# Latent Chain-of-Thought for Visual Reasoning

arXiv: https://arxiv.org/abs/2510.23925 · HTML: https://mlsanity.com/paper/2510.23925/

## Key facts

- Cited by 29 papers in the map (data through 14 Oct 2026).
- Placed in Multimodal reasoning, one of 17 directions in LLM reasoning.
- Builds on 7 of its 58 references that are in the map.
- First posted on arXiv 27 Oct 2025.

## Placed in

- LLM reasoning / [Multimodal reasoning](https://mlsanity.com/lr/multimodal-reasoning/)
- Multimodal LLMs / vision-language models / [Embodied and 3D vision-language models](https://mlsanity.com/mllm/embodied-and-3d-vision-language-models/)

## Builds on (most cited first, 7 of 7)

- [GPT-4 Technical Report](https://mlsanity.com/paper/2303.08774/) (2023)
- [Proximal Policy Optimization Algorithms](https://mlsanity.com/paper/1707.06347/) (2017)
- [DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning](https://mlsanity.com/paper/2501.12948/) (2025)
- [Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolution](https://mlsanity.com/paper/2409.12191/) (2024)
- [OpenAI o1 System Card](https://mlsanity.com/paper/2412.16720/) (2024)
- [Multimodal Chain-of-Thought Reasoning in Language Models](https://mlsanity.com/paper/2302.00923/) (2023)
- [MC-CoT: A Modular Collaborative CoT Framework for Zero-shot Medical-VQA with LLM and MLLM Integration](https://mlsanity.com/paper/2410.04521/) (2024)

## Built on by (most cited first, 29 of 29)

- [The Latent Space: Foundation, Evolution, Mechanism, Ability, and Outlook](https://mlsanity.com/paper/2604.02029/) (2026)
- [VisMem: Latent Vision Memory Unlocks Potential of Vision-Language Models](https://mlsanity.com/paper/2511.11007/) (2025)
- [Reasoning Within the Mind: Dynamic Multimodal Interleaving in Latent Space](https://mlsanity.com/paper/2512.12623/) (2025)
- [Latent Reasoning VLA: Latent Thinking and Prediction for Vision-Language-Action Models](https://mlsanity.com/paper/2602.01166/) (2026)
- [MIRA: Multimodal Iterative Reasoning Agent for Image Editing](https://mlsanity.com/paper/2511.21087/) (2025)
- [Latent Chain-of-Thought World Modeling for End-to-End Driving](https://mlsanity.com/paper/2512.10226/) (2025)
- [Thinking in Uncertainty: Mitigating Hallucinations in MLRMs with Latent Entropy-Aware Decoding](https://mlsanity.com/paper/2603.13366/) (2026)
- [MementoGUI: Learning Agentic Multimodal Memory Control for Long-Horizon GUI Agents](https://mlsanity.com/paper/2605.18652/) (2026)
- [Residual Context Diffusion Language Models](https://mlsanity.com/paper/2601.22954/) (2026)
- [Dual Latent Memory for Visual Multi-agent System](https://mlsanity.com/paper/2602.00471/) (2026)
- [ChartNet: A Million-Scale, High-Quality Multimodal Dataset for Robust Chart Understanding](https://mlsanity.com/paper/2603.27064/) (2026)
- [Toward Cognitive Supersensing in Multimodal Large Language Model](https://mlsanity.com/paper/2602.01541/) (2026)
- [Video-R4: Reinforcing Text-Rich Video Reasoning with Visual Rumination](https://mlsanity.com/paper/2511.17490/) (2025)
- [Towards Explainable Industrial Anomaly Detection via Knowledge-Guided Latent Reasoning](https://mlsanity.com/paper/2602.09850/) (2026)
- [The Perceptual Bandwidth Bottleneck in Vision-Language Models: Active Visual Reasoning via Sequential Experimental Design](https://mlsanity.com/paper/2605.01345/) (2026)
- [Latent Thought Flow: Efficient Latent Reasoning in Large Language Models](https://mlsanity.com/paper/2606.16222/) (2026)
- [What Makes Effective Supervision in Latent Chain-of-Thought: An Information-Theoretic Analysis](https://mlsanity.com/paper/2606.20075/) (2026)
- [CoLT: Teaching Multi-Modal Models to Think with Chain of Latent Thoughts](https://mlsanity.com/paper/2606.31986/) (2026)
- [DAVE: A VLM Vision Encoder for Document Understanding and Web Agents](https://mlsanity.com/paper/2512.17221/) (2025)
- [ReGuLaR: Relation-Grounded Latent Reasoning for Large Vision-Language Models](https://mlsanity.com/paper/2605.30587/) (2026)
- [Agent Skills Should Go Beyond Text: The Case for Visual Skills](https://mlsanity.com/paper/2606.01414/) (2026)
- [VideoSearch-R1: Iterative Video Retrieval and Reasoning via Soft Query Refinement](https://mlsanity.com/paper/2607.00446/) (2026)
- [LatentGeo: Learnable Auxiliary Constructions in Latent Space for Multimodal Geometric Reasoning](https://mlsanity.com/paper/2603.12166/) (2026)
- [Reflection Anchors for Propagation-Aware Visual Retention in Long-Chain Multimodal Reasoning](https://mlsanity.com/paper/2605.09614/) (2026)
- [Information-Regularized Attention for Visual-Centric Reasoning](https://mlsanity.com/paper/2607.00434/) (2026)
- [Latent-Centroid Steering: Single-Pass Classifier-Free Guidance for Command-Aligned Autonomous Driving](https://mlsanity.com/paper/2608.00237/) (2026)
- [On the Token Value Inequality in Efficient Reasoning](https://mlsanity.com/paper/2609.33970/) (2026)
- [Soft Spatial Reasoning](https://mlsanity.com/paper/2609.38717/) (2026)
- [SLVR: Structured Latent Visual Reasoning via Human-like Reasoning Flows](https://mlsanity.com/paper/2610.10563/) (2026)

## Abstract (arXiv)

Chain-of-thought (CoT) reasoning is critical for improving the interpretability and reliability of Large Vision-Language Models (LVLMs). However, existing training algorithms such as SFT, PPO, and GRPO may not generalize well across unseen reasoning tasks and heavily rely on a biased reward model. To address this challenge, we reformulate reasoning in LVLMs as posterior inference and propose a scalable training algorithm based on amortized variational inference. By leveraging diversity-seeking reinforcement learning algorithms, we introduce a novel sparse reward function for token-level learning signals that encourage diverse, high-likelihood latent CoT, overcoming deterministic sampling limitations and avoiding reward hacking. Additionally, we implement a Bayesian inference-scaling strategy that replaces costly Best-of-N and Beam Search with a marginal likelihood to efficiently rank optimal rationales and answers. We empirically demonstrate that the proposed method enhances the state-of-the-art LVLMs on seven reasoning benchmarks, in terms of effectiveness, generalization, and interpretability.