Importance Sampling for Multi-Negative Multimodal Direct Preference Optimization

Documenti analoghi

• SceneAlign: Aligning Multimodal Reasoning to Scene Graphs in Complex Visual Scenes
  di: Wang, Chuhan, et al.
  Pubblicazione: (2026)
• CoMMIT: Coordinated Multimodal Instruction Tuning
  di: Li, Xintong, et al.
  Pubblicazione: (2024)
• MASS-DPO: Multi-negative Active Sample Selection for Direct Policy Optimization
  di: Surana, Rohan, et al.
  Pubblicazione: (2026)
• Explainable Chain-of-Thought Reasoning: An Empirical Analysis on State-Aware Reasoning Dynamics
  di: Yu, Sheldon, et al.
  Pubblicazione: (2025)
• Active Learning for Direct Preference Optimization
  di: Kveton, Branislav, et al.
  Pubblicazione: (2025)