Liang He

The paper introduces AgentSchool, an advanced LLM-powered multi-agent simulator that models learning as state transitions to provide a robust, ethically viable testbed for educational research and pedagogical reform.

BudgetDraft introduces an acceptance-aware multi-view training method that trains a sparse-KV speculative decoder to maintain high acceptance rates across varying context lengths and sparsity levels, achieving significant speedups in memory-constrained, long-context inference.

The paper proposes TRACE, a novel agentic jailbreaking framework that successfully bypasses safety mechanisms of advanced LLM agents by decomposing malicious tasks and disguising harmful subtasks within task-aware, iteratively evolved scenarios.

MemPro introduces a system-level evolution framework that treats the entire memory construction-retrieval pipeline as an evolvable program, significantly improving long-horizon agent performance over fixed-pipeline baselines.

The paper introduces CaDDTree, a cost-aware method that optimizes token throughput by jointly selecting the tree structure and node budget for speculative decoding, outperforming existing methods like DDTree.

MLEvolve is a novel self-evolving multi-agent framework that enables LLM agents to discover and optimize machine learning algorithms for complex, long-horizon tasks.

This paper introduces Agents-K1, an end-to-end knowledge orchestration pipeline that converts raw documents into agent-native scientific knowledge graphs.