Kai Li

The paper proposes Information Sufficiency (IS) as a comprehensive framework for privacy-preserving LLM communication, demonstrating that free-text pseudonymization outperforms existing suppression and generalization methods, especially in multi-turn conversations.

The paper introduces PSR extsuperscript{2}, a novel static analysis framework that significantly improves the detection of atomicity violations in smart contracts by combining structural path searching with deep semantic reasoning.

NFTDELTA is a novel framework that uses multi-view learning on static code analysis to detect permission control vulnerabilities in NFT contracts with high accuracy.

SAFEDREAM introduces a lightweight, external world-model framework that proactively detects multi-turn jailbreak attacks by modeling cumulative safety erosion and predicting early failure points.

This paper introduces Heimdallr, a novel framework that characterizes and detects LLM-induced security risks by analyzing the full execution chain of LLM integrations within GitHub CI workflows.

The paper proposes an Augmented Model maniPulation (AugMP) strategy, utilizing graph representation learning, to effectively and stealthily manipulate federated fine-tuning of LLMs, significantly degrading global model performance while evading standard defenses.

MIRA proposes a novel source-aware filtering framework that discovers and anchors evaluation rubrics during data selection, significantly improving code-oriented mid-training data quality while reducing token usage.

This paper analyzes the decoding process of masked diffusion models for graph-to-text generation, finding that structural fine-tuning disrupts natural entity-first generation and proposing a structural decoding method to fix it.

The paper proposes a novel trace-aware decoding framework, combining Temporal-Spatial Parallel Decoding (TSPD) and Confidence Extrapolation (CE), to significantly accelerate the inference of diffusion-based LLMs by identifying and fixing converged tokens early.

The paper proposes a novel render-free framework that conditions video diffusion models directly on compressed 3D human mesh tokens, enabling robust 3D-aware human motion control without relying on rendered 2D guidance.

The paper introduces Humanoid-GPT, a large-scale generative Transformer model that achieves robust zero-shot motion tracking and control by training on a massive, unified corpus of motion data.

This paper proposes a new router redesign for Mixture-of-Experts models using Manifold Power Iteration to align router rows with the principal singular directions of associated experts.

A framework called DexCompose is proposed to reuse pretrained dexterous policies for multi-task manipulation with explicit finger-level action ownership.

This paper introduces Chronos, a physics-informed framework for non-Markovian long-horizon manipulation, which elevates observation history to the latent state of the policy dynamics and achieves higher success rates and fewer parameters than Markovian VLA baselines in both simulated and real-world experiments.

The paper proposes FedLAB, a traceable semantic codebook framework for federated multimodal graph foundation learning, which organizes multimodal graph knowledge into hierarchical codebooks and refines them through federated semantic barycenter pre-training.

The paper proposes FocalSE, a method for enhancing speech in neural speech codecs by performing feature denoising, separation, and recognition in the continuous embedding space.

This paper proposes Wat3R, a cross-domain semi-supervised learning framework for adapting 3D reconstruction models from air to underwater scenes using unlabeled real underwater video footage and a teacher-student architecture.

The paper introduces DexVerse, a large-scale and modular benchmark for dexterous manipulation with 100 tasks, 3 robot arms, 6 hands, and configurable visual variations.

This paper introduces the REAL-TSE Challenge, a satellite challenge on target speaker extraction from real conversational recordings, and describes its task definition, datasets, evaluation protocol, and submitted systems.

A single robot platform, Handroid, is introduced that can function as both a dexterous hand and a humanoid robot, with interchangeable control and learning frameworks.