影响力指数

63.4/100

前 3.9%

全站排名 #2,504

发表论文11 篇

平均评分5.7

年均产出3.7 篇/年

Shuo He

Postdoc@Nanyang Technological University·新加坡·OpenReview

研究方向

Weakly-supervised machine learning · Security in multimodal large language models · RL training for LLM-based agents

6.5

Improving Generalizability and Undetectability for Targeted Adversarial Attacks on Multimodal Pre-trained Models

ICLR 2026Withdrawn

7.2

Representation Surgery in Model Merging with Probabilistic Modeling

ICML 2025Poster

二作

6.3

A Closer Look at Backdoor Attacks on CLIP

ICML 2025Poster

一作

6.1

Test-Time Multimodal Backdoor Detection by Contrastive Prompting

ICML 2025Poster

二作

6.0

Defending Multimodal Backdoored Models by Repulsive Visual Prompt Tuning

NeurIPS 2025Poster

二作

5.8

BDetCLIP: Multimodal Prompting Contrastive Test-Time Backdoor Detection

ICLR 2025Rejected

二作

5.3

A Closer Look at Backdoor Attacks on CLIP

合作者 (20)

Shuo He

Hierarchy-of-Groups Policy Optimization for Long-Horizon Agentic Tasks

Solving Puzzles? Jailbreaking Multimodal Large Language Models!

Reasoning-Preserved Safety Alignment for Large Reasoning Models

Improving Generalizability and Undetectability for Targeted Adversarial Attacks on Multimodal Pre-trained Models

Representation Surgery in Model Merging with Probabilistic Modeling

A Closer Look at Backdoor Attacks on CLIP

Test-Time Multimodal Backdoor Detection by Contrastive Prompting

Defending Multimodal Backdoored Models by Repulsive Visual Prompt Tuning

BDetCLIP: Multimodal Prompting Contrastive Test-Time Backdoor Detection

A Closer Look at Backdoor Attacks on CLIP