Bin Zhang

The paper introduces extsc{Spore}, a novel, training-free, and highly efficient privacy extraction attack that targets sensitive information stored in the memory of LLM agents during inference, outperforming existing state-of-the-art methods.

HASTE introduces group-shared fixed fan-in sparsity for multi-label classification, achieving significant wall-clock speedups (up to 25x in backward pass) by enabling efficient GPU execution while maintaining high accuracy.

DFlare introduces a lightweight layer-wise fusion mechanism to overcome the narrow conditioning bottleneck of existing block diffusion methods, enabling the scaling of draft models and achieving superior speculative decoding speedups across multiple LLMs.

HarnessForge introduces a meta-adaptive framework that jointly evolves the execution structure (harness) and the reasoning policy of LLM agents, significantly improving overall system performance across diverse tasks.

This paper proposes a hybrid two-stage diffusion transformer architecture for instruction-guided audio editing, balancing performance and efficiency.

The paper introduces GigaSpeechBench, a comprehensive multilingual and multidimensional ASR & AST benchmark with 680 hours of human-annotated speech, featuring 12 low-resource languages, 6 Chinese dialects, 6 English accents, dense terminology, older adult and child speech, and human-annotated translations.

This paper presents a product coding scheme that converts bit-level reliability into block-level reliability, achieving the same asymptotic rate.

Log-Insight is an automated incident-diagnosis system for large-scale microservice systems that reduces raw events by 1,000-7,000x while preserving statistically significant failure signals, achieving high accuracy in under a minute.

This paper explores safety risks in humorization of large language models (LLMs) and introduces HumorSafe framework for evaluating latent safety risks.

MagicSelector is a framework for tool retrieval in agents using counterfactual task decomposition, progressive reranking, and dynamic Top-K.

This paper introduces UniRank, an open benchmark for comparing and studying unified ranking models that combine sequential modeling and feature interaction.