提出首个9类任务分类体系，诊断Transformer在抽象推理中的组合能力短板。
A Neural Affinity Framework for Abstract Reasoning: Diagnosing the Compositional Gap in Transformer Architectures via Procedural Task Taxonomy
- 构建9类任务分类体系，通过规则代码分析验证准确率达97.5%
- 发现69.5%任务中模型局部准确但全局合成失败，存在神经亲和力天花板
- 适用于诊断模型架构适配性，尤其对高阶抽象推理任务优化有指导意义
针对Hodel等人（2024）关于re-arc任务相关性的形式化定义呼吁，我们首次构建了涵盖全部400个任务的9类别分类体系，并通过规则代码分析验证其准确率达97.5%。通过在原始网格像素上训练CNN（S3任务达95.24%准确率，整体36.25%，为随机水平的3.3倍），证实该分类体系具有视觉一致性。将其应用于原ARC-AGI-2测试集进行诊断，发现35.3%的任务对Transformer表现出低神经亲和性，分布特征与原数据集一致。对170万参数Transformer在302个任务上的微调显示：210个任务（69.5%）局部细胞准确率超80%，但整体网格准确率低于10%，揭示显著的组合差距。此结果直接证明神经亲和力天花板效应——性能受限于架构适配性而非训练量。将框架应用于Li等人的独立ViTARC研究（400个专家模型，每个100万样本），验证其预测能力：低亲和力任务平均51.9%，高亲和力达77.7%（p<0.001），甚至出现0%表现。分类体系可精准诊断：低亲和力任务（A2）触及硬天花板，高亲和力任务（C1）可达99.8%。结果表明，进展需依赖亲和力匹配的混合架构。我们公开已验证的分类体系。
原文摘要 · Abstract (English)
Responding to Hodel et al.'s (2024) call for a formal definition of task relatedness in re-arc, we present the first 9-category taxonomy of all 400 tasks, validated at 97.5% accuracy via rule-based code analysis. We prove the taxonomy's visual coherence by training a CNN on raw grid pixels (95.24% accuracy on S3, 36.25% overall, 3.3x chance), then apply the taxonomy diagnostically to the original ARC-AGI-2 test set. Our curriculum analysis reveals 35.3% of tasks exhibit low neural affinity for Transformers--a distributional bias mirroring ARC-AGI-2. To probe this misalignment, we fine-tuned a 1.7M-parameter Transformer across 302 tasks, revealing a profound Compositional Gap: 210 of 302 tasks (69.5%) achieve >80% cell accuracy (local patterns) but <10% grid accuracy (global synthesis). This provides direct evidence for a Neural Affinity Ceiling Effect, where performance is bounded by architectural suitability, not curriculum. Applying our framework to Li et al.'s independent ViTARC study (400 specialists, 1M examples each) confirms its predictive power: Very Low affinity tasks achieve 51.9% versus 77.7% for High affinity (p<0.001), with a task at 0% despite massive data. The taxonomy enables precise diagnosis: low-affinity tasks (A2) hit hard ceilings, while high-affinity tasks (C1) reach 99.8%. These findings indicate that progress requires hybrid architectures with affinity-aligned modules. We release our validated taxonomy,
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。