Bo Li

This survey provides a comprehensive, structured review of safety research in Embodied AI, analyzing attacks and defenses across the entire embodied pipeline to guide the development of safe, robust, and reliable real-world agents.

MARD introduces a multi-agent framework that combines Large Language Models (LLMs) with traditional static analysis engines to achieve robust and highly interpretable Android malware detection with low computational cost.

The paper introduces ML-Bench, a policy-grounded multilingual safety benchmark, and ML-Guard, a superior guardrail model that enables culturally and legally aligned safety assessment for LLMs across 14 languages.

The paper introduces TurnGate, a response-aware defense mechanism that detects the earliest turn in a multi-turn dialogue where the accumulated interaction enables a harmful action, significantly improving malicious intent detection.

The paper proposes using GUI agents, both as objective evaluators and subjective playtesters, to significantly improve the generation of playable games from prompts, demonstrating a 66.8% rubric pass-rate with a novel iterative framework.

The paper proposes using GUI agents, both as objective evaluators and subjective playtesters, to significantly improve the generation of playable games from prompts, demonstrating a 66.8% rubric pass-rate with a novel iterative framework.

The paper introduces DrugClaw, a multi-agent system, and DrugAudit, a new benchmark, demonstrating that DrugClaw excels at answering drug-related questions by grounding answers in primary regulatory sources.

The paper introduces DrugClaw, a multi-agent system, and DrugAudit, a new benchmark, demonstrating that DrugClaw excels at answering drug-related questions by grounding answers in primary regulatory sources.

This paper introduces a stakeholder-centric benchmark to evaluate the vulnerabilities of large language models in web agent systems.

The paper presents PULSE, an automatic pipeline-parallel training strategy for diffusion models that reduces communication overhead and increases training throughput.

This paper introduces GEAR, a method for training a vector-quantized tokenizer and an autoregressive generator jointly and end-to-end, resolving the issue of non-differentiable VQ indices.

This paper proposes SkillOpt-Lite, a minimal viable pipeline for skill optimization in autonomous agents, which accelerates convergence and outperforms full SkillOpt.

This paper proposes MineValiCoder, a collaborative closed-loop TDD framework using mutual reinforcement of test-case quality and code quality to address stochasticity in Large Language Model-based Test-Driven Development.

This paper introduces AgentSysBench, a benchmark suite and measurement toolkit for agentic applications, and identifies six properties that distinguish agentic workloads from conventional LLM serving.

The paper investigates the effectiveness of automatic red teaming using large language models and proposes a role-separated approach.