profile image

About Me

Researchmap / Google Scholar

奥野 （ おくの ） 彰文 （ あきふみ ） ：博士（情報学）

統計数理研究所 統計思考院 助教
(兼) 総合研究大学院大学 統計科学専攻 助教
(兼) 理研AIP 高次元統計モデリングチーム 客員研究員

履歴書（英語）

Tips

okuno[at]ism.ac.jp
〒190-0014 東京都立川市緑町10-3

@public_aokn

Akifumi OKUNO (Doctor of Informatics)

Assistant Professor at School of Statistical Thinking, The Institute of Statistical Mathematics,
(conc. with) Assistant Professor at Department of Statistical Science, The Graduate University for Advanced Studies, SOKENDAI,
(conc. with) Visiting Scientist at High-Dimensional Statistical Modeling Team, RIKEN AIP.

Download CV

okuno[at]ism.ac.jp
Midori-machi 10-3, Tachikawa, Tokyo, Japan. Postal code: 190-0014.

@public_aokn

Research Interest

When I was a graduate student in Osaka Univeristy and Kyoto University, I studied robust statistics, model selection, correlation analysis, graph embedding, and nonparametric classification (see my presentation in [English]) with my supervisor Prof. Shimodaira. Particularly I am interested in their theoretical aspects.

Soon after I moved to the institute of statistical mathematics as an assistant professor, I started four research projects:

  1. design-dependence of asymptotic variance of link prediction problem (with K. Yano),
  2. generalization gap estimation for overparameterized models (including neural networks) using Langevin dynamics and WAIC (with K. Yano and R. Hataya; see my presentations in [English] and [Japanese]),
  3. minimax analysis for learning invertible regression functions (with M. Imaizumi; see my presentation in [English]),
  4. clustering analysis with a specified uncertainty of covariates and its application to astronomy (with K. Hattori and R. Ian U.; see my presentation in [English]).

Recently, I am also interested in robust optimization, physics-informed machine learning, and time series analysis including point process.

Publications

Preprints
  1. Kohei Hattori, Akifumi Okuno, and Ian U. Roederer. "Finding r-II sibling stars in the Milky Way with the Greedy Optimistic Clustering algorithm", 36 pages. In revision.
    arXiv:2207.04110
  2. Akifumi Okuno and Kohei Hattori. "A Greedy and Optimistic Approach to Clustering with a Specified Uncertainty of Covariates", 20 pages. Under review.
    arXiv:2204.08205 GitHub
  3. Ruixing Cao*, Akifumi Okuno*, Kei Nakagawa and Hidetoshi Shimodaira. "Improving Nonparametric Classification via Local Radial Regression with an Application to Stock Prediction", 23 pages. (*The first two authors R. Cao and A. Okuno contributed equally to this work.) Under review.
    arXiv:2112.13951
  4. Akifumi Okuno and Keisuke Yano. "A generalization gap estimation for overparameterized models via the Langevin functional variance", 21 pages. In revision.
    arXiv:2112.03660
  5. Akifumi Okuno and Masaaki Imaizumi. "Minimax Analysis for Inverse Risk in Nonparametric Planer Invertible Regression", 29 pages. In revision.
    arXiv:2112.00213
  6. Morihiro Mizutani, Akifumi Okuno, Geewook Kim and Hidetoshi Shimodaira. "Stochastic Neighbor Embedding of Multimodal Relational Data for Image-Text Simultaneous Visualization", 20 pages. Submitted.
    arXiv:2005.00670

Journal Paper (Refereed)
  1. Akifumi Okuno and Keisuke Yano. "Dependence of variance on covariate design in nonparametric link regression", accepted to Statistics and Probability Letters.
    Paper arXiv:2012.13106
  2. Akifumi Okuno and Hidetoshi Shimodaira. "Hyperlink Regression via Bregman Divergence," Neural Networks, June 2020, Volume 126, pages 362-383.
    Paper arXiv:1908.02573

Conference Papers (Refereed)
  1. Akifumi Okuno and Hidetoshi Shimodaira. "Extrapolation Towards Imaginary 0-Nearest Neighbour and Its Improved Convergence Rate," Advances in Neural Information Processing Systems 33 (NeurIPS 2020), pages 21889-21899.
    Paper arXiv:2002.03054
  2. Geewook Kim, Akifumi Okuno, Kazuki Fukui, and Hidetoshi Shimodaira. "Representation Learning with Weighted Inner Product for Universal Approximation of General Similarities," Proceedings of the Twenty-Eighth International Joint Conference on Artificial Intelligence (IJCAI), 2019, pages 5031-5038.
    Paper arXiv:1902.10409
  3. Akifumi Okuno and Hidetoshi Shimodaira. "Robust Graph Embedding with Noisy Link Weights." 22nd International Conference on Artificial Intelligence and Statistics (AISTATS), Okinawa, Japan, 2019, PMLR 89:664-673.
    Paper arXiv:1902.08440
  4. Akifumi Okuno, Geewook Kim, and Hidetoshi Shimodaira. "Graph Embedding with Shifted Inner Product Similarity and Its Improved Approximation Capability," 22nd International Conference on Artificial Intelligence and Statistics (AISTATS), Okinawa, Japan, 2019, PMLR 89:644-653.
    Paper arXiv:1810.03463
  5. Akifumi Okuno, Tetsuya Hada, and Hidetoshi Shimodaira. "A probabilistic framework for multi-view feature learning with many-to-many associations via neural networks," 35th International Conference on Machine Learning (ICML), Stockholm, Sweden, 2018, PMLR 80:3888-3897. (selected for long talk)
    Paper arXiv:1802.04630
  6. Kazuki Fukui, Akifumi Okuno, and Hidetoshi Shimodaira. "Image and tag retrieval by leveraging image-group links with multi-domain graph embedding," Proceedings of the 2016 IEEE International Conference on Image Processing (ICIP), Phoenix, AZ, USA, 2016, pp. 221-225. doi: 10.1109/ICIP.2016.7532351.
    Paper

    7. Miscellaneous

Presentations

Invited talks are also listed in my CV. See my list of presentations excluding above publications.

国内学会では主に

ほか大学や研究所のセミナーなどに時々参加しています．

Grants/Projects

As a PI:
  • JSPS KAKENHI 21K17718 (Apr.2021-Mar.2025; Grant-in-Aid for Young Scientists, 4,555,000JPY)
  • JSPS KAKENHI 17J03623 (Apr.2017-Mar.2019; Grant-in-Aid for JSPS Fellow DC2, 1,900,000JPY)
As a research collaborator (without grant):

Awards

  • Nov. 2019. Excellent Presentation Award from IBISML2019.
  • Feb. 2019. Excellent Research Award from ICT-13 held in Kyoto Unviersity.
  • May 2017. Full Repayment Exemption of JASSO Type-I Scholarship (2,000,000JPY≈20,000USD) for Students with Excellent Grades from Japan Student Services Organization (JASSO).
  • Mar. 2017. Excellent Presentation Award from Japan Statistical Society.
  • Mar. 2016. Graduate School of Engineering Science Award (Graduated Top of the List) from the Divison of Mathematical Science, Osaka University.
  • Mar. 2014. Engineering Science Department Award (Graduated Top of the List) from Mathematical Science Course, Osaka University.
  • Mar. 2010. Kimura Award from Ritsumeikan Academy.
  • Dec. 2007. Asahi Shimbun Award from JSEC2007.

Misc.

Amazon wish list of books represents my interests.

Datasets: UCI Machine Learning Repository KEEL Data Set Repository kaggle Open Datasets Open Data on AWS Awesome Public Datasets List of datasets for ML research (wikipedia) arXiv Times (Datasets) Google Dataset Search MSR Open Data

Calenders: wikiCFP (Machine Learning) AI Conference Deadlines International Conferences in AI, ML, CV and DM ML+CompBio Conference Deadlines World Conference Calendar

Time-zone converter: timeanddata.com

Bookstores: Springer Link Book Depository 四方堂書店

Personally, I like playing puzzle game. While I used to play several authentic puzzles such as Rubik's cube, Sudoku, Nonogram, ..., I was heavy into Tetris99 during the stay-at-home period of COVID-19.