# Latent Visual Contrastive Reasoning with Multi-Branch Exploration

Preprint · Kegeng Tang, Jingbo Wang, Yangming Zhang, Huanshu Zhang, Shaogang Ren, Zihao Wang

- Archive id: arxiv:2610.36809
- Published: 2026-10-27
- arXiv: https://arxiv.org/abs/2610.36809
- Citations: 0
- Source URL: https://arxiv.org/abs/2610.36809
- PDF: https://arxiv.org/pdf/2610.36809
- Provenance: arxiv

## Abstract

Multimodal large language models have made substantial progress in visual question answering, but complex visual reasoning often requires locating and comparing multiple pieces of question-relevant visual evidence. Existing methods that move intermediate reasoning into the visual latent space usually follow a single linear reasoning trajectory, which can be limiting when multiple visual hypotheses or competing evidence regions are relevant. We propose Latent Visual Contrastive Reasoning, a branch-aware latent reasoning framework that generates multiple question-conditioned latent visual branches as soft hypotheses over image tokens. To train these branches, our method uses bounding-box evidence annotations to build target evidence representations and introduces a soft contrastive branch-evidence alignment objective. This objective encourages evidence-relevant branches to align with annotated visual regions while contrasting them against negative evidence. We instantiate our method on Qwen2.5-VL backbones and train it with standard supervised fine-tuning. Experiments on perception-intensive visual reasoning benchmarks show that the proposed method achieves overall improvements over the corresponding backbone and single-path latent reasoning baselines at both 3B and 7B scales.

_Abstract reproduced under CC0-1.0 (arXiv descriptive metadata)._

Topics: [Generative & Multimodal Models](https://space-kitty.com/topics/generative-multimodal)

---
Space Kitty Archive · https://space-kitty.com · API: https://space-kitty.com/api/v1 · MCP: https://space-kitty.com/mcp · Premium: $0.20/request via x402 (Base Sepolia (testnet)). Metadata: arXiv (CC0), OpenAlex (CC0), Europe PMC (CC BY/CC0 works), ClinicalTrials.gov.