Development of machine learning algorithms based on discrete convex analysis

2017 Fiscal Year Final Research Report

• Project/Area Number: 26280086
• Research Category: Grant-in-Aid for Scientific Research (B)
• Allocation Type: Partial Multi-year Fund
• Section: 一般
• Research Field: Intelligent informatics
• Research Institution: Osaka University
• Principal Investigator: Kawahara Yoshinobu 大阪大学, 産業科学研究所, 准教授 (00514796)
• Co-Investigator(Kenkyū-buntansha): 永野 清仁 群馬大学, 社会情報学部, 准教授 (20515176) ; 岩田 具治 日本電信電話株式会社NTTコミュニケーション科学基礎研究所, 上田特別研究室, 主任研究員 (70396159)
• Co-Investigator(Renkei-kenkyūsha): HIRAI Hiroshi 東京大学, 大学院情報理工学系研究科, 准教授 (20378962) ; KANEMURA Atsunori 産業技術総合研究所, 情報数理研究グループ, 研究員 (50580297) ; ISHIHATA Masakazu 日本電信電話株式会社, NTTコミュニケーション科学研究所, 研究員 (80726563) ; TAKEUCHI Koh 日本電信電話株式会社, NTTコミュニケーション科学研究所, 研究員 (30726568)
• Project Period (FY): 2014-04-01 – 2018-03-31
• Keywords: 機械学習 / 組合せ最適化

Outline of Final Research Achievements

In this study, we developed several machine learning algorithms based on discrete convexity such as submodularity. In particular, we developed efficient learning algorithm with structured sparsity, which is formulated with continuous relaxations of submodular functions. We applied those to problems in several engineering fields, and confirmed the proposed methods effectiveness in those problems.

Free Research Field

機械学習