[KongchangAI]
Concept多模态思维链 / Multimodal Chain-of-Thought

M-CoT

A dataset or method that extends chain-of-thought reasoning to multimodal settings, containing step-by-step reasoning paths spanning vision and text. It can incorporate error-awareness information to strengthen a model's understanding of perceptual error patterns.

Timeline (last 90 days)

Nov 6

PSGAgent adopts a two-stage training pipeline: the first stage builds an error-aware multimodal chain-of-thought (M-CoT) dataset for cold-start learning; the second stage uses a hierarchical reward mechanism combined with group sequence policy optimization (GSPO) to learn the tool-use policy.

Unverified50%

All Facts (1)

Source Articles