Tongliang Liu

Tug-of-War No More: Harmonizing Accuracy and Robustness in Vision-Language Models via Stability-Aware Task Vector Merging

On the Thinking-Language Modeling Gap in Large Language Models

Memorize to Forget: Machine Unlearning without Gradient Ascent via Model Extrapolation

Mitigating Mismatch within Reference-based Preference Optimization

WRF4CIR: Weight-Regularized Fine-Tuning Network for Composed Image Retrieval

Concept Concentration for Faithful Representation Intervention

Unifying Stable Optimization and Reference Regularization in RLHF

Detecting Generated Images via Machine Unlearning

Learning Locally, Revising Globally: Global Reviser for Federated Learning with Noisy Labels

FORCE: Transferable Visual Jailbreaking Attacks via Feature Over-Reliance CorrEction

AdLift: Lifting Adversarial Perturbations to Safeguard 3D Gaussian Splatting Assets Against Instruction-Driven Editing

FRAG: Filtering Noise Using Snippet-Level Query Relevance

CoDaPO: Confidence and Difficulty-Adaptive Policy Optimization for Language Models

Causal Effect Identification in the Presence of Latent Confounding with a Single Imperfect Proxy Variable

A Closer Look at In-context Learning of LLMs in Simple Classification Tasks

JailbreakLoRA: Your Downloaded LoRA from Sharing Platforms might be Unsafe

R²AG: Learning to Reason, Retrieve, and Self-Evolve through a Multi-Branch Retrieval Tree

Do Vision-Language Models Reason Like Humans? Exploring the Functional Roles of Attention Heads

FATE: Feature-Wise Graph Attention with Multi-Period Temporal Encoding for Stock Return Forecasting

MaGA: Machine-Guided Amnesiac Unlearning through Target Feature Disentanglement

Lifelong Unlearning for Multimodal Large Language Models

Editing by Reconstruction: Background Preservation for Instruction-based Autoregressive Image Editing

Symmetry-Guaranteed Prediction of High-Order Tensor Properties for Crystalline Materials via Irreducible Decomposition

LeGIT: LLM Guided Intervention Targeting for Online Causal Discovery

REL-RAG: Relation-Aware Retrieval-Augmented Generation for Generalizable Knowledge Graph Question Answering

Plan before Solving: Problem-Aware Strategy Routing for Mathematical Reasoning

Forgetting: A New Mechanism Towards Better Large Language Model Fine-tuning

Combining-on-Graph: Structured Reasoning of LLMs with Knowledge Graph

From Static to Dynamic: Adaptive Monte Carlo Search for Mathematical Process Supervision

Reinforcement Learning with Verifiable yet Noisy Rewards under Imperfect Verifiers

Generative Model Inversion Through the Lens of the Manifold Hypothesis

RankMatch: A Novel Approach to Semi-Supervised Label Distribution Learning Leveraging Rank Correlation between Labels

Instance-dependent Early Stopping

Cognitive Mirrors: Exploring the Diverse Functional Roles of Attention Heads in LLM Reasoning

DEEM: Diffusion models serve as the eyes of large language models for image perception

Epistemic Uncertainty for Generated Image Detection

Pruning Spurious Subgraphs for Graph Out-of-Distribution Generalization

Revealing Multimodal Causality with Large Language Models

Chain-of-Focus Prompting: Leveraging Sequential Visual Cues to Prompt Large Autoregressive Vision Models

A Robust Method to Discover Causal or Anticausal Relation

Pruning Spurious Subgraphs for Graph Out-of-Distribtuion Generalization

A Sample Efficient Conditional Independence Test in the Presence of Discretization

Aligning What Matters: Masked Latent Adaptation for Text-to-Audio-Video Generation

MLLM-For3D: Adapting Multimodal Large Language Model for 3D Reasoning Segmentation

Flow: Modularized Agentic Workflow Automation

A Lens into Interpretable Transformer Mistakes via Semantic Dependency

From Debate to Equilibrium: Belief‑Driven Multi‑Agent LLM Reasoning via Bayesian Nash Equilibrium

Efficient and Trustworthy Causal Discovery with Latent Variables and Complex Relations

Recovery of Causal Graph Involving Latent Variables via Homologous Surrogates

Noisy Test-Time Adaptation in Vision-Language Models

Learning Graph Invariance by Harnessing Spuriosity

Understanding and Enhancing the Transferability of Jailbreaking Attacks

Towards Out-of-Modal Generalization without Instance-level Modal Correspondence

AgentAuditor: Human-level Safety and Security Evaluation for LLM Agents

Diversifying Spurious Subgraphs for Graph Out-of-Distribution Generalization

Enhancing Sample Selection Against Label Noise by Cutting Mislabeled Easy Examples

Surrogate Prompt Learning: Towards Efficient and Diverse Prompt Learning for Vision-Language Models

Towards Effective Evaluations and Comparisons for LLM Unlearning Methods

Can Dependencies Induced by LLM-Agent Workflows Be Trusted?

Detecting Discrepancies Between Generated and Natural Images Using Uncertainty

Exploring the Discriminative Capability of LLMs in In-context Learning

On the Language of Thoughts in Large Language Models

Exploring Criteria of Loss Reweighting to Enhance LLM Unlearning

Single-Step Diffusion Model-Based Generative Model Inversion Attacks

Ranked from Within: Ranking Large Multimodal Models Without Labels

Understanding Mistakes in Transformers through Token-level Semantic Dependencies

Nested Gloss Makes Large Language Models Lost

When Data-Free Knowledge Distillation Meets Non-Transferable Teacher: Escaping Out-of-Distribution Trap is All You Need

Can Large Language Models Help Experimental Design for Causal Discovery?

Revisiting the expressiveness of CNNs: a mathematical framework for feature extraction

Safeguarding System Prompts: A Surrogate-Based Defense Against Injection Attacks