用多维度评分提升多能力模型训练效果
SkillRater: Untangling Capabilities in Multimodal Data
- 为每项能力设专用评分器，通过元学习分别训练
- 在多个任务上提升2%-5.6%，且评分信号基本独立
- 适合需要多种能力协同的视觉语言模型训练
数据筛选通常只给样本一个单一质量分数，但我们认为这种标量框架本质受限：当训练需多种不同能力时，单一评分器无法同时最大化所有能力的信号。质量应被视为多维的，每个维度对应模型需掌握的一种能力。我们提出SkillRater框架，将数据过滤分解为针对每种能力的专用评分器，每个评分器通过元学习在互不重叠的验证目标上训练，并通过渐进式选择规则组合得分：在每个训练阶段，只要任一评分器将样本评分超过逐渐收紧的阈值，该样本即被保留。早期保留多样性，后期聚焦高价值样本。我们在视觉语言模型上验证此方法，将质量分解为视觉理解、OCR和STEM推理三个维度。在20亿参数规模下，SkillRater相比未过滤基线，在视觉理解上提升5.63%，在OCR上提升2.00%，在STEM上提升3.53%。学习到的评分信号近乎正交，证实分解捕捉了真正独立的质量维度，解释了其为何优于未过滤训练和单体学习筛选。
原文摘要 · Abstract (English)
Data curation methods typically assign samples a single quality score. We argue this scalar framing is fundamentally limited: when training requires multiple distinct capabilities, a monolithic scorer cannot maximize useful signals for all of them simultaneously. Quality is better understood as multidimensional, with each dimension corresponding to a capability the model must acquire. We introduce SkillRater, a framework that decomposes data filtering into specialized raters - one per capability, each trained via meta-learning on a disjoint validation objective - and composes their scores through a progressive selection rule: at each training stage, a sample is retained if any rater ranks it above a threshold that tightens over time, preserving diversity early while concentrating on high-value samples late. We validate this approach on vision language models, decomposing quality into three capability dimensions: visual understanding, OCR, and STEM reasoning. At 2B parameters, SkillRater improves over unfiltered baselines by 5.63% on visual understanding, 2.00% on OCR, and 3.53% on STEM on held out benchmarks. The learned rater signals are near orthogonal, confirming that the decomposition captures genuinely independent quality dimensions and explaining why it outperforms both unfiltered training and monolithic learned filtering.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。