探究知识图谱零样本组合推理失败原因，发现核心瓶颈是记忆检索能力不足。
Holographic Memory for Zero-Shot Compositional Reasoning in Knowledge Graphs: A Mechanistic Study of Where and Why It Fails
- 用全息表示与霍普菲尔德清理结合，测试零样本多跳推理性能。
- 单跳预测表现良好（过滤MRR超0.35），但零样本组合准确率仍处于随机水平。
- 故障根源在记忆检索容量与干扰，而非绑定解绑或清理机制本身。
知识图谱嵌入模型虽能有效预测单跳链接，却缺乏对零样本组合查询的处理能力——即训练中从未出现的关系链组合。全息缩减表示（HRR）通过循环卷积实现符号绑定与解绑，理论上具备可逆性与结合性，是理想候选方案。本文在FB15k-237数据集上，对实值HRR与相位仅限傅里叶HRR（FHRR）两种变体，分别搭配现代霍普菲尔德清理机制，进行五次种子实验。结果表明：第一，两者单跳检索性能优异（过滤MRR分别为0.358±0.002和0.350±0.021）；第二，二者均无法实现零样本组合推理，准确率始终维持在随机水平；第三，关键发现：对第一跳中间实体的探测显示，记忆能高保真恢复正确实体（MRR 0.896±0.002），但即使输入正确中间项，组合仍失败；进一步探测发现，将真实第二跳事实作为独立原子查询时，其召回率仅为原子平均准确率的0.26至0.48倍，且不受关系扇出影响，说明问题本质在于叠加状态下记忆检索的固有难度，即容量与干扰效应已存在于单跳阶段；第四，证明了FHRR的softmax清理不满足相位等变性，加剧了少数首跳出错链路的失败。修复零样本组合需提升叠加状态下的检索容量，而非仅重构清理模块。
原文摘要 · Abstract (English)
Knowledge graph embedding (KGE) models predict single-hop links well but have no mechanism for zero-shot compositional queries: multi-hop questions whose relation chains never appeared during training. Holographic Reduced Representations (HRR), which bind and unbind symbols via circular convolution, are a theoretically attractive candidate, since binding is approximately invertible and associative. We test whether this promise holds. We study two holographic memory variants, real-valued HRR and phase-only Fourier HRR (FHRR), each with a modern Hopfield cleanup, on FB15k-237 over five seeds. Four findings follow. First, both are competitive single-hop retrievers (filtered MRR 0.358 +/- 0.002 for HRR, 0.350 +/- 0.021 for FHRR). Second, neither composes zero-shot: accuracy stays at chance across all cleanup temperatures. Third, the main contribution, we localise the failure mechanistically. A hop-1 probe shows the memory recovers the correct intermediate entity with high fidelity (MRR 0.896 +/- 0.002 for HRR), yet composition still fails even with a verified-correct intermediate. A second probe shows why: posing the ground-truth second-hop fact as a standalone atomic query, bypassing composition entirely, already recovers it at only 0.26 to 0.48x average atomic accuracy, uniformly across relation fan-out. The bottleneck is not the bind-unbind algebra or the cleanup; it is that facts compositional chains pass through are intrinsically harder for the superposed memory to retrieve, a capacity and interference effect present already at a single hop. Fourth, we prove (Lemma 4.1) that FHRR's softmax cleanup is not phase-equivariant, compounding the primary failure on the minority of chains where hop-1 itself errs. Fixing zero-shot composition requires improving retrieval capacity under superposition, not just redesigning the cleanup.
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。