Rongzhe Wei

Long-Horizon Plan Execution in Large Tool Spaces through Entropy-Guided Branching

Rongzhe Wei, Ge Shi, Min Cheng, Na Zhang et autres

Large Language Models (LLMs) have significantly advanced tool-augmented agents, enabling autonomous reasoning via API interactions. However, executing multi-step tasks within massive tool libraries remains challenging due to two critical bottlenecks: (1) the absence of rigorous, plan-level evaluation frameworks and (2) the computational …

Long-Horizon Plan Execution in Large Tool Spaces through Entropy-Guided Branching

Rongzhe Wei, Ge Shi, Min Cheng, Na Zhang et autres

Large Language Models (LLMs) have significantly advanced tool-augmented agents, enabling autonomous reasoning via API interactions. However, executing multi-step tasks within massive tool libraries remains challenging due to two critical bottlenecks: (1) the absence of rigorous, plan-level evaluation frameworks and (2) the computational …

us (code pays fourni par la source)

The Trojan Knowledge: Bypassing Commercial LLM Guardrails via Harmless Prompt Weaving and Adaptive Tree Search

Rongzhe Wei, Peizhi Niu, Xinjie Shen, Tiankai Tu et autres

Large language models (LLMs) remain vulnerable to jailbreak attacks that bypass safety guardrails to elicit harmful outputs. Existing approaches overwhelmingly operate within the prompt-optimization paradigm: whether through traditional algorithmic search or recent agent-based workflows, the resulting prompts typically retain malicious semantic signals …

Model Generalization on Text Attribute Graphs: Principles with Large Language Models

Haoyu Wang, Shikun Liu, Rongzhe Wei, Pan Li

Large language models (LLMs) have recently been introduced to graph learning, aiming to extend their zero-shot generalization success to tasks where labeled graph data is scarce. Among these applications, inference over text-attributed graphs (TAGs) presents unique challenges: existing methods struggle with LLMs' …

Do LLMs Really Forget? Evaluating Unlearning with Knowledge Correlation and Confidence Awareness

Rongzhe Wei, Peizhi Niu, Ruihan Wu, Haoteng Yin et autres

Machine unlearning techniques aim to mitigate unintended memorization in large language models (LLMs). However, existing approaches predominantly focus on the explicit removal of isolated facts, often overlooking latent inferential dependencies and the non-deterministic nature of knowledge within LLMs. Consequently, facts presumed forgotten …

us, cn (code pays fourni par la source)

Differentially Private Relational Learning with Entity-level Privacy Guarantees

Yinan Huang, Haoteng Yin, Eli Chien, Rongzhe Wei et autres

Learning with relational and network-structured data is increasingly vital in sensitive domains where protecting the privacy of individual entities is paramount. Differential Privacy (DP) offers a principled approach for quantifying privacy risks, with DP-SGD emerging as a standard mechanism for private model …

us (code pays fourni par la source)

Underestimated Privacy Risks for Minority Populations in Large Language Model Unlearning

Rongzhe Wei, Michelle Li, Mohsen Ghassemi, Eleonora Kreačić et autres

Large Language Models (LLMs) embed sensitive, human-generated data, prompting the need for unlearning methods. Although certified unlearning offers strong privacy guarantees, its restrictive assumptions make it unsuitable for LLMs, giving rise to various heuristic approaches typically assessed through empirical evaluations. These standard …

Privately Learning from Graphs with Applications in Fine-tuning Large Language Models

Haoteng Yin, Rongzhe Wei, Eli Chien, Pan Li

Graphs offer unique insights into relationships between entities, complementing data modalities like text and images and enabling AI models to extend their capabilities beyond traditional tasks. However, learning from graphs often involves handling sensitive relationships in the data, raising significant privacy concerns. …

SLA2P: Self-Supervised Anomaly Detection With Adversarial Perturbation

Yizhou Wang, Can Qin, Rongzhe Wei, Yi Xu et autres

Anomaly detection is a foundational yet difficult problem in machine learning. In this work, we propose a new and effective framework, dubbed as SLA2P, for unsupervised anomaly detection. Following the extraction of delegate embeddings from raw data, we implement random projections on …

us (code pays fourni par la source)

Differentially Private Graph Diffusion with Applications in Personalized PageRanks

Rongzhe Wei, Eli Chien, Pan Li

Graph diffusion, which iteratively propagates real-valued substances among the graph, is used in numerous graph/network-involved applications. However, releasing diffusion vectors may reveal sensitive linking information in the data such as transaction information in financial network data. However, protecting the privacy of graph …

Guarding Multiple Secrets: Enhanced Summary Statistic Privacy for Data Sharing

Shuaiqi Wang, Rongzhe Wei, Mohsen Ghassemi, Eleonora Kreačić et autres

Data sharing enables critical advances in many research areas and business applications, but it may lead to inadvertent disclosure of sensitive summary statistics (e.g., means or quantiles). Existing literature only focuses on protecting a single confidential quantity, while in practice, data sharing …

Learning Scalable Structural Representations for Link Prediction with Bloom Signatures

Tianyi Zhang, Haoteng Yin, Rongzhe Wei, Pan Li et autres

Graph neural networks (GNNs) have shown great potential in learning on graphs, but they are known to perform sub-optimally on link prediction tasks. Existing GNNs are primarily designed to learn node-wise representations and usually fail to capture pairwise relations between target nodes, …

us (code pays fourni par la source)