近三年，在无向图模型的可信赖性以及无向图模型诱导的概念类的复杂性等问题上，取得了一些成果。本人和杨有龙教授合作证明了无向离散图模型诱导的概念类的 VC 维数、欧氏嵌入维数、与模型自身的维数以及相应矩阵的秩是相同的，这为 VC 维数的计算奠定了理论基础；整理出的文章已在 Pattern Rcognition 发表，审稿时得到三位匿名审稿人的一致高度评价。如审稿人1：“Paper represents a lot of good work. The result is important to many people, who may not necessarily need to know how it was obtained.” 审稿人 2：“...I think those highlighted contributions especially the theoretical side have been adequately presented and justified. By showing enough details of proofs and derivations, very interesting and promsing conclutions have been reached...”