Paper Overview
Field: Computer Vision Authors: Ritesh Thawkar, Shravan Venkatraman, Omkar Thawkar Published: 2026-06-27 arXiv: 2606.27376
Abstract
Most unified large multimodal models (LMMs) that support both visual understanding and image generation still rely on curated post-training supervision, such as human annotations, preference labels, or external reward models. We ask whether a unified LMM can improve both abilities autonomously using only unlabeled images. We propose a self-evolving training framework with three internal roles: a Proposer that generates visual questions, a Solver that answers and evaluates them, and a Generator that synthesizes images. Training uses only self-derived consistency signals, without human annotations, preference labels, or task-trained external reward/judge models.
To stabilize learning, we introduce Solver Token Entropy (STE), a continuous difficulty signal based on token-level prediction uncertainty. For image generation, we design a multi-scale internal evaluation scheme that combines question-answering fidelity scoring with cycle-consistency captioning.
The framework is applied to BLIP3o, BAGEL, and VARGPT-v1.1 architectures while keeping the same role decomposition and reward logic. Across eight understanding benchmarks, our method consistently outperforms the base models. On BAGEL, we achieve a +3.5% absolute gain on MMMU and improve GenEval image generation performance from 82% to 85%.
---
*Auto-collected on 2026-06-27*