RubiCap: Rubric-Guided Reinforcement Learning for Dense Image Captioning

Dimensions overall score 8.0

GitHub Code Pulse

No public code linked for this paper yet.

• RubiCap achieves the highest win rates on CapArena benchmark compared to supervised distillation, prior RL methods, human-expert annotations, and GPT-4V-augmented outputs.result 90%
• RubiCap demonstrates superior word efficiency on CaptionQA benchmark, with the 7B model matching Qwen2.5-VL-32B-Instruct performance.result 90%
• RubiCap-3B model surpasses the 7B counterpart of Qwen2.5-VL on CaptionQA benchmark.result 90%
• Using RubiCap-3B as a captioner produces stronger pretrained vision-language models than those trained on captions from proprietary models.result 85%
• RubiCap uses LLM-written rubrics to derive fine-grained, sample-specific reward signals for reinforcement learning in dense image captioning.method 90%
• RubiCap replaces coarse scalar rewards with structured, multi-faceted evaluations by decomposing holistic quality assessment.technical 85%
• Supervised distillation for synthetic captioning via vision-language models yields limited output diversity and weak generalization.limitation 80%
• Reinforcement learning successes in captioning have been limited by the lack of deterministic checkers available in verifiable domains.limitation 80%

Competitive landscape

Competitor map is still being generated for this paper. Enable generation or check back soon.

Startup potential card