Visual Backdoor Attacks on MLLM Embodied Decision Making via Contrastive Trigger Learning
Anzeige
Ähnliche Artikel
arXiv – cs.LG
•
SmoothGuard: Defending Multimodal Large Language Models with Noise Perturbation and Clustering Aggregation
arXiv – cs.AI
•
DeepCompress: A Dual Reward Strategy for Dynamically Exploring and Compressing Reasoning Chains
arXiv – cs.AI
•
BMGQ: A Bottom-up Method for Generating Complex Multi-hop Reasoning Questions from Semi-structured Data
arXiv – cs.AI
•
ssToken: Self-modulated and Semantic-aware Token Selection for LLM Fine-tuning
arXiv – cs.LG
•
UniRL-Zero: Reinforcement Learning on Unified Models with Joint Language Model and Diffusion Model Experts
arXiv – cs.AI
•
VERITAS: Leveraging Vision Priors and Expert Fusion to Improve Multimodal Data