When Graph-JEPA Learns the Wrong Thing: Diagnosing and Repairing Category-Conditional Collapse

📄 arXiv: 2608.20516v1 📥 PDF

作者: Gollam Rabby, Sören Auer

分类: cs.LG

发布日期: 2026-08-20


💡 一句话要点

提出Graph-JEPA修复类别条件崩溃问题的诊断方法

🎯 匹配领域: 支柱二:RL算法与架构 (RL & Architecture)

关键词: 联合嵌入 信息恢复 科学推理 方差分配 模型修复 知识图谱 数据分析

📋 核心要点

  1. 现有的联合嵌入预测架构在某些情况下表现出失效,尽管线性探测和有效秩显示健康,但信息表示却无用。
  2. 论文提出了一种修复方法,重点在于方差分配的重新调整,以提高信息恢复能力。
  3. 实验结果显示,修复后的配置在信息恢复上达到了14.377比特,显著高于原始配置的0.307比特。

📝 摘要(中文)

本文探讨了联合嵌入预测架构在特定情况下的失效,尽管线性探测和有效秩显示健康状态,但表示却没有可用的实例信息。通过对57,903篇文章的科学推理图进行实验,发现修复后的度量在目标上饱和,且缺乏结构信息。研究揭示了方差分配的问题,提出了一种修复配置,显著提高了信息恢复能力,并发布了可审计的工具以评估目标的可约性。

🔬 方法详解

问题定义:本文解决的是Graph-JEPA在类别条件崩溃情况下的失效问题,现有方法在某些情况下无法有效利用实例信息。

核心思路:通过分析方差分配,论文提出了一种修复策略,旨在重新分配信息方差,以提高模型的推理能力和信息恢复效果。

技术框架:整体架构包括数据预处理、模型训练和修复配置的评估。主要模块包括方差分析、修复算法和信息恢复测试。

关键创新:最重要的创新点在于识别并修复方差分配不当的问题,提出了一种新的目标配置,使得信息恢复能力显著提升。

关键设计:在修复过程中,采用了特定的学习率调度和损失函数设计,确保模型在训练过程中能够有效地调整方差分配。

🖼️ 关键图片

fig_0
fig_1
fig_2

📊 实验亮点

实验结果表明,修复后的模型在信息恢复上达到了14.377比特,接近于13.865比特的理论上限,显著高于未修复模型的0.307比特,验证了修复方法的有效性。

🎯 应用场景

该研究的潜在应用领域包括科学文献分析、知识图谱构建和信息检索等。通过提升模型的推理能力,能够更好地支持复杂数据的分析与理解,具有重要的实际价值和未来影响。

📄 摘要(原文)

Joint-embedding predictive architectures are selected almost universally by linear probing and effective rank. We report a case where both read healthily while the representation carries zero usable instance information. We repair it, and a second failure appears: the repaired metric saturates on a target carrying no structural information. Our corpus is a scientific-reasoning graph over 57,903 articles, each a subgraph. A Graph-JEPA predicts one masked aspect from a subgraph's remaining aspects, attaining linear-probe accuracy 0.871 and effective rank 18-47, yet retrieval recovers 0.00 of 14.4 bits (MRR 1.9e-4 vs chance 1.99e-4, p=0.98). Three upper bounds on the same pool and code recover nearly everything (+14.28, +14.34, +14.22 bits), ruling out corpus, masking, pool, and metric as causes. We trace this to variance allocation - frozen inputs place 86.05% of variance on subgraph identity and 0.40% on aspect identity, while trained latents place 0.39% and 99.61%. This is a property of the objective's optimum: the degenerate solution is a global minimum of the coupled predictor/EMA-target objective, present already at init. A repaired configuration reaches 14.377 of 14.379 bits, above the 13.865-bit oracle; reverting the loss to regression drops it to 0.307 bits, confirming it. Yet the repair licenses nothing about reasoning: the target is reducible, since intra-subgraph edges are a deterministic function of node census. The oracle reaches 96.4% of the ceiling, and our largest effect is the learning-rate schedule, not architecture. Bits and a reasoning probe show no relation across ten cells. A data-derived target fails a quality gate - 25.96% of nodes are duplicate placeholders, and the rest is more generic than supporting evidence. Rank, probes, and metrics can all saturate on an unsupportive evaluation. We release a harness with a reducibility audit and target gate.