Han Bao of the University of Tokyo graduate student, belonging to Imperfect Information Learning Team (Team Leader:Masashi Sugiyama) RIKEN Center for Advanced Intelligence Project, received AIP Network Lab Award (1st) at the 2017 AIP Challenge Program. Han Bao participated in the AIP Challenge Program as a researcher of Tatsuya Harada Team of CREST “Big Data Core Technology” (Research Supervisor: Masaru Kitsuregawa).
AIP challenge program aims to support individual research for young researchers who belong to CREST team. In 2017, 40 researchers were selected, and Han Bao got the 1st prize in the final presentation by mutual election.
Title: Classification from Pairwise Similarity and Unlabeled Data
Abstract: In this research, our goal is to make classifiers from information which is easier to obtain than class labels, in order to reduce labeling cost in the traditional supervised learning. It is shown that classifiers can be learned from pairwise similar data, i.e., two data belong to the same class. This method gives good performances for many datasets.
