Dongrui Liu

Attributing Emergence in Million-Agent Systems

Ling Tang, Jilin Mei, Qian Chen, Qihan Ren et autres

Large language models (LLMs) can simulate human-like reasoning and decision-making in individual agents. LLM-powered multi-agent systems (MAS) combine such agents to simulate population-scale social phenomena such as polarization, information cascades, and market panics. Such studies require attributing macro emergence to individual agents, …

Attributing Emergence in Million-Agent Systems

Ling Tang, Jilin Mei, Qian Chen, Qihan Ren et autres

Large language models (LLMs) can simulate human-like reasoning and decision-making in individual agents. LLM-powered multi-agent systems (MAS) combine such agents to simulate population-scale social phenomena such as polarization, information cascades, and market panics. Such studies require attributing macro emergence to individual agents, …

cn (code pays fourni par la source)

IS-Bench: Evaluating Interactive Safety of VLM-Driven Embodied Agents in Daily Household Tasks

Xiaoya Lu, Zeren Chen, Xuhao Hu, Yijin Zhou et autres

Flawed planning from VLM-driven embodied agents poses significant safety hazards, hindering their deployment in real-world household tasks. However, existing static, termination-oriented evaluation paradigms fail to adequately assess risks within these interactive environments, since they cannot simulate dynamic risks that emerge from an …

cn (code pays fourni par la source)

Discovery and Activity Evaluation of Novel Dibenzoxazepinone Derivatives as Glycogen Phosphorylase Inhibitors

Dongrui Liu, Zhiwei Yan, Youde Wang, Shuai Li et autres

Inhibition of glycogen phosphorylases (GP) has been regarded as a therapeutic strategy for blood glucose control in diabetes. In this study, a series of novel dibenzoxazepinone derivatives was synthesized. The in vitro activity screening results indicated that compound Id most significantly inhibited …

cn (code pays fourni par la source)

AI Deception: Risks, Dynamics, and Controls

Boyuan Chen, S. S. Fang, Jiaming Ji, Yanxu Zhu et autres

As intelligence increases, so does its shadow. AI deception, in which systems induce false beliefs to secure self-beneficial outcomes, has evolved from a speculative concern to an empirically demonstrated risk across language models, AI agents, and emerging frontier systems. This project provides …

us (code pays fourni par la source)

ExGRPO: Learning to Reason from Experience

Runzhe Zhan, Yafu Li, Zhi-Wei Wang, Xiaoye Qu et autres

Reinforcement learning from verifiable rewards (RLVR) is an emerging paradigm for improving the reasoning ability of large language models. However, standard on-policy training discards rollout experiences after a single update, leading to computational inefficiency and instability. While prior work on RL has …

Transcriptome analysis provides insights into the regulatory role of phytohormone in second flowering of pear

Dongrui Liu, Cong Shi, Jie Ding, Xuepeng Sun et autres

cn (code pays fourni par la source)

The Devil behind the mask: An emergent safety vulnerability of Diffusion LLMs

Zichen Wen, Jinping Qu, Dongrui Liu, Zhiyuan Liu et autres

Diffusion-based large language models (dLLMs) have recently emerged as a powerful alternative to autoregressive LLMs, offering faster inference and greater interactivity via parallel decoding and bidirectional modeling. However, despite strong performance in code generation and text infilling, we identify a fundamental safety …

IS-Bench: Evaluating Interactive Safety of VLM-Driven Embodied Agents in Daily Household Tasks

Xiaoya Lu, Zeren Chen, Xinye Hu, Weichen Zhang et autres

Flawed planning from VLM-driven embodied agents poses significant safety hazards, hindering their deployment in real-world household tasks. However, existing static, non-interactive evaluation paradigms fail to adequately assess risks within these interactive environments, since they cannot simulate dynamic risks that emerge from an …

Accelerating Diffusion Large Language Models with SlowFast Sampling: The Three Golden Principles

Yaojie Zhang, Zhiyuan Liu, Dongrui Liu, Linfeng Zhang

Diffusion-based language models (dLLMs) have emerged as a promising alternative to traditional autoregressive LLMs by enabling parallel token generation and significantly reducing inference latency. However, existing sampling strategies for dLLMs, such as confidence-based or semi-autoregressive decoding, often suffer from static behavior, leading …

Decouple-Then-Merge: Finetune Diffusion Models as Multi-Task Learning

Qianli Ma, Xuefei Ning, Dongrui Liu, Li Niu et autres

Diffusion models are trained by learning a sequence of models that reverse each step of noise corruption. Typically, the model parameters are fully shared across multiple timesteps to enhance training efficiency. However, since the denoising tasks differ at each timestep, the gradients …

cn (code pays fourni par la source)

The light-harvesting complex protein LHCB4 fine-tunes temperature-regulated flowering via chloroplast retrograde signaling in pear

Zeyu Jiang, Dongrui Liu, Siting Zhu, Biying Zhao et autres

Flowering is crucial for plant reproductive success and is regulated by both endogenous and external factors. However, the mechanisms by which ambient temperature influences flowering in pear (Pyrus spp.) remain poorly understood. In this study, we observed that elevated temperatures induce early …

cn (code pays fourni par la source)