MAPO improves multimodal chain-of-thought reasoning by requiring explicit textual descriptions of visual tool results and using a novel advantage estimator that combines semantic alignment with task rewards.
Varr(ˆgout|τ) =∥∇ θ logπ(τ)∥ 2 ·Var(r out) =C τ ·σ 2 out (21) Varr(ˆgsem|τ) =∥∇ θ logπ(τ)∥ 2 ·Var(r sem) =C τ ·σ 2 sem (22) whereC τ =∥∇ θ logπ(τ)∥ 2
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Walk the Talk: Bridging the Reasoning-Action Gap for Thinking with Images via Multimodal Agentic Policy Optimization
MAPO improves multimodal chain-of-thought reasoning by requiring explicit textual descriptions of visual tool results and using a novel advantage estimator that combines semantic alignment with task rewards.