# Self-Evolving Visual Questioner

arXiv: https://arxiv.org/abs/2606.13929 · HTML: https://mlsanity.com/paper/2606.13929/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in VLM benchmarks and visual reasoning, one of 16 directions in Multimodal LLMs / vision-language models.
- Builds on 16 of its 28 references that are in the map.
- First posted on arXiv 11 Jun 2026.

## Placed in

- Multimodal LLMs / vision-language models / [VLM benchmarks and visual reasoning](https://mlsanity.com/mllm/vlm-benchmarks-and-visual-reasoning/)
- Synthetic data and self-improvement for LLMs / [Self-improving MLLMs and synthetic data](https://mlsanity.com/synth/self-improving-mllms-and-synthetic-data/)

## Builds on (most cited first, 16 of 16)

- [Cambrian-1: A Fully Open, Vision-Centric Exploration of Multimodal LLMs](https://mlsanity.com/paper/2406.16860/) (2024)
- [V*: Guided Visual Search as a Core Mechanism in Multimodal LLMs](https://mlsanity.com/paper/2312.14135/) (2023)
- [MLLM-as-a-Judge: Assessing Multimodal LLM-as-a-Judge with Vision-Language Benchmark](https://mlsanity.com/paper/2402.04788/) (2024)
- [From Flatland to Space: Teaching Vision-Language Models to Perceive and Reason in 3D](https://mlsanity.com/paper/2503.22976/) (2025)
- [Enhancing Large Vision Language Models with Self-Training on Image Comprehension](https://mlsanity.com/paper/2405.19716/) (2024)
- [Vision-Language Models Can Self-Improve Reasoning via Reflection](https://mlsanity.com/paper/2411.00855/) (2024)
- [LOVA3: Learning to Visual Question Answering, Asking and Assessment](https://mlsanity.com/paper/2405.14974/) (2024)
- [Guiding Visual Question Generation](https://mlsanity.com/paper/2110.08226/) (2021)
- [Through the Lens of Contrast: Self-Improving Visual Reasoning in VLMs](https://mlsanity.com/paper/2603.02556/) (2026)
- [K-VQG: Knowledge-aware Visual Question Generation for Common-sense Acquisition](https://mlsanity.com/paper/2203.07890/) (2022)
- [Learning to Instruct for Visual Instruction Tuning](https://mlsanity.com/paper/2503.22215/) (2025)
- [Savaal: Scalable Concept-Driven Question Generation to Enhance Human Learning](https://mlsanity.com/paper/2502.12477/) (2025)
- [CYCLE-INSTRUCT: Fully Seed-Free Instruction Tuning via Dual Self-Training and Cycle Consistency](https://mlsanity.com/paper/2508.16100/) (2025)
- [Location-Aware Visual Question Generation with Lightweight Models](https://mlsanity.com/paper/2310.15129/) (2023)
- [ConVQG: Contrastive Visual Question Generation with Multimodal Guidance](https://mlsanity.com/paper/2402.12846/) (2024)
- [Ask Questions with Double Hints: Visual Question Generation with Answer-awareness and Region-reference](https://mlsanity.com/paper/2407.05100/) (2024)

## Built on by (most cited first, 1 of 1)

- [On-Policy Self-Distillation without Any Supervision](https://mlsanity.com/paper/2608.06296/) (2026)

## Abstract (arXiv)

Vision-language models (VLMs) are typically trained as passive answerers, while their ability to actively ask diverse, non-trivial, visual-centric and grounded questions remains underexplored. Existing visual questioners' performance is bottlenecked by the availability of high-quality training data or the cost of curating them. We show that a VLM can continuously improve itself as a visual questioner without any external supervision. We propose a self-evolving framework that uses a VLM itself as both a proposer and a filter to produce harder, more informative, and visual-centric questions, while maintaining their exploration diversity to avoid training collapse. These questions are then used to train the VLM in both questioner and answerer modes. To evaluate the questioner, we introduce an agentic protocol that assesses questions along perception, reasoning, and diversity dimensions. Experiments across various backbone VLMs show that our method substantially enhances the quality and substantially expands the difficulty boundary of autonomous question generation. Under the same budget, our self-supervision is more effective than training on the static source data. Moreover, the self-evolving questioner remains a competitive or even better answerer.