Yuan Li

The paper proposes an operation-centric, TEE-backed isolation model to constrain self-hosted computer-use agents, preventing malicious or unsafe host-level operations without sacrificing general functionality.

The paper introduces EditRisk-Bench, a novel benchmark designed to systematically evaluate the safety risks and downstream reasoning corruption caused by malicious knowledge editing in large language models.

The paper introduces TESLA, a novel, contactless electromagnetic (EM) side-channel attack that exploits inherent EM emanations from capacitive touchscreens to extract highly sensitive user data like PIN codes and keystrokes.

This paper proposes a novel data-driven image encryption framework that learns the chaotic map dynamics directly from the image data, enhancing security beyond traditional fixed-map schemes.

The paper proposes TimeGuard, a novel channel-wise pool training defense, to significantly improve the robustness of time series forecasting against backdoor attacks by addressing signal dilution and loss degeneration.

VFEAgent is a novel multi-agent framework that automates the entire Finite Element Analysis (FEA) workflow, achieving high success rates in generating complete and physically valid simulations directly from multimodal inputs.

The paper proposes ESRT, an edge-cloud framework that achieves state-of-the-art, bandwidth-efficient, and privacy-preserving many-to-many speech translation across 45 languages by splitting the model inference.

Persona prompting does not universally improve LLM performance; instead, it systematically trades increased expertise depth for reduced clarity, making multi-metric evaluation essential.

The paper proposes Cert-LAS, a novel certified method for verifying model ownership in text-to-image diffusion models, which is robust against malicious signal removal attacks.

UniAudio-Token is a framework that enhances existing semantic speech tokenizers with general audio perception, allowing them to handle diverse audio types while maintaining high-fidelity speech capabilities.

This paper addresses the challenge of achieving optimal fairness and accuracy simultaneously in multi-class classification by proposing novel in-processing and post-processing algorithms that converge to the optimal Pareto frontier.

The paper introduces Dr. DocBench, a difficulty-aware, comprehensive benchmark designed to rigorously test expert-level and challenging document parsing capabilities for VLMs, demonstrating that current state-of-the-art models fail on complex, domain-specific structures.

The paper proposes PG-RSSNN, a physics-guided recurrent state-space neural network that improves multi-step prediction stability and accuracy compared to both pure black-box and pure physics models, especially with limited data.

The paper proposes a novel, theoretically-grounded algorithm (HAMU) that addresses the challenge of machine unlearning by guaranteeing specified improvements in forget quality while minimizing retain utility degradation.

EVA-Net proposes a two-stage framework that uses action videos as semantic priors to achieve strong subject-independent EEG motor decoding, significantly outperforming text-based methods.

MOSS-Audio is a unified audio-language model designed for comprehensive understanding of speech, environmental sounds, and music, achieving strong performance across various audio-grounded tasks.

THRD introduces a novel, training-free framework that models temporal risk accumulation to effectively defend against multi-turn jailbreak attacks on LLMs, significantly reducing attack success rates while maintaining model utility.

The paper demonstrates that explicit gender cues systematically affect LLM value trade-offs, causing decision flips that are often masked or misattributed by the models themselves.

NeuroArmor is a white-box runtime defense that uses prompt-specific safe variants to selectively detect and mitigate jailbreak attacks, significantly reducing attack success rates while maintaining a low false positive rate.

The paper proposes OneReason, a framework that enhances the reasoning capability of generative recommendation models by focusing on improving item perception and structuring user behavior into coherent latent interests.