15 papers
Are Vision-Language-Action Models Robust to One-Step Observation Perturbations?
Shojiro Yamabe, Jun Sakuma
Understanding the safety risks of vision-language-action (VLA) models is essential for their deployment in the physical world. Existing safety research has mainly considered persis…
Capability-Routed Guard: Defending Large Reasoning Models Against Reasoning-Centric Jailbreaks
Yiyong Liu, Yixin Wu, Jun Sakuma
Large reasoning models (LRMs) expose a new safety failure mode: adversarial prompts can manipulate reasoning context, task decomposition, or capability interpretation so that harmf…
When Benchmarks Leak: Inference-Time Decontamination for LLMs
Jianzhe Chai, Yu Zhe, Jun Sakuma
Benchmark-based evaluation is the de facto standard for comparing large language models (LLMs). However, its reliability is increasingly threatened by test set contamination, where…
Model-agnostic Adversarial Attack and Defense for Vision-Language-Action Models
Haochuan Xu, Yun Sing Koh, Shuhuai Huang +4
Vision-Language-Action (VLA) models have achieved revolutionary progress in robot learning, enabling robots to execute complex physical robot tasks from natural language instructio…
Pattern Enhanced Multi-Turn Jailbreaking: Exploiting Structural Vulnerabilities in Large Language Models
Ragib Amin Nihal, Rui Wen, Kazuhiro Nakadai +1
Large language models (LLMs) remain vulnerable to multi-turn jailbreaking attacks that exploit conversational context to bypass safety constraints gradually. These attacks target d…
Differentially Private Synthetic Text Generation for Retrieval-Augmented Generation (RAG)
Junki Mori, Kazuya Kakizaki, Taiki Miyagawa +1
Retrieval-Augmented Generation (RAG) enhances large language models (LLMs) by grounding them in external knowledge. However, its application in sensitive domains is limited by priv…
Toward Safer Diffusion Language Models: Discovery and Mitigation of Priming Vulnerability
Shojiro Yamabe, Jun Sakuma
Diffusion language models (DLMs) generate tokens in parallel through iterative denoising, which can reduce latency and enable bidirectional conditioning. However, the safety risks…
Data Poisoning Attacks to Locally Differentially Private Range Query Protocols
Ting-Wei Liao, Chih-Hsun Lin, Yu-Lin Tsai +5
Local Differential Privacy (LDP) has been widely adopted to protect user privacy in decentralized data collection. However, recent studies have revealed that LDP protocols are vuln…
Disrupting Model Merging: A Parameter-Level Defense Without Sacrificing Accuracy
Wei Junhao, Yu Zhe, Sakuma Jun
Model merging is a technique that combines multiple finetuned models into a single model without additional training, allowing a free-rider to cheaply inherit specialized capabilit…
BADTV: Unveiling Backdoor Threats in Third-Party Task Vectors
Chia-Yi Hsu, Yu-Lin Tsai, Yu Zhe +6
Task arithmetic in large-scale pre-trained models enables agile adaptation to diverse downstream tasks without extensive retraining. By leveraging task vectors (TVs), users can per…
Beyond Full Poisoning: Effective Availability Attacks with Partial Perturbation
Yu Zhe, Jun Sakuma
The widespread use of publicly available datasets for training machine learning models raises significant concerns about data misuse. Availability attacks have emerged as a means f…
Zero-shot domain adaptation based on dual-level mix and contrast
Yu Zhe, Jun Sakuma
Zero-shot domain adaptation (ZSDA) is a domain adaptation problem in the situation that labeled samples for a target task (task of interest) are only available from the source doma…
Robust Deep Reinforcement Learning against Adversarial Behavior Manipulation
Shojiro Yamabe, Kazuto Fukuchi, Jun Sakuma
This study investigates behavior-targeted attacks on reinforcement learning and their countermeasures. Behavior-targeted attacks aim to manipulate the victim's behavior as desired…
Adversarial Attacks on Hidden Tasks in Multi-Task Learning
Yu Zhe, Rei Nagaike, Daiki Nishiyama +2
Deep learning models are susceptible to adversarial attacks, where slight perturbations to input data lead to misclassification. Adversarial attacks become increasingly effective w…
Harnessing the Power of Vicinity-Informed Analysis for Classification under Covariate Shift
Mitsuhiro Fujikawa, Yohei Akimoto, Jun Sakuma +1
Transfer learning enhances prediction accuracy on a target distribution by leveraging data from a source distribution, demonstrating significant benefits in various applications. T…