arXiv subjects

Wenyi Zhang

Publications and source records attributed to Wenyi Zhang.

At least 55 records · Page 3Linked to original sources

arXiv2025

AB-Cache: Training-Free Acceleration of Diffusion Models via Adams-Bashforth Cached Feature Reuse

Diffusion models have demonstrated remarkable success in generative tasks, yet their iterative denoising process results in slow inference, limiting their practicality. While existing acceleration methods exploit the well-known U-shaped similarity pattern between adjacent steps through caching mechanisms, they lack theoretical foundation and rely on simplistic computation reuse, often leading to performance degradation. In this work, we provide a theoretical understanding by analyzing the denoising process through the second-order Adams-Bashforth method, revealing a linear relationship between the outputs of consecutive steps. This analysis explains why the outputs of adjacent steps exhibit a U-shaped pattern. Furthermore, extending Adams-Bashforth method to higher order, we propose a novel caching-based acceleration approach for diffusion models, instead of directly reusing cached results, with a truncation error bound of only $O(h^k)$ where $h$ is the step size. Extensive validation across diverse image and video diffusion models (including HunyuanVideo and FLUX.1-dev) with various schedulers demonstrates our method's effectiveness in achieving nearly $3\times$ speedup while maintaining original performance levels, offering a practical real-time solution without compromising generation quality.

stat.ML↗

arXiv2025

Data-Driven Neural Estimation of Indirect Rate-Distortion Function

The rate-distortion function (RDF) has long been an information-theoretic benchmark for data compression. As its natural extension, the indirect rate-distortion function (iRDF) corresponds to the scenario where the encoder can only access an observation correlated with the source, rather than the source itself. Such scenario is also relevant for modern applications like remote sensing and goal-oriented communication. The iRDF can be reduced into a standard RDF with the distortion measure replaced by its conditional expectation conditioned upon the observation. This reduction, however, leads to a non-trivial challenge when one needs to estimate the iRDF given datasets only, because without statistical knowledge of the joint probability distribution between the source and its observation, the conditional expectation cannot be evaluated. To tackle this challenge, starting from the well known fact that conditional expectation is the minimum mean-squared error estimator and exploiting a Markovian relationship, we identify a functional equivalence between the reduced distortion measure in the iRDF and the solution of a quadratic loss minimization problem, which can be efficiently approximated by neural network approach. We proceed to reformulate the iRDF as a variational problem corresponding to the Lagrangian representation of the iRDF curve, and propose a neural network based approximate solution, integrating the aforementioned distortion measure estimator. Asymptotic analysis guarantees consistency of the solution, and numerical experimental results demonstrate the accuracy and effectiveness of the algorithm.

cs.IT↗

arXiv2024

Dual-Zone Hard-Core Model for RTS/CTS Handshake Analysis in WLANs

This paper introduces a new stochastic geometry-based model to analyze the Request-to-Send/Clear-to-Send (RTS/CTS) handshake mechanism in wireless local area networks (WLANs). We develop an advanced hard-core point process model, termed the dual-zone hard-core process (DZHCP), which extends traditional hard-core models to capture the spatial interactions and exclusion effects introduced by the RTS/CTS mechanism. This model integrates key parameters accounting for the thinning effects imposed by RTS/CTS, enabling a refined characterization of active transmitters in the network. Analytical expressions are derived for the intensity of the DZHCP, the mean interference, and an approximation of the success probability, providing insight into how network performance depends on critical design parameters. Our results provide better estimates of interference levels and success probability, which could inform strategies for better interference management and improved performance in future WLAN designs.

cs.NI↗

arXiv2024

Geometric constellation shaping for wireless optical intensity channels: An information-theoretic approach

A simple geometric shaping method is proposed for optical wireless communication systems based on intensity modulation and direct detection (IM/DD) from an information-theoretic perspective. Constellations consisting of equiprobable levels with exponential-like distribution are obtained, which possesses asymptotic optimality in the sense that the high-SNR capacity of average-intensity constrained optical intensity channel can be approached by such constellations with increasing size. All $2^b$ levels ($b\in\mathbb N$) of the obtained constellation can be represented by a basic level and $b+2$ bits, thereby reducing the required resolution of the digital-to-analog converter (DAC) without affecting the asymptotic optimality. Achievable information rate evaluations verify the asymptotic optimality. As an example, error performance results of a simple $16$-level LDPC coded modulation scheme show that a shaping gain of $0.65$ dB can be obtained by applying the proposed constellation design. This method can also be applied to more specific IM/DD channel models, since it only requires a near-optimal continuous input distribution.

cs.IT↗

arXiv2024

Information Rates of Channels with Additive White Cauchy Noise

Information transmission over discrete-time channels with memoryless additive noise obeying a Cauchy, rather than Gaussian, distribution, are studied. The channel input satisfies an average power constraint. Upper and lower bounds to such additive white Cauchy noise (AWCN) channel capacity are established. In the high input power regime, the gap between upper and lower bounds is within 0.5 nats per channel use, and the lower bound can be achieved with Gaussian input. In the lower input power regime, the capacity can be asymptotically approached by employing antipodal input. It is shown that the AWCN decoder can be applied to additive white Gaussian noise (AWGN) channels with negligible rate loss, while the AWGN decoder when applied to AWCN channels cannot ensure reliable decoding. For the vector receiver case, it is shown that a linear combining receiver front end loses the channel combining gain, a phenomenon drastically different from AWGN vector channels.

cs.IT↗

arXiv2024

A Rate-Distortion Analysis for Composite Sources Under Subsource-Dependent Fidelity Criteria

A composite source, consisting of multiple subsources and a memoryless switch, outputs one symbol at a time from the subsource selected by the switch. If some data should be encoded more accurately than other data from an information source, the composite source model is suitable because in this model different distortion constraints can be put on the subsources. In this context, we propose subsource-dependent fidelity criteria for composite sources and use them to formulate a rate-distortion problem. We solve the problem and obtain a single-letter expression for the rate-distortion function. Further rate-distortion analysis characterizes the performance of classify-then-compress (CTC) coding, which is frequently used in practice when subsource-dependent fidelity criteria are considered. Our analysis shows that CTC coding generally has performance loss relative to optimal coding, even if the classification is perfect. We also identify the cause of the performance loss, that is, class labels have to be reproduced in CTC coding. Last but not least, we show that the performance loss is negligible for asymptotically small distortion if CTC coding is appropriately designed and some mild conditions are satisfied.

cs.IT↗

arXiv2024

Alternating Maximization Algorithm for Mismatch Capacity with Oblivious Relaying

Reliable communication over a discrete memoryless channel with the help of a relay has aroused interest due to its widespread applications in practical scenarios. By considering the system with a mismatched decoder, previous works have provided optimization models to evaluate the mismatch capacity in these scenarios. The proposed models, however, are difficult due to the complicated structure of the mismatched decoding problem with the information flows in hops given by the relay. Existing methods, such as the grid search, become impractical as they involve finding all roots of a nonlinear system, with the growing size of the alphabet. To address this problem, we reformulate the max-min optimization model as a consistent maximization form, by considering the dual form of the inner minimization problem and the Lagrangian with a fixed multiplier. Based on the proposed formulation, an alternating maximization framework is designed, which provides the closed-form solution with simple iterations in each step by introducing a suitable variable transformation. The effectiveness of the proposed approach is demonstrated by the simulations over practical scenarios, including Quaternary and Gaussian channels. Moreover, the simulation results of the transitional probability also shed light on the promising application attribute to the quantizer design in the relay node.

cs.IT↗

arXiv2024

A Survey of Foundation Models for Music Understanding

Music is essential in daily life, fulfilling emotional and entertainment needs, and connecting us personally, socially, and culturally. A better understanding of music can enhance our emotions, cognitive skills, and cultural connections. The rapid advancement of artificial intelligence (AI) has introduced new ways to analyze music, aiming to replicate human understanding of music and provide related services. While the traditional models focused on audio features and simple tasks, the recent development of large language models (LLMs) and foundation models (FMs), which excel in various fields by integrating semantic information and demonstrating strong reasoning abilities, could capture complex musical features and patterns, integrate music with language and incorporate rich musical, emotional and psychological knowledge. Therefore, they have the potential in handling complex music understanding tasks from a semantic perspective, producing outputs closer to human perception. This work, to our best knowledge, is one of the early reviews of the intersection of AI techniques and music understanding. We investigated, analyzed, and tested recent large-scale music foundation models in respect of their music comprehension abilities. We also discussed their limitations and proposed possible future directions, offering insights for researchers in this field.

cs.SD↗

arXiv2024

An In-Depth Investigation of LEO Satellite Topology Design Parameters

Low Earth Orbit (LEO) satellite networks are rapidly gaining traction today. Although several real-world deployments exist, our preliminary analysis of LEO topology performance with the soon-to-be operational Inter-Satellite Links (ISLs) reveals several interesting characteristics that are difficult to explain based on our current understanding of topologies. For example, a real-world satellite shell with a low density of satellites offers better latency performance than another shell with nearly double the number of satellites. In this work, we conduct an in-depth investigation of LEO satellite topology design parameters and their impact on network performance while using the ISLs. In particular, we focus on three design parameters: the number of orbits in a shell, the inclination of orbits, and the number of satellites per orbit. Through an extensive analysis of real-world and synthetic satellite configurations, we uncover several interesting properties of satellite topologies. Notably, there exist thresholds for the number of satellites per orbit and the number of orbits below which the latency performance degrades significantly. Moreover, network delay between a pair of traffic endpoints depends on the alignment of the satellite's orbit (Inclination) with the geographic locations of endpoints.

cs.NI↗

arXiv2024

Rethinking the Key Factors for the Generalization of Remote Sensing Stereo Matching Networks

Stereo matching, a critical step of 3D reconstruction, has fully shifted towards deep learning due to its strong feature representation of remote sensing images. However, ground truth for stereo matching task relies on expensive airborne LiDAR data, thus making it difficult to obtain enough samples for supervised learning. To improve the generalization ability of stereo matching networks on cross-domain data from different sensors and scenarios, in this paper, we dedicate to study key training factors from three perspectives. (1) For the selection of training dataset, it is important to select data with similar regional target distribution as the test set instead of utilizing data from the same sensor. (2) For model structure, cascaded structure that flexibly adapts to different sizes of features is preferred. (3) For training manner, unsupervised methods generalize better than supervised methods, and we design an unsupervised early-stop strategy to help retain the best model with pre-trained weights as the basis. Extensive experiments are conducted to support the previous findings, on the basis of which we present an unsupervised stereo matching network with good generalization performance. We release the source code and the datasets at https://github.com/Elenairene/RKF_RSSM to reproduce the results and encourage future work.

cs.CV↗

arXiv2024

On Convergence of Discrete Schemes for Computing the Rate-Distortion Function of Continuous Source

Computing the rate-distortion function for continuous sources is commonly regarded as a standard continuous optimization problem. When numerically addressing this problem, a typical approach involves discretizing the source space and subsequently solving the associated discrete problem. However, existing literature has predominantly concentrated on the convergence analysis of solving discrete problems, usually neglecting the convergence relationship between the original continuous optimization and its associated discrete counterpart. This neglect is not rigorous, since the solution of a discrete problem does not necessarily imply convergence to the solution of the original continuous problem, especially for non-linear problems. To address this gap, our study employs rigorous mathematical analysis, which constructs a series of finite-dimensional spaces approximating the infinite-dimensional space of the probability measure, establishing that solutions from discrete schemes converge to those from the continuous problems.

cs.IT↗

arXiv2024

A Double Maximization Approach for Optimizing the LM Rate of Mismatched Decoding

An approach is established for maximizing the Lower bound on the Mismatch capacity (hereafter abbreviated as LM rate), a key performance bound in mismatched decoding, by optimizing the channel input probability distribution. Under a fixed channel input probability distribution, the computation of the corresponding LM rate is a convex optimization problem. When optimizing the channel input probability distribution, however, the corresponding optimization problem adopts a max-min formulation, which is generally non-convex and is intractable with standard approaches. To solve this problem, a novel dual form of the LM rate is proposed, thereby transforming the max-min formulation into an equivalent double maximization formulation. This new formulation leads to a maximization problem setup wherein each individual optimization direction is convex. Consequently, an alternating maximization algorithm is established to solve the resultant maximization problem setup. Each step of the algorithm only involves a closed-form iteration, which is efficiently implemented with standard optimization procedures. Numerical experiments show the proposed approach for optimizing the LM rate leads to noticeable rate gains.

cs.IT↗

arXiv2024

An Expectation-Maximization Relaxed Method for Privacy Funnel

The privacy funnel (PF) gives a framework of privacy-preserving data release, where the goal is to release useful data while also limiting the exposure of associated sensitive information. This framework has garnered significant interest due to its broad applications in characterization of the privacy-utility tradeoff. Hence, there is a strong motivation to develop numerical methods with high precision and theoretical convergence guarantees. In this paper, we propose a novel relaxation variant based on Jensen's inequality of the objective function for the computation of the PF problem. This model is proved to be equivalent to the original in terms of optimal solutions and optimal values. Based on our proposed model, we develop an accurate algorithm which only involves closed-form iterations. The convergence of our algorithm is theoretically guaranteed through descent estimation and Pinsker's inequality. Numerical results demonstrate the effectiveness of our proposed algorithm.

cs.IT↗

arXiv2024

ORBGRAND: Achievable Rate for General Bit Channels and Application in BICM

Guessing random additive noise decoding (GRAND) has received widespread attention recently, and among its variants, ordered reliability bits GRAND (ORBGRAND) is particularly attractive due to its efficient utilization of soft information and its amenability to hardware implementation. It has been recently shown that ORBGRAND is almost capacity-achieving in additive white Gaussian noise channels under antipodal input. In this work, we first extend the analysis of ORBGRAND achievable rate to memoryless binary-input bit channels with general output conditional probability distributions. The analytical result also sheds insight into understanding the gap between the ORBGRAND achievable rate and the channel mutual information. As an application of the analysis, we study the ORBGRAND achievable rate of bit-interleaved coded modulation (BICM). Numerical results indicate that for BICM, the gap between the ORBGRAND achievable rate and the channel mutual information is typically small, and hence suggest the feasibility of ORBGRAND for channels with high-order coded modulation schemes.

cs.IT↗

arXiv2024

Approaching Maximum Likelihood Decoding Performance via Reshuffling ORBGRAND

Guessing random additive noise decoding (GRAND) is a recently proposed decoding paradigm particularly suitable for codes with short length and high rate. Among its variants, ordered reliability bits GRAND (ORBGRAND) exploits soft information in a simple and effective fashion to schedule its queries, thereby allowing efficient hardware implementation. Compared with maximum likelihood (ML) decoding, however, ORBGRAND still exhibits noticeable performance loss in terms of block error rate (BLER). In order to improve the performance of ORBGRAND while still retaining its amenability to hardware implementation, a new variant of ORBGRAND termed RS-ORBGRAND is proposed, whose basic idea is to reshuffle the queries of ORBGRAND so that the expected number of queries is minimized. Numerical simulations show that RS-ORBGRAND leads to noticeable gains compared with ORBGRAND and its existing variants, and is only 0.1dB away from ML decoding, for BLER as low as $10^{-6}$.

cs.IT↗

arXiv2024

PAC Learnability for Reliable Communication over Discrete Memoryless Channels

In practical communication systems, knowledge of channel models is often absent, and consequently, transceivers need be designed based on empirical data. In this work, we study data-driven approaches to reliably choosing decoding metrics and code rates that facilitate reliable communication over unknown discrete memoryless channels (DMCs). Our analysis is inspired by the PAC (probably approximately correct) learning theory and does not rely on any assumptions on the statistical characteristics of DMCs. We show that a naive plug-in algorithm for choosing decoding metrics is likely to fail for finite training sets. We propose an alternative algorithm called the virtual sample algorithm and establish a non-asymptotic lower bound on its performance. The virtual sample algorithm is then used as a building block for constructing a learning algorithm that chooses a decoding metric and a code rate using which a transmitter and a receiver can reliably communicate at a rate arbitrarily close to the channel mutual information. Therefore, we conclude that DMCs are PAC learnable.

cs.IT↗

arXiv2024

Efficient and Provably Convergent Computation of Information Bottleneck: A Semi-Relaxed Approach

Information Bottleneck (IB) is a technique to extract information about one target random variable through another relevant random variable. This technique has garnered significant interest due to its broad applications in information theory and deep learning. Hence, there is a strong motivation to develop efficient numerical methods with high precision and theoretical convergence guarantees. In this paper, we propose a semi-relaxed IB model, where the Markov chain and transition probability condition are relaxed from the relevance-compression function. Based on the proposed model, we develop an algorithm, which recovers the relaxed constraints and involves only closed-form iterations. Specifically, the algorithm is obtained by analyzing the Lagrangian of the relaxed model with alternating minimization in each direction. The convergence property of the proposed algorithm is theoretically guaranteed through descent estimation and Pinsker's inequality. Numerical experiments across classical and discrete distributions corroborate the analysis. Moreover, our proposed algorithm demonstrates notable advantages in terms of computational efficiency, evidenced by significantly reduced run times compared to existing methods with comparable accuracy.

cs.IT↗

arXiv2024

A Constrained BA Algorithm for Rate-Distortion and Distortion-Rate Functions

The Blahut-Arimoto (BA) algorithm has played a fundamental role in the numerical computation of rate-distortion (RD) functions. This algorithm possesses a desirable monotonic convergence property by alternatively minimizing its Lagrangian with a fixed multiplier. In this paper, we propose a novel modification of the BA algorithm, wherein the multiplier is updated through a one-dimensional root-finding step using a monotonic univariate function, efficiently implemented by Newton's method in each iteration. Consequently, the modified algorithm directly computes the RD function for a given target distortion, without exploring the entire RD curve as in the original BA algorithm. Moreover, this modification presents a versatile framework, applicable to a wide range of problems, including the computation of distortion-rate (DR) functions. Theoretical analysis shows that the outputs of the modified algorithms still converge to the solutions of the RD and DR functions with rate $O(1/n)$, where $n$ is the number of iterations. Additionally, these algorithms provide $\varepsilon$-approximation solutions with $O\left(\frac{MN\log N}{\varepsilon}(1+\log |\log \varepsilon|)\right)$ arithmetic operations, where $M,N$ are the sizes of source and reproduced alphabets respectively. Numerical experiments demonstrate that the modified algorithms exhibit significant acceleration compared with the original BA algorithms and showcase commendable performance across classical source distributions such as discretized Gaussian, Laplacian and uniform sources.

cs.IT↗