# BalCapRL: A Balanced Framework for RL-Based MLLM Image Captioning

arXiv: https://arxiv.org/abs/2605.07394 · HTML: https://mlsanity.com/paper/2605.07394/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 23 of its 29 references that are in the map.
- First posted on arXiv 8 May 2026.

## Builds on (most cited first, 23 of 23)

- [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://mlsanity.com/paper/2402.03300/) (2024)
- [Improved Baselines with Visual Instruction Tuning](https://mlsanity.com/paper/2310.03744/) (2023)
- [CLIPScore: A Reference-free Evaluation Metric for Image Captioning](https://mlsanity.com/paper/2104.08718/) (2021)
- [HybridFlow: A Flexible and Efficient RLHF Framework](https://mlsanity.com/paper/2409.19256/) (2024)
- [Understanding R1-Zero-Like Training: A Critical Perspective](https://mlsanity.com/paper/2503.20783/) (2025)
- [SPICE: Semantic Propositional Image Caption Evaluation](https://mlsanity.com/paper/1607.08822/) (2016)
- [Aligning Large Multimodal Models with Factually Augmented RLHF](https://mlsanity.com/paper/2309.14525/) (2023)
- [Video-R1: Reinforcing Video Reasoning in MLLMs](https://mlsanity.com/paper/2503.21776/) (2025)
- [Rubrics as Rewards: Reinforcement Learning Beyond Verifiable Domains](https://mlsanity.com/paper/2507.17746/) (2025)
- [GDPO: Group reward-Decoupled Normalization Policy Optimization for Multi-reward RL Optimization](https://mlsanity.com/paper/2601.05242/) (2026)
- [Mastering Complex Control in MOBA Games with Deep Reinforcement Learning](https://mlsanity.com/paper/1912.09729/) (2019)
- [RLAIF-V: Open-Source AI Feedback Leads to Super GPT-4V Trustworthiness](https://mlsanity.com/paper/2405.17220/) (2024)
- [FaithScore: Fine-grained Evaluations of Hallucinations in Large Vision-Language Models](https://mlsanity.com/paper/2311.01477/) (2023)
- [CapArena: Benchmarking and Analyzing Detailed Image Captioning in the LLM Era](https://mlsanity.com/paper/2503.12329/) (2025)
- [CAPTURe: Evaluating Spatial Reasoning in Vision Language Models via Occluded Object Counting](https://mlsanity.com/paper/2504.15485/) (2025)
- [DLER: Doing Length pEnalty Right - Incentivizing More Intelligence per Token via Reinforcement Learning](https://mlsanity.com/paper/2510.15110/) (2025)
- [Prism: A Framework for Decoupling and Assessing the Capabilities of VLMs](https://mlsanity.com/paper/2406.14544/) (2024)
- [CAPability: A Comprehensive Visual Caption Benchmark for Evaluating Both Correctness and Thoroughness](https://mlsanity.com/paper/2502.14914/) (2025)
- [Painting with Words: Elevating Detailed Image Captioning with Benchmark and Alignment Learning](https://mlsanity.com/paper/2503.07906/) (2025)
- [SC-Captioner: Improving Image Captioning with Self-Correction by Reinforcement Learning](https://mlsanity.com/paper/2508.06125/) (2025)
- [CaptionQA: Is Your Caption as Useful as the Image Itself?](https://mlsanity.com/paper/2511.21025/) (2025)
- [RubiCap: Rubric-Guided Reinforcement Learning for Dense Image Captioning](https://mlsanity.com/paper/2603.09160/) (2026)
- [CCCaption: Dual-Reward Reinforcement Learning for Complete and Correct Image Captioning](https://mlsanity.com/paper/2602.21655/) (2026)

## Built on by (most cited first, 2 of 2)

- [Momentum-Coupled Rubric Adaptation for Detailed Image Captioning](https://mlsanity.com/paper/2609.36893/) (2026)
- [LineupRL: Verifiable Reinforcement Learning for Time Series Captioning via Caption-to-Series Identification](https://mlsanity.com/paper/2610.01800/) (2026)

## Abstract (arXiv)

Image captioning is one of the most fundamental tasks in computer vision. Owing to its open-ended nature, it has received significant attention in the era of multimodal large language models (MLLMs). In pursuit of ever more detailed and accurate captions, recent work has increasingly turned to reinforcement learning (RL). However, existing captioning-RL methods and evaluation metrics often emphasize a narrow notion of caption quality, inducing trade-offs across core dimensions of captioning. For example, utility-oriented objectives can encourage noisy, hallucinated, or overlong captions that improve downstream question answering while harming fluency, whereas arena-style objectives can favor fluent but generic descriptions with limited usefulness. To address this, we propose a more balanced RL framework that jointly optimizes utility-aware correctness, reference coverage, and linguistic quality. In order to effectively optimize the resulting continuous multi-objective reward formulation, we apply GDPO-style reward-decoupled normalization to continuous-valued captioning rewards and show that it improves performance over vanilla GRPO. Additionally, we introduce length-conditional reward masking, yielding a more suitable length penalty for captioning. Across LLaVA-1.5-7B and Qwen2.5-VL 3B and 7B base models, our method consistently improves caption quality, with peak gains of +13.6 DCScore, +9.0 CaptionQA, and +29.0 CapArena across different models.