慶應義塾大学 理工学部 数理科学科

理研AIPセミナー

タイトル A derived isometry theorem of Berkouk-Ginot
開催日時 2019年7月11日 13:00 - 14:00 + 30 min
主催者
講演者 Kei Hagihara 氏
場所 Keio University, Yagami-campus Bldg.14th, 6F
Room 631A/B
内容 To define a good distance between "topological datasets" is one of the most important themes in topological data analysis.
With this motivation, Kashiwara and Schapira introduced the convolution distance between complexes of sheaves with the language of sheaves and derived categories, in the paper "Persistent homology and microlocal sheaf theory".
In this talk, we give an overview of Berkouk-Ginot's preprint "A derived isometry theorem for sheaves"(arXiv: 1805:09694), where they give an explicit way for the computations of the convolution distance in terms of the combinatorial objects called "graded barcodes".
資料
URL