Transformers Provably Learn Directed Acyclic Graphs via Kernel-Guided Mutual Information
机构 * Department of Mathematics, National University of Singapore, Singapore(新加坡国立大学数学系) ; Department of Statistics and Data Science, University of Pennsylvania, USA(宾夕法尼亚大学统计与数据科学系) ; Department of Electrical and Computer Engineering, The Ohio State University, USA(俄亥俄州立大学电子与计算机工程系)