33 citations · 101 across the 15 of their papers we have counts for
15 papers
TMPO: Trajectory Matching Policy Optimization for Diverse and Efficient Diffusion Alignment
Jiaming Li, Chenyu Zhu, Nanxi Yi +9
Reinforcement learning (RL) has shown extraordinary potential in aligning diffusion models to downstream tasks, yet most of them still suffer from significant reward hacking, which…
Position: The Turing-Completeness of Autoregressive Transformers Relies Heavily on Context Management
Guanyu Cui, Zhewei Wei, Kun He
Many works make the eye-catching claim that Transformers are Turing-complete. However, the literature often conflates two distinct settings: (i) a fixed Transformer system setting,…
KVSmooth: Mitigating Hallucination in Multi-modal Large Language Models through Key-Value Smoothing
Siyu Jiang, Feiyang Chen, Xiaojin Zhang +1
Despite the significant progress of Multimodal Large Language Models (MLLMs) across diverse tasks, hallucination -- corresponding to the generation of visually inconsistent objects…
ViT-EnsembleAttack: Augmenting Ensemble Models for Stronger Adversarial Transferability in Vision Transformers
Hanwen Cao, Haobo Lu, Xiaosen Wang +1
Ensemble-based attacks have been proven to be effective in enhancing adversarial transferability by aggregating the outputs of models with various architectures. However, existing…
VisCRA: A Visual Chain Reasoning Attack for Jailbreaking Multimodal Large Language Models
Bingrui Sima, Linhua Cong, Wenxuan Wang +1
The emergence of Multimodal Large Language Models (MLRMs) has enabled sophisticated visual reasoning capabilities by integrating reinforcement learning and Chain-of-Thought (CoT) s…
Refining Adaptive Zeroth-Order Optimization at Ease
Yao Shu, Qixin Zhang, Kun He +1
Recently, zeroth-order (ZO) optimization plays an essential role in scenarios where gradient information is inaccessible or unaffordable, such as black-box systems and resource-con…
Improving Adversarial Transferability with Scheduled Step Size and Dual Example
Zeliang Zhang, Peihan Liu, Xiaosen Wang +1
Deep neural networks are widely known to be vulnerable to adversarial examples, especially showing significantly poor performance on adversarial examples generated under the white-…
Triangle Attack: A Query-efficient Decision-based Adversarial Attack
Xiaosen Wang, Zeliang Zhang, Kangheng Tong +4
Decision-based attack poses a severe threat to real-world applications since it regards the target model as a black box and only accesses the hard prediction label. Great efforts h…
Multi-stage Optimization based Adversarial Training
Xiaosen Wang, Chuanbiao Song, Liwei Wang +1
In the field of adversarial robustness, there is a common practice that adopts the single-step adversarial training for quickly developing adversarially robust models. However, the…
Boosting Adversarial Transferability through Enhanced Momentum
Xiaosen Wang, Jiadong Lin, Han Hu +2
Deep learning models are known to be vulnerable to adversarial examples crafted by adding human-imperceptible perturbations on benign images. Many existing adversarial attack metho…
Enhancing the Transferability of Adversarial Attacks through Variance Tuning
Xiaosen Wang, Kun He
Deep neural networks are vulnerable to adversarial examples that mislead the models with imperceptible perturbations. Though adversarial attacks have achieved incredible success ra…
Admix: Enhancing the Transferability of Adversarial Attacks
Xiaosen Wang, Xuanran He, Jingdong Wang +1
Deep neural networks are known to be extremely vulnerable to adversarial examples under white-box setting. Moreover, the malicious adversaries crafted on the surrogate (source) mod…
Adversarial Training with Fast Gradient Projection Method against Synonym Substitution based Text Attacks
Xiaosen Wang, Yichen Yang, Yihe Deng +1
Adversarial training is the most empirically successful approach in improving the robustness of deep neural networks for image classification.For text classification, however, exis…
Natural Language Adversarial Defense through Synonym Encoding
Xiaosen Wang, Hao Jin, Yichen Yang +1
In the area of natural language processing, deep learning models are recently known to be vulnerable to various types of adversarial perturbations, but relatively few works are don…
A New Anchor Word Selection Method for the Separable Topic Discovery
Kun He, Wu Wang, Xiaosen Wang +1
Separable Non-negative Matrix Factorization (SNMF) is an important method for topic modeling, where "separable" assumes every topic contains at least one anchor word, defined as a…