2019 Fiscal Year Research-status Report

A Sequence-to-sequence Model based Dissimilarity Measurement for Clustering Structural Data

Research Project

Project/Area Number	18K18068
Research Institution	Tokyo University of Agriculture and Technology
Principal Investigator	NGUYENTUAN CUONG 東京農工大学, 工学(系)研究科(研究院), 特任助教 (10814246)
Project Period (FY)	2018-04-01 – 2021-03-31
Keywords	handwriting / mathematical expression / clustering
Outline of Annual Research Achievements	We have published our works "CNN based spatial classification features for clustering offline handwritten mathematical expressions" on Pattern Recognition Letters. We also prepared a method for clustering online handwritten mathematical expression using BLSTM-CTC for recognizing label sequence and pyramid histogram of characters for sequence embedding.
Current Status of Research Progress	Current Status of Research Progress 2: Research has progressed on the whole more than it was originally planned. Reason We published a Journal paper on the topic and continue to develop new method for solving the problem.
Strategy for Future Research Activity	We will publish the result of online handwritten mathematical expression recognition. We continue develop online and offline handwritten mathematical expression recognition using seq2seq with attention approach. Then we derive the dissimilarity of patterns by its recognition results. We also make more research to other metric learning methods and unsupervised learning to improve the robustness of the clustering method.

(2 results)