两篇新研究论文介绍了图表示学习的新方法。其中一篇论文《Graphical einops》提出了一种用于张量编程的正式图演算，它连接了张量网络和计算图，实现了可证明的图示和稀疏注意力块的高效实现。另一篇论文《View Space》提出了一个通过形式化“视图空间”来学习任意图表示的框架，该空间允许统一表示具有异构特征的图，并在节点分类基准测试中取得了最先进的结果。 AI
影响 这些论文为图表示学习引入了新的理论框架和模型，有可能提高处理复杂关系数据的任务的性能。
排序理由 arXiv上发表的两篇不同的学术论文。
在 Hugging Face Daily Papers 阅读 →
AI 生成摘要 · Google Gemini · 来自 4 个来源。 我们如何撰写摘要 →
arXiv:2605.31485v1 Announce Type: new Abstract: Architecture diagrams are ubiquitous in deep learning, but they are usually only representational: the tensor-program identities they suggest are still proved by prose and tensor-axis manipulation. We introduce a formal graphical ca…
arXiv:2512.11561v2 Announce Type: replace Abstract: Generalizing pretrained models to unseen datasets without retraining is a central challenge toward foundation models. Achieving fully inductive inference on numerical data is particularly difficult due to large variations in fea…
Architecture diagrams are ubiquitous in deep learning, but they are usually only representational: the tensor-program identities they suggest are still proved by prose and tensor-axis manipulation. We introduce a formal graphical calculus for the structural fragment of tensor pro…
Traditional network analysis focuses on single-layer networks, real-world systems often form multilayer networks with multiple relationship types. However, existing methods typically fail to capture complex inter-layer dependencies by treating layers independently or aggregating …