Han Li

The paper introduces Plan, a structured agentic behavior that decomposes multi-hop questions into ordered sub-questions before retrieval, and proposes a self-bootstrapping paradigm to train it without relying on model distillation.

The paper introduces a histogram-regularized latent diffusion model to synthesize highly realistic and subtype-specific pulmonary nodules in 3D CT volumes, addressing the limitations of existing methods that fail to capture accurate lesion-level intensity distributions.

This paper introduces a new evaluation framework, SpatialUncertain, demonstrating that current Vision-Language Models (VLMs) are prone to overconfident and incorrect answers to spatial questions when visual evidence is incomplete or misleading.

HoliTok introduces a novel continuous holistic tokenization model that provides a unified, high-fidelity latent representation for simultaneously supporting both speech generation and speech understanding tasks.

The paper introduces Source-Grounded Semantic Reinforcement Learning (SG-SRL), a framework that leverages abundant source-language monolingual data to improve target-language generation in low-resource settings by providing cross-lingual semantic supervision.

The paper introduces an adaptive interview framework to gather rich persona context, demonstrating that LLMs improve decision alignment in moral dilemmas only when they selectively ground their decisions in follow-up-derived, user-specific evidence.

The paper introduces DiffErase, a black-box attack that effectively removes inaudible audio watermarks while preserving perceptual quality by utilizing diffusion models.

The paper introduces PIGMENT, a physics-informed foundation model that enables reliable quantitative mapping of brain microstructure from extremely sparse or challenging diffusion MRI scans.

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 OPD+, a corrected on-policy distillation framework that mathematically proves the bias of standard stop-gradient methods and improves the stability and performance of knowledge transfer from teacher to student models.

LongAttnComp introduces a novel, two-stage fine-tuning framework for context compression that significantly improves long-context reasoning performance, matching or exceeding full-context accuracy on demanding tasks like code debugging.

The paper reframes Parameter-Efficient Fine-Tuning (PEFT) from a mere cost-saving alternative to a robust architecture for creating persistent, personalized models that layer specific behaviors onto large shared foundation models.

The paper introduces U4D, an uncertainty-aware framework that synthesizes 4D LiDAR scenes by prioritizing the reconstruction of geometrically difficult and uncertain regions first, leading to state-of-the-art fidelity and temporal consistency.

The paper introduces X-Stream, a new benchmark for multi-stream video understanding, and finds that current state-of-the-art MLLMs perform poorly when required to process multiple concurrent video streams.

The paper introduces MMG2Skill, a closed-loop framework that converts noisy, human-oriented web guides into editable, executable skills, significantly improving agent performance across diverse tasks.

The paper proposes Credit-Attenuated Privileged Feedback (CAPF), a training-time mechanism that uses verifier-side information to guide LLM search agents, significantly improving their performance on complex QA tasks.

The paper proposes DySCo, a dynamic trust-aware sparse consensus mechanism, to efficiently manage communication in multi-agent LLM systems by selectively connecting agents based on real-time value, thus reducing overhead while maintaining critical cross-validation.

The paper introduces ContinuousBench, a novel benchmark designed to rigorously test if differentially private (DP) synthetic text can genuinely transfer new knowledge, finding that state-of-the-art DP synthesis methods generally fail to achieve this capability gain.

The paper introduces ContinuousBench, a dynamic benchmark designed to rigorously test if differentially private (DP) synthetic text can genuinely transfer new knowledge and capabilities from sensitive source corpora, finding that current state-of-the-art DP methods generally fail to achieve this.

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.