用于视觉微调的高质量数据游戏化众包
Gamified crowd-sourcing of high-quality data for visual fine-tuning
AI总结:
本研究提出游戏化对抗提示框架GAP,通过游戏化众包从5万余参与者收集针对模型知识缺口的高质量视觉微调数据,显著提升了多款多模态模型的性能,接近GPT-4V基准水平。
摘要:
本文提出了Gamified Adversarial Prompting(GAP,游戏化对抗提示)框架,用于为大型多模态模型的视觉指令微调众包高质量数据。GAP将数据收集过程转化为具有吸引力的游戏,激励玩家提供针对模型知识缺口的细粒度、高难度问答对。我们的贡献包括:(1)一种从人类获取直接针对模型知识薄弱点的问答对的方法;(2)一种评估并奖励玩家的机制,可有效激励玩家提交高质量内容;(3)一个可扩展的游戏化平台,仅用数周就成功从超过50000名参与者处收集到这类数据。我们实现的GAP显著提升了小型多模态模型MiniCPM-Llama3-V-2.5-8B的准确率,使其在我们的数据集上的GPT得分从0.147提升至0.477,接近规模大得多的GPT-4V所设的基准。此外,我们证明用MiniCPM-Llama3-V-2.5-8B生成的数据也提升了该模型在其他基准上的表现,并且展现出跨模型增益:具体而言,相同数据提升了QWEN2-VL-2B和QWEN2-VL-7B在相同多个基准上的性能。
英文摘要:
This paper introduces Gamified Adversarial Prompting (GAP), a framework that crowd-sources high-quality data for visual instruction tuning of large multimodal models. GAP transforms the data collection process into an engaging game, incentivizing players to provide fine-grained, challenging questions and answers that target gaps in the model's knowledge. Our contributions include (1) an approach to capture question-answer pairs from humans that directly address weaknesses in a model's knowledge, (2) a method for evaluating and rewarding players that successfully incentivizes them to provide high-quality submissions, and (3) a scalable, gamified platform that succeeds in collecting this data from over 50,000 participants in just a few weeks. Our implementation of GAP has significantly improved the accuracy of a small multimodal model, namely MiniCPM-Llama3-V-2.5-8B, increasing its GPT score from 0.147 to 0.477 on our dataset, approaching the benchmark set by the much larger GPT-4V. Moreover, we demonstrate that the data generated using MiniCPM-Llama3-V-2.5-8B also enhances its performance across other benchmarks, and exhibits cross-model benefits. Specifically, the same data improves the performance of QWEN2-VL-2B and QWEN2-VL-7B on the same multiple benchmarks.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。