GraphMERT: Efficient and Scalable Distillation of Reliable Knowledge Graphs from Unstructured Data
GraphMERT:从非结构化数据中高效可扩展地蒸馏可靠知识图谱
机构 * Department of Electrical and Computer Engineering, Princeton University(普林斯顿大学电气与计算机工程系)
AI总结 本文提出GraphMERT,一种高效的图神经网络模型,通过蒸馏高质量知识图谱提升领域知识的可靠性与有效性。
Comments Camera-ready version. Published in Transactions on Machine Learning Research (TMLR), 2026. Reviewed on OpenReview: https://openreview.net/forum?id=tnXSdDhvqc
Journal ref Transactions on Machine Learning Research, 2026