspects of the information. Our model assumes separate fidelity and secrecy constraints on each source component and, in addition, encompasses two cases for the source output, in order to evaluate the performance gains if the encoder has an extended access to the source. Specifically, in Case 1, the encoder has direct access only to the samples from a single (observed) source component, while in Case 2 it has additional direct access to the samples of the underlaying semantic information. We derive single-letter converse and achievability bounds on the rate-distortion-equivocation region. The converse bound explicitly contains rate-distortion functions, making it easy to evaluate, especially for some common distributions. The proposed achievability coding scheme involves novel stochastic superposition coding with two private parts to enable analysis of the equivocation for each source component, separately. Our results generalise some of the previously established source and source-channel coding problems. The general results are further specialised to Gaussian and Bernoulli sources transmitted over Gaussian and binary wiretap channels, respectively. The numerical evaluations illustrate the derived bounds for these distributions.","PeriodicalId":73295,"journal":{"name":"IEEE journal on selected areas in information theory","volume":"7 ","pages":"120-133"},"PeriodicalIF":2.2,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147796120","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

Degrees of Freedom of Cache-Aided Interference Channels Assisted by Active Intelligent Reflecting Surfaces 主动智能反射面辅助的缓存辅助干涉通道的自由度

IF 2.2

IEEE journal on selected areas in information theory Pub Date : 2026-01-01 Epub Date: 2026-03-27 DOI: 10.1109/JSAIT.2026.3678472

Abolfazl Changizi;Ali H. Abdollahi Bafghi;Mahtab Mirmohseni;Masoumeh Nasiri-Kenari

{"title":"Degrees of Freedom of Cache-Aided Interference Channels Assisted by Active Intelligent Reflecting Surfaces","authors":"Abolfazl Changizi;Ali H. Abdollahi Bafghi;Mahtab Mirmohseni;Masoumeh Nasiri-Kenari","doi":"10.1109/JSAIT.2026.3678472","DOIUrl":"https://doi.org/10.1109/JSAIT.2026.3678472","url":null,"abstract":"This paper studies cache-aided wireless networks in the presence of active intelligent reflecting surfaces (IRSs) from an information-theoretic perspective. Specifically, we investigate interference management in a cache-aided wireless network assisted by an active IRS to enhance the achievable degrees of freedom (DoF). To this end, we jointly design the content placement, delivery phase, and IRS coefficients, and propose a one-shot achievability scheme. Our scheme exploits transmitters’ cooperation, cache contents, interference alignment, and IRS capabilities, based on the network parameters. We derive the achievable one-shot sum-DoF for different cache sizes, network configurations, and numbers of IRS elements, followed by an upper bound. Our results highlight the potential of deploying an IRS in cache-aided wireless communication systems. In particular, they underscore the enhancement of achievable DoF for various parameter regimes, especially when cache sizes are inadequate. Notably, we show that access to an IRS with a sufficient number of elements enables the achievement of the maximum possible DoF for various parameter regimes of interest.","PeriodicalId":73295,"journal":{"name":"IEEE journal on selected areas in information theory","volume":"7 ","pages":"192-204"},"PeriodicalIF":2.2,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148003500","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

Efficient Near-Field Channel Feature Estimation: Criteria and Algorithms 有效的近场信道特征估计：准则和算法

IF 2.2

IEEE journal on selected areas in information theory Pub Date : 2026-01-01 Epub Date: 2025-12-31 DOI: 10.1109/JSAIT.2025.3649807

Hyeonjin Chung;Hanvit Kim;Sunwoo Kim;Andrea Conti;Moe Z. Win

{"title":"Efficient Near-Field Channel Feature Estimation: Criteria and Algorithms","authors":"Hyeonjin Chung;Hanvit Kim;Sunwoo Kim;Andrea Conti;Moe Z. Win","doi":"10.1109/JSAIT.2025.3649807","DOIUrl":"https://doi.org/10.1109/JSAIT.2025.3649807","url":null,"abstract":"When the radio signal propagates in the near-field (NF), the distance can be estimated by the curvature of the wavefront. This brings new opportunities for sensing and communication, but a computational burden as well. This paper first establishes a theoretical foundation for computationally efficient NF feature (i.e., angle-of-arrival (AOA) and distance) estimation. We reveal that the computation of feature estimates can be reduced without sacrificing the estimation accuracy, if the distance between the receiver and transmitter (or scatterer) exceeds approximately one quarter of the Rayleigh distance. Based on this criterion, we introduce a zone-wise search-based algorithm that reduces computation by restricting the high-dimensional (HD) search to the angle-distance search space where the criterion is not met. In the other search space, sequential low-dimensional (LD) searches are carried out. Along with the algorithm, the estimation-theoretic analysis for NF feature estimation is provided to present the theoretical performance limit. The simulation results demonstrate that even under severe NF effect, the accuracy of the proposed algorithm reaches that of the HD search-based algorithm, which represents a realistically achievable upper bound. Nevertheless, the computation complexity of the proposed algorithm is significantly lower than the HD search-based algorithm.","PeriodicalId":73295,"journal":{"name":"IEEE journal on selected areas in information theory","volume":"7 ","pages":"75-90"},"PeriodicalIF":2.2,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147665231","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

Fundamental Limits of Learning Under Erasure-Constrained Communication Channels 在擦除约束的通信渠道下学习的基本限制

IF 2.2

IEEE journal on selected areas in information theory Pub Date : 2026-01-01 Epub Date: 2026-03-25 DOI: 10.1109/JSAIT.2026.3677413

Merve Karakas;Osama Hanna;Lin F. Yang;Christina Fragouli

引用次数: 0

A Spatio-Temporal Model for Information Freshness in Massive Random Access 海量随机访问下信息新鲜度的时空模型

IF 2.3

IEEE journal on selected areas in information theory Pub Date : 2026-01-01 Epub Date: 2026-07-17 DOI: 10.1109/JSAIT.2026.3714595

Andrea Munari;Alessandro Buratto;Federico Chiariotti;Leonardo Badia;Petar Popovski

{"title":"A Spatio-Temporal Model for Information Freshness in Massive Random Access","authors":"Andrea Munari;Alessandro Buratto;Federico Chiariotti;Leonardo Badia;Petar Popovski","doi":"10.1109/JSAIT.2026.3714595","DOIUrl":"https://doi.org/10.1109/JSAIT.2026.3714595","url":null,"abstract":"Massive connectivity, a key building block of 5G, is expected to play an important role in the next generation of wireless systems, and its expected requirements are being revolutionized through the modeling of the information dynamics related to the vast numbers of Internet of things (IoT) devices. Motivated by this, the present paper introduces a model that captures the spatio-temporal nature of freshness of information sent via random access channel policies from an extremely large set of IoT devices via simple scalar parameters, i.e., the probability of success and accuracy of received updates. There are many information freshness metrics, starting from the age of information (AoI), all of which are proxies for the actual application performance, characterized over the temporal dimension. Our model adds the spatial dimension to this picture, observing that sensors distributed over the same area may have a strong correlation, and information from multiple closeby sensors may improve the overall accuracy of the receiver. We focus on characterizing the uncertainty of the receiver, expressed through the conditional entropy, considering a network of partially reliable, spatially distributed sensors observing the same process and reporting their measurements over a slotted ALOHA channel. We consider a simple forgetful receiver and a more complete model which accounts for the full history of past observations, deriving their performance, and optimizing the transmission probability of nodes to minimize the expected uncertainty.","PeriodicalId":73295,"journal":{"name":"IEEE journal on selected areas in information theory","volume":"7 ","pages":"333-353"},"PeriodicalIF":2.3,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11614095","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148626449","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

The Thermodynamic Costs of Simple Linear Regression 简单线性回归的热力学代价

IF 2.3

IEEE journal on selected areas in information theory Pub Date : 2026-01-01 Epub Date: 2026-09-11 DOI: 10.1109/JSAIT.2026.3732971

Samuel H. D’Ambrosia;Sultan M. Daniels;Michael R. DeWeese;Anant Sahai

引用次数: 0

Coherence-Aware Over-the-Air Distributed Learning Under Heterogeneous Link Impairments 异构链路损伤下的相干感知空中分布式学习

IF 2.2

IEEE journal on selected areas in information theory Pub Date : 2026-01-01 Epub Date: 2026-03-11 DOI: 10.1109/JSAIT.2026.3672794

Mehdi Karbalayghareh;David J. Love;Christopher G. Brinton

{"title":"Coherence-Aware Over-the-Air Distributed Learning Under Heterogeneous Link Impairments","authors":"Mehdi Karbalayghareh;David J. Love;Christopher G. Brinton","doi":"10.1109/JSAIT.2026.3672794","DOIUrl":"https://doi.org/10.1109/JSAIT.2026.3672794","url":null,"abstract":"Distributed machine learning (ML) over wireless networks hinges on accurate channel state information (CSI) and efficient exchange of high-dimensional model updates. These demands are governed by channel coherence time and bandwidth, which vary across devices (links) due to heterogeneous mobility and scattering, causing degraded downlink delivery and distorted uplink over-the-air (OTA) aggregation. We propose a coherence-aware federated learning (FL) framework that jointly addresses impairments on downlink and uplink with communication-efficient strategies. In the downlink, we employ product superposition to multiplex global model symbols for long-coherence (static) devices onto the pilot tones required by short-coherence (dynamic) devices for channel estimation, turning pilot overhead into payload while preserving estimation fidelity. In the proposed scheme, an orthogonal frequency-division multiplexing (OFDM) super-block is partitioned into sub-blocks aligned with the smallest coherence time and bandwidth, enabling consistent channel estimation and stabilizing OTA aggregation across heterogeneous devices. Partial model reception at dynamic devices is mitigated via previous local model filling (PLMF), which reuses prior updates. We establish convergence guarantees under heterogeneous link impairments, imperfect CSI, and aggregation noise. The proposed framework enables efficient scheduling under coherence heterogeneity; analysis and experiments demonstrate notable gains in communication efficiency, latency, and learning accuracy over conventional FL baselines.","PeriodicalId":73295,"journal":{"name":"IEEE journal on selected areas in information theory","volume":"7 ","pages":"46-61"},"PeriodicalIF":2.2,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147557852","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

On the Convergence Theory of Pipeline Gradient-Based Analog In-Memory Training 基于管道梯度的模拟内存训练收敛理论研究

IF 2.3

IEEE journal on selected areas in information theory Pub Date : 2026-01-01 Epub Date: 2026-04-29 DOI: 10.1109/JSAIT.2026.3688610

Zhaoxian Wu;Quan Xiao;Tayfun Gokmen;Hsinyu Tsai;Kaoutar El Maghraoui;Tianyi Chen

{"title":"On the Convergence Theory of Pipeline Gradient-Based Analog In-Memory Training","authors":"Zhaoxian Wu;Quan Xiao;Tayfun Gokmen;Hsinyu Tsai;Kaoutar El Maghraoui;Tianyi Chen","doi":"10.1109/JSAIT.2026.3688610","DOIUrl":"https://doi.org/10.1109/JSAIT.2026.3688610","url":null,"abstract":"Aiming to accelerate the training of large deep neural networks (DNN) in an energy-efficient way, analog in-memory computing (AIMC) emerges as a solution with immense potential. AIMC accelerator keeps model weights in memory without moving them from memory to processors during training, reducing overhead dramatically. Despite its efficiency, scaling up AIMC systems presents significant challenges. Since weight copying is expensive and inaccurate, data parallelism is less efficient on AIMC accelerators. It necessitates the exploration of pipeline parallelism, particularly asynchronous pipeline parallelism, which utilizes all available accelerators during the training process. This paper examines the convergence theory of stochastic gradient descent on AIMC hardware with an asynchronous pipeline (<monospace>Analog-SGD-AP</monospace>). Although there is empirical exploration of AIMC accelerators, the theoretical understanding of how analog hardware imperfections in weight updates affect the training of multi-layer DNN models remains underexplored. Furthermore, the asynchronous pipeline parallelism results in stale weights issues, which render the update signals no longer valid gradients. To close the gap, this paper investigates the convergence properties of <monospace>Analog-SGD-AP</monospace> on multi-layer DNN training. We show that the <monospace>Analog-SGD-AP</monospace> converges with iteration complexity <inline-formula> <tex-math>${mathcal {O}}left ({ varepsilon ^{-2}+varepsilon ^{-1}}right)$ </tex-math></inline-formula> despite the aforementioned issues, which matches the complexities of digital SGD and Analog SGD with synchronous pipeline, except the non-dominant term <inline-formula> <tex-math>${mathcal {O}}(varepsilon ^{-1})$ </tex-math></inline-formula>. It implies that AIMC training benefits from asynchronous pipelining almost for free compared with the synchronous pipeline by overlapping computation.","PeriodicalId":73295,"journal":{"name":"IEEE journal on selected areas in information theory","volume":"7 ","pages":"279-289"},"PeriodicalIF":2.3,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148433329","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