Summary
This paper introduces a self-evolving training framework for unified large multimodal models (LMMs) that support both visual understanding and image generation, eliminating the need for curated post-training supervision such as human annotations, preference labels, or external reward models. The framework assigns three internal roles to the model: a Proposer that generates visual questions, a Solver that answers and evaluates them, and a Generator that produces images based on the questions. Through self-consistency rewards, the model iteratively improves its visual understanding and generation capabilities using only unlabeled images. The authors report significant improvements on both understanding and generation benchmarks, demonstrating that self-evolution is a viable path for unified multimodal models. The work was published on arXiv (2606.27376) by Ritesh Thawkar, Shravan Venkatraman, and Omkar Thawkar in the computer vision field.
Paper Overview
Research Area: Computer Vision (CV)
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. The authors ask whether a unified LMM can improve both abilities autonomously using only unlabeled images.
They propose a self-evolving training framework with three internal roles:
- Proposer: generates visual questions
- Solver: answers and evaluates them
- Generator: generates images based on the questions
Through self-consistency rewards, the model iteratively improves its visual understanding and generation capabilities without external supervision. Experiments show significant improvements on both understanding and generation tasks, demonstrating that self-evolution is a viable path for unified multimodal models.
---
*Auto-collected on 2026-06-27*
This page is an English static mirror generated for search and AI citation.
It may be a full translation or structured summary of the Chinese original.
Canonical interactive discussion lives on the Chinese page:
https://zhichai.net/topic/178208182