Qihan Ren

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)

The Why Behind the Action: Unveiling Internal Drivers via Agentic Attribution

Chen Qian, Peng Wang, Dongrui Liu, Junyao Yang et autres

Large Language Model (LLM)-based agents are widely used in real-world applications such as customer service, web navigation, and software engineering. As these systems become more autonomous and are deployed at scale, understanding why an agent takes a particular action becomes increasingly important …

The Why Behind the Action: Unveiling Internal Drivers via Agentic Attribution

Chen Qian, Peng Wang, Dongrui Liu, Junyao Yang et autres

Large Language Model (LLM)-based agents are widely used in real-world applications such as customer service, web navigation, and software engineering. As these systems become more autonomous and are deployed at scale, understanding why an agent takes a particular action becomes increasingly important …

Interpretable Rotation-Equivariant Multiary-Valued Network for Attribute Obfuscation

Quanshi Zhang, Hao Zhang, Yi-Ting Chen, Qihan Ren et autres

This paper focuses on the problem of preventing information leakage in neural networks, i.e., assuming that attackers have obtained intermediate-layer features of a neural network, and preventing attackers from inverting these features to the input with private information. We propose a generic …

cn (code pays fourni par la source)

Towards the first principles of explaining DNNs: interactions explain the learning dynamics

Huilin Zhou, Qihan Ren, Junpeng Zhang, Quanshi Zhang

Most explanation methods are designed in an empirical manner, so exploring whether there exists a first-principles explanation of a deep neural network (DNN) becomes the next core scientific problem in explainable artificial intelligence (XAI). Although it is still an open problem, in …

cn (code pays fourni par la source)

Benchmarking Multimodal Knowledge Conflict for Large Multimodal Models

Yifan Jia, Kailin Jiang, Y. T. Liang, Qihan Ren et autres

Large Multimodal Models(LMMs) face notable challenges when encountering multimodal knowledge conflicts, particularly under retrieval-augmented generation(RAG) frameworks where the contextual information from external sources may contradict the model's internal parametric knowledge, leading to unreliable outputs. However, existing benchmarks fail to reflect such realistic …

Alita: Generalist Agent Enabling Scalable Agentic Reasoning with Minimal Predefinition and Maximal Self-Evolution

J. F. Qiu, Qi Xuan, Tongcheng Zhang, Xinzhe Juan et autres

Recent advances in large language models (LLMs) have enabled agents to autonomously perform complex, open-ended tasks. However, many existing frameworks depend heavily on manually predefined tools and workflows, which hinder their adaptability, scalability, and generalization across domains. In this work, we introduce …

Revisiting Generalization Power of a DNN in Terms of Symbolic Interactions

Lei Cheng, Junpeng Zhang, Qihan Ren, Quanshi Zhang

This paper aims to analyze the generalization power of deep neural networks (DNNs) from the perspective of interactions. Unlike previous analysis of a DNN's generalization power in a highdimensional feature space, we find that the generalization power of a DNN can be …

Towards the Dynamics of a DNN Learning Symbolic Interactions

Qihan Ren, Junpeng Zhang, Xu Yang, Xin Yue et autres

This study proves the two-phase dynamics of a deep neural network (DNN) learning interactions. Despite the long disappointing view of the faithfulness of post-hoc explanation of a DNN, a series of theorems have been proven in recent years to show that for …

Interpretable Rotation-Equivariant Quaternion Neural Networks for 3D Point Cloud Processing

Wen Shen, Zhihua Wei, Qihan Ren, Binbin Zhang et autres

This study proposes a set of generic rules to revise existing neural networks for 3D point cloud processing to rotation-equivariant quaternion neural networks (REQNNs), in order to make feature representations of neural networks to be rotation-equivariant and permutation-invariant. Rotation equivariance of features …

cn (code pays fourni par la source)

Towards the Dynamics of a DNN Learning Symbolic Interactions

Qihan Ren, Junpeng Zhang, Yang Xu, Dongrui Liu et autres