arXiv subjects

Wei Kang

Publications and source records attributed to Wei Kang.

At least 19 recordsLinked to original sources

arXiv2026

Polar Code Based Federated Learning: Convergence Analysis and Resource Allocation

Federated learning (FL) enables collaborative model training across distributed devices without sharing raw data; however, it faces significant communication bottlenecks and channel impairments in practice. Conventional network layer treatments either idealize the channel as error free or apply equal error protection (EEP) to transmitted model updates, failing to account for the inherently unequal importance of quantization bits within a single local model. To address this limitation, we propose a cross layer polar code based FL scheme that leverages the unequal error protection (UEP) property of polar codes under finite block lengths. Specifically, the proposed design selectively protects more significant quantization bits, thereby mitigating the detrimental effects of channel noise. We further provide a rigorous convergence analysis of the proposed scheme, deriving an upper bound on the convergence gap, which we then jointly optimize over the number of quantization bits and the polar code block length across all training iterations. Experimental results demonstrate that both constant and variable block length configurations of our polar code based scheme consistently achieve substantial performance gains over uncoded and LDPC-based EEP benchmarks, with the advantage becoming increasingly pronounced as the channel quality deteriorating. These findings confirm the efficacy of our cross-layer design in enhancing FL robustness and efficiency under realistic channel conditions.

cs.LG↗

arXiv2026

Robust Indicators of Spatial Association

The Moran statistic, and its accompanying local statistics, are one of the most extensively used exploratory spatial data analysis tools for assessing global and local spatial autocorrelation. The paired visualizations for these statistics, the Moran Scatterplot and LISA map, are likewise central to spatial analysis. Together, these statistics and visualizations are used to identify spatial clusters, regions of a map where observations are similar to one another, or spatial outliers, observations that differ sharply from their surroundings. However, the use of Moran statistics to detect spatial outliers is complicated by their high sensitivity to *distributional* outliers: observations that are extreme relative to the overall data distribution, regardless of their spatial context. Indeed, a single distributional outlier can (I) distort local statistics across the entire map and (II) bias the global estimate of spatial association. Recent work has begun to address (I) and (II) separately using plug-in robust estimators for location, scale, and spatial correlation. In this paper, we offer the first systematic evaluation of robust LISA and global spatial association measures, using variety of plug-in robust estimators, a trimmed least squares (TLS) estimator, and a Theil-Sen-style estimator. We also outline a visualization strategy to create Robust Moran Scatterplots/LISA maps for each. Out of all considered approaches, we find that the Theil-Sen Moran estimator is a better default for exploratory spatial data analysis and visualization, while robust plug-in estimators also offer acceptable performance in large datasets.

stat.ME↗

arXiv2026

On Demand-Private Coded Caching With Multiple Demands

We consider a coded caching problem with multiple demands under a privacy constraint. In this problem, a server with access to $N$ files serves $K$ users over a shared link, and each user requests $L$ distinct files. The privacy constraint requires that each user obtain no information about the demands of the other users. We propose a new achievable scheme for arbitrary numbers of files and users. The scheme is obtained via a transformation from a non-private coded caching scheme under uncoded placement for $N$ files and $K \cdot \min\{N,KL\}$ users, where each user requests one file and the demands are restricted to a subset of all possible demands. We then derive a converse bound, and the proposed scheme is shown to be order optimal within a factor of 6 of this bound.

cs.IT↗

arXiv2026

OmniVoice: Towards Omnilingual Zero-Shot Text-to-Speech with Diffusion Language Models

We present OmniVoice, a massively multilingual zero-shot text-to-speech (TTS) model that scales to over 600 languages. At its core is a novel diffusion language model-style discrete non-autoregressive (NAR) architecture. Unlike conventional discrete NAR models that suffer from performance bottlenecks in complex two-stage (text-to-semantic-to-acoustic) pipelines, OmniVoice directly maps text to multi-codebook acoustic tokens. This simplified approach is facilitated by two key technical innovations: (1) a full-codebook random masking strategy for efficient training, and (2) initialization from a pre-trained LLM to ensure superior intelligibility. By leveraging a 581k-hour multilingual dataset curated entirely from open-source data, OmniVoice achieves the broadest language coverage to date and delivers state-of-the-art performance across Chinese, English, and diverse multilingual benchmarks. Our code and pre-trained models are publicly available at https://github.com/k2-fsa/OmniVoice.

cs.CL↗

arXiv2026

Movable Antenna Assisted Dual-Polarized Multi-Cell Cooperative AirComp: An Alternating Optimization Approach

Over-the-air computation (AirComp) is a key enabler for distributed optimization, since it leverages analog waveform superposition to perform aggregation and thereby mitigates the communication bottleneck caused by iterative information exchange. However, AirComp is sensitive to wireless environment and conventional systems with fixed single-polarized base-station arrays cannot fully exploit spatial degrees of freedom while also suffering from polarization mismatch. To overcome these limitations, this paper proposes a multi-cell cooperative air-computation framework assisted by dual-polarized movable antennas (D-PMA), and formulates a mean squared error (MSE) minimization problem by jointly optimizing the combining matrix, polarization vectors, antenna positions, and user transmit coefficients. The resulting problem is highly nonconvex, so an alternating algorithm is developed in which closed-form updates are obtained for the combining matrix and transmit coefficients. Then a method based on successive convex approximation (SCA) and semidefinite relaxation (SDR) is proposed to refine polarization vectors, and the antenna positions are updated using a gradient-based method. In addition, we develop a statistical-channel-based scheme for optimizing the antenna locations, and we further present the corresponding algorithm to efficiently obtain the solution. Numerical results show that the proposed movable dual-polarized scheme consistently outperforms movable single-polarized and fixed-antenna baselines under both instantaneous and statistical channels.

cs.IT↗

arXiv2025

Sum Rate optimization for RIS-Aided RSMA system with Movable Antenna

Rate-Splitting Multiple Access (RSMA) is a key enabling technique for sixth-generation (6G) wireless systems due to its powerful interference management, and Reconfigurable Intelligent Surface (RIS) improves communication performance by shaping wireless propagation. However, conventional RSMA--RIS architectures employ fixed antennas, limiting spatial degrees of freedom and system performance. To address this, we propose a movable-antenna (MA) assisted RSMA--RIS framework and formulate a sum-rate maximization problem that jointly optimizes the transmit beamforming matrix, RIS reflection matrix, common-rate partition, and MA positions. After yielding a closed-form solution for common rate splitting, the problem is transformed via fractional programming (FP). Using Karush--Kuhn--Tucker (KKT) conditions, we give iterative updates for Lagrange multipliers and beamforming matrix, obtain the RIS reflection matrix via the dual problem, and determine optimal antenna positions via gradient ascent. Numerical results show that with the existence of RIS, integrating MA yields additional gains of approximately 33.3\% for SDMA and 35.6\% for RSMA.

cs.IT↗

arXiv2025

Flow2GAN: Hybrid Flow Matching and GAN with Multi-Resolution Network for Few-step High-Fidelity Audio Generation

Existing dominant methods for audio generation include Generative Adversarial Networks (GANs) and diffusion-based methods like Flow Matching. GANs suffer from slow convergence during training, while diffusion methods require multi-step inference that introduces considerable computational overhead. In this work, we introduce Flow2GAN, a two-stage framework that combines Flow Matching training for learning generative capabilities with GAN fine-tuning for efficient few-step inference. Specifically, given audio's unique properties, we first improve Flow Matching for audio modeling through: 1) reformulating the objective as endpoint estimation, avoiding velocity estimation difficulties when involving empty regions; 2) applying spectral energy-based loss scaling to emphasize perceptually salient quieter regions. Building on these Flow Matching adaptations, we demonstrate that a further stage of lightweight GAN fine-tuning enables us to obtain few-step (e.g., 1/2/4 steps) generators that produce high-quality audio. In addition, we develop a multi-branch network architecture that processes Fourier coefficients at different time-frequency resolutions, which improves the modeling capabilities compared to prior single-resolution designs. Experimental results indicate that our Flow2GAN delivers high-fidelity audio generation from Mel-spectrograms or discrete audio tokens, achieving highly favorable quality-efficiency trade-offs compared to existing state-of-the-art GAN-based and Flow Matching-based methods. Online demo samples are available at https://flow2gan.github.io, and the source code is released at https://github.com/k2-fsa/Flow2GAN.

eess.AS↗

arXiv2025

EvoMail: Self-Evolving Cognitive Agents for Adaptive Spam and Phishing Email Defense

Modern email spam and phishing attacks have evolved far beyond keyword blacklists or simple heuristics. Adversaries now craft multi-modal campaigns that combine natural-language text with obfuscated URLs, forged headers, and malicious attachments, adapting their strategies within days to bypass filters. Traditional spam detection systems, which rely on static rules or single-modality models, struggle to integrate heterogeneous signals or to continuously adapt, leading to rapid performance degradation. We propose EvoMail, a self-evolving cognitive agent framework for robust detection of spam and phishing. EvoMail first constructs a unified heterogeneous email graph that fuses textual content, metadata (headers, senders, domains), and embedded resources (URLs, attachments). A Cognitive Graph Neural Network enhanced by a Large Language Model (LLM) performs context-aware reasoning across these sources to identify coordinated spam campaigns. Most critically, EvoMail engages in an adversarial self-evolution loop: a ''red-team'' agent generates novel evasion tactics -- such as character obfuscation or AI-generated phishing text -- while the ''blue-team'' detector learns from failures, compresses experiences into a memory module, and reuses them for future reasoning. Extensive experiments on real-world datasets (Enron-Spam, Ling-Spam, SpamAssassin, and TREC) and synthetic adversarial variants demonstrate that EvoMail consistently outperforms state-of-the-art baselines in detection accuracy, adaptability to evolving spam tactics, and interpretability of reasoning traces. These results highlight EvoMail's potential as a resilient and explainable defense framework against next-generation spam and phishing threats.

cs.LG↗

arXiv2025

Controlling Language Difficulty in Dialogues with Linguistic Features

Large language models (LLMs) have emerged as powerful tools for supporting second language acquisition, particularly in simulating interactive dialogues for speaking practice. However, adapting the language difficulty of LLM-generated responses to match learners' proficiency levels remains a challenge. This work addresses this issue by proposing a framework for controlling language proficiency in educational dialogue systems. Our approach leverages three categories of linguistic features, readability features (e.g., Flesch-Kincaid Grade Level), syntactic features (e.g., syntactic tree depth), and lexical features (e.g., simple word ratio), to quantify and regulate text complexity. We demonstrate that training LLMs on linguistically annotated dialogue data enables precise modulation of language proficiency, outperforming prompt-based methods in both flexibility and stability. To evaluate this, we introduce Dilaprix, a novel metric integrating the aforementioned features, which shows strong correlation with expert judgments of language difficulty. Empirical results reveal that our approach achieves superior controllability of language proficiency while maintaining high dialogue quality.

cs.CL↗

arXiv2025

ZipVoice-Dialog: Non-Autoregressive Spoken Dialogue Generation with Flow Matching

Generating spoken dialogue is inherently more complex than monologue text-to-speech (TTS), as it demands both realistic turn-taking and the maintenance of distinct speaker timbres. While existing autoregressive (AR) models have made progress, they often suffer from high inference latency and stability issues. To overcome these limitations, we propose ZipVoice-Dialog, a non-autoregressive (NAR) zero-shot spoken dialogue generation model based on flow-matching. Observing that applying vanilla flow-matching to dialogue generation leads to poor speech intelligibility and turn-taking precision, we introduce two simple yet effective methods to adapt flow-matching architectures for dialogue generation: (1) a curriculum learning strategy to ensure robust speech-text alignment, and (2) speaker-turn embeddings to govern precise speaker turn-taking. Additionally, we introduce dedicated strategies to support stereo dialogue generation. Recognizing the lack of training datasets in this field, we curate and release OpenDialog, the first large-scale (6.8k hours) open-source spoken dialogue dataset derived from in-the-wild speech data. Moreover, for fair and rigorous evaluations, we established a benchmark to comprehensively evaluate dialogue generation models. Experiments demonstrate the effectiveness of the proposed methods and dataset, showing that ZipVoice-Dialog achieves superior performance in inference speed, intelligibility, speaker turn-taking accuracy, and speaker similarity. Our code, model checkpoints, and the OpenDialog dataset are publicly available at https://github.com/k2-fsa/ZipVoice.

eess.AS↗

arXiv2025

HiPreNets: High-Precision Neural Networks through Progressive Training

Deep neural networks are powerful tools for solving nonlinear problems in science and engineering, but training highly accurate models becomes challenging as problem complexity increases. Non-convex optimization and sensitivity to hyperparameters make consistent performance improvement difficult, and traditional approaches prioritize minimizing mean squared error while overlooking the $L^{\infty}$ norm error that is critical in safety-sensitive applications. To address these challenges, we present HiPreNets, a progressive framework for training high-precision neural networks through sequential residual refinements. Starting from an initial network, each stage trains a refinement network on the normalized residuals of the ensemble so far, systematically reducing both average and worst-case error. A key theme throughout the framework is concentrating training effort on high-error regions of the input domain, which we pursue through complementary techniques including loss function design, adaptive data sampling, localized patching, and boundary-aware training. We validate the framework on benchmark regression problems from the Feynman dataset, where it consistently outperforms standard fully connected networks and reported Kolmogorov-Arnold Networks results, with accuracy approaching machine precision depending on select problems. We further apply the framework to learning the flow map of a 20-dimensional power system ODE, which appears to be the highest dimensional problem studied using this class of multistage methods, achieving substantial reductions in both RMSE and $L^{\infty}$ norm error while enabling a surrogate that predicts system state $238\times$ faster than direct numerical simulation.

cs.LG↗

arXiv2025

ZipVoice: Fast and High-Quality Zero-Shot Text-to-Speech with Flow Matching

Existing large-scale zero-shot text-to-speech (TTS) models deliver high speech quality but suffer from slow inference speeds due to massive parameters. To address this issue, this paper introduces ZipVoice, a high-quality flow-matching-based zero-shot TTS model with a compact model size and fast inference speed. Key designs include: 1) a Zipformer-based vector field estimator to maintain adequate modeling capabilities under constrained size; 2) Average upsampling-based initial speech-text alignment and Zipformer-based text encoder to improve speech intelligibility; 3) A flow distillation method to reduce sampling steps and eliminate the inference overhead associated with classifier-free guidance. Experiments on 100k hours multilingual datasets show that ZipVoice matches state-of-the-art models in speech quality, while being 3 times smaller and up to 30 times faster than a DiT-based flow-matching baseline. Codes, model checkpoints and demo samples are publicly available at https://github.com/k2-fsa/ZipVoice.

eess.AS↗

arXiv2025

Ion Track Formation via Electric-Field-Enhanced Energy Deposition

High-energy ion irradiation deposits extreme energy in a narrow range (1-10 nm) along ion trajectories in solid through electronic energy loss, producing unique irradiation effects such as ion tracks. However, intrinsic velocity effects impose an upper limit on electronic energy loss that cannot be overcome by adjusting irradiation parameters. We introduce a method using electric fields during irradiation to enhance nanoscale energy deposition by accelerating ion-excited electrons within sub-picosecond timescales.Our extended thermal spike model quantitatively describes this enhancement and predicts a significant reduction in the electronic energy loss required for ion track formation in amorphous SiO2, which is in excellent agreement with experimental observations. This work provides a new approach to control energy deposition during irradiation and boosts the wide application of ion tracks in material modification and nanoengineering to much broader extents.

physics.app-ph↗

arXiv2025

Characterizing the Optimal Memory-Rate Tradeoff in Secure Coded Caching for Small Buffer or Small Rate

We consider the secure coded caching problem proposed by Ravindrakumar et. al where no user can obtain information about files other than the one requested. We first propose three new schemes for the three cases of cache size $M=1$, $N=2$ files and arbitrary $K$ users, delivery rate $ R=1$, arbitrary $N$ files and $K$ users, and the general case for arbitrary $N$ files and $K$ users, respectively. Then we derive converse results by characterizing new properties of secure coded caching schemes. As a result, we characterize the two end-points of the optimal memory-rate tradeoff curve for arbitrary number of users and files. Furthermore, for the case of $N=2$ files and arbitrary number of users, we also characterize a segment of the optimal memory-rate tradeoff curve, where the cache size is relatively small.

cs.IT↗

arXiv2025

Demand Private Coded Caching: Small Cache Size

We investigate the demand private coded caching problem, which is an $(N,K)$ coded caching problem with $N$ files, $K$ users, each equipped with a cache of size $M$, and an additional privacy constraint on user demands, i.e., each user can not gain any information about the demands of other users. We focus on scenarios where the size of users' caches is small, aiming to further characterize the fundamental limits of this problem. We first present a new virtual-user-based achievable scheme for arbitrary number of users and files, and two MDS-code-based achievable schemes for the case $N \le K$. With a newly derived converse bound for the case $N \le K$, these proposed schemes lead to the optimal memory-rate tradeoff of the demand private coded caching problem for $M \in \big[0, \frac{N}{(K+1)(N-1)} \big] $ where $N \le K \le 2N-2$, and the optimal memory-rate tradeoff for $M \in \big[0, \frac{1}{K+1} \big] $ where $ K > 2N-2$. Moreover, for the case of 2 files and arbitrary number of users, by deriving another new converse bound, the optimal memory-rate tradeoff is characterized for $M\in \big[0,\frac{2}{K}\big] \cup \big[\frac{2(K-1)}{K+1},2\big]$. Finally, we provide the optimal memory-rate tradeoff of the demand private coded caching problem for 2 files and 3 users.

cs.IT↗

arXiv2025

EPhishCADE: A Privacy-Aware Multi-Dimensional Framework for Email Phishing Campaign Detection

Phishing attacks, typically carried out by email, remain a significant cybersecurity threat with attackers creating legitimate-looking websites to deceive recipients into revealing sensitive information or executing harmful actions. In this paper, we propose {\bf EPhishCADE}, the first {\em privacy-aware}, {\em multi-dimensional} framework for {\bf E}mail {\bf Phish}ing {\bf CA}mpaign {\bf DE}tection to automatically identify email phishing campaigns by clustering seemingly unrelated attacks. Our framework employs a hierarchical architecture combining a structural layer and a contextual layer, offering a comprehensive analysis of phishing attacks by thoroughly examining both structural and contextual elements. Specifically, we implement a graph-based contextual layer to reveal hidden similarities across multiple dimensions, including textual, numeric, temporal, and spatial features, among attacks that may initially appear unrelated. Our framework streamlines the handling of security threat reports, reducing analysts' fatigue and workload while enhancing protection against these threats. Another key feature of our framework lies in its sole reliance on phishing URLs in emails without the need for private information, including senders, recipients, content, etc. This feature enables a collaborative identification of phishing campaigns and attacks among multiple organizations without compromising privacy. Finally, we benchmark our framework against an established structure-based study (WWW \textquotesingle 17) to demonstrate its effectiveness.

cs.CR↗

arXiv2024

Capacity of Hierarchical Secure Coded Gradient Aggregation with Straggling Communication Links

The growing privacy concerns in distributed learning have led to the widespread adoption of secure aggregation techniques in distributed machine learning systems, such as federated learning. Motivated by a coded gradient aggregation problem in a user-helper-master hierarchical network setting with straggling communication links, we formulate a new secure hierarchical coded gradient aggregation problem. In our setting, $ K $ users communicate with the master through an intermediate layer of $ N $ helpers, who can communicate with each other. With a resiliency threshold of $ N_r $ for straggling communication links, and at most $ T $ colluding helpers and any number of colluding users, the master aims to recover the sum of all users' gradients while remaining unaware of any individual gradient that exceeds the expected sum. In addition, helpers cannot infer more about users' gradients than what is already known by the colluding users. We propose an achievable scheme where users' upload messages are based on a globally known Vandermonde matrix, and helper communication is facilitated using an extended Vandermonde matrix with special structural properties. A matching converse bound is also derived, establishing the optimal result for this hierarchical coded gradient aggregation problem.

cs.IT↗

arXiv2024

k2SSL: A Faster and Better Framework for Self-Supervised Speech Representation Learning

Self-supervised learning (SSL) has achieved great success in speech-related tasks. While Transformer and Conformer architectures have dominated SSL backbones, encoders like Zipformer, which excel in automatic speech recognition (ASR), remain unexplored in SSL. Concurrently, inefficiencies in data processing within existing SSL training frameworks, such as fairseq, pose challenges in managing the growing volumes of training data. To address these issues, we propose k2SSL, an open-source framework that offers faster, more memory-efficient, and better-performing self-supervised speech representation learning, focusing on downstream ASR tasks. The optimized HuBERT and proposed Zipformer-based SSL systems exhibit substantial reductions in both training time and memory usage during SSL training. Experiments on LibriSpeech demonstrate that Zipformer Base significantly outperforms HuBERT and WavLM, achieving up to a 34.8% relative WER reduction compared to HuBERT Base after fine-tuning, along with a 3.5x pre-training speedup in GPU hours. When scaled to 60k hours of LibriLight data, Zipformer Large exhibits remarkable efficiency, matching HuBERT Large's performance while requiring only 5/8 pre-training steps.

eess.AS↗