日本語/English

AKIFUMI OKUNO

I am a doctoral student of Shimodaira Lab., Graduate School of Informatics, Kyoto Univeresity, Japan,
a JSPS research fellow (DC2) and a trainee of RIKEN center for AIP.
mailto: okuno [at] sys.i.kyoto-u.ac.jp
Here is my wish list of books (including possession), that represents my research interests.
Biography
Education
Mar. 2016 M.Eng. Graduate school of Engineering Science, Osaka University, Japan
Graduated at the top of the list with Thesis: "Robust Graph Embedding" (Supervisor: Hidetoshi Shimodaira).
Mar. 2014 B.Eng. Engineering Science Department, Osaka University, Japan
Graduated at the top of the list with Thesis: "Robust Information Criterion and Its Application" (Supervusir: Hidetoshi Shimodaira).
Mar. 2010 graduated Ritsumeikan Junior and Senior High School
Overseas Experience
Oct. 2016 - Nov. 2016 UC Berkeley supported by MMDS scholarship [Photos]
Feb. 2016 - May. 2016 Ohio State University Prof. Belkin's lab. supported by MMDS scholarship [Photos]
Awards
May. 2016 Eligibility for Exemption from repayment of Scholarship Loan ($20,000) Japan Student Services Organization (JASSO)
Mar. 2016 Graduate school of Engineering Science Award Graduate School of Engineering Science, Osaka University
Mar. 2014 Engineering Science Department Award Engineering Science Department, Osaka University
Mar. 2010 Kimura Award Ritsumeikan Academy
Teaching/Research Experience
Apr. 2018-(Current) Research part-timer RIKEN AIP
May. 2017-Mar. 2018 Trainee RIKEN AIP
Apr. 2017-(Current) JSPS Reseach Fellow (DC2) JSPS
Feb. 2017-Mar. 2017 Research part-timer RIKEN AIP
Apr. 2016-Feb. 2017 Sigma Research Assistant Graduate school of Engineering Science, Osaka University
Oct. 2015-Mar. 2016 Tutor OUSSEP
Apr. 2014-Feb. 2017 Library Junior TA Main Library, Osaka University
Apr. 2014-May. 2016 Junior TAs Osaka University
Presentations
(arXiv preprint)
  1. Akifumi Okuno, Geewook Kim, and Hidetoshi Shimodaira. Graph Embedding with Shifted Inner Product Similarity and Its Improved Approximation Capability. [arXiv:1810.03463]
(International Conferences/Symposiums)
  1. Akifumi Okuno and Hidetoshi Shimodaira. On representation power of neural network-based graph embedding and beyond. ICML Workshop (Theoretical Foundations and Applications of Deep Generative Models), Stockholm, July 2018. [Proceedings (14.)] [arXiv:1805.12332] [Poster (pdf)]
  2. Akifumi Okuno, Tetsuya Hada and Hidetoshi Shimodaira. A probabilistic framework for multi-view feature learning with many-to-many associations via neural networks. International Conference on Machine Learning (ICML), Stockholm, July 2018. [Proceedings] [arXiv:1802.04630] [Poster (pdf)] (Long talk)
  3. ◯Akifumi Okuno and Hidetoshi Shimodaira. Leveraging local data structure for multi-view analysis with many-to-many associations, IFCS 2017, Tokyo, Aug. 2017. [Poster]
  4. ◯Akifumi Okuno and Hidetoshi Shimodaira. Statistical consistency of multi-view correlation analysis with many-to-many associations, Joint Statistical Meeting 2017, Baltimore (USA), Aug 2017. [Abst] [Poster]
  5. ◯Tetsuya Hada, Akifumi Okuno and Hidetoshi Shimodaira. Deep Multi-view Representation Learning Based on Adaptive Weighted Similarity, First International Workshop on Symbolic-Neural Learning, Nagoya, Aug 2017. [Abst]
  6. ◯Akifumi Okuno and Hidetoshi Shimodaira. Robust Multi-view Graph Embedding, International Conference on Robust Statistics 2017, Wollongong (Australia), July 2017. (TBA)
  7. ◯Kazuki Fukui, Akifumi Okuno, and Hidetoshi Shimodaira. Image and tag retrieval by leveraging image-group links with multi-domain graph embedding, IEEE The International Conference on Image Processing, Arizona (USA), September 2016. [Link]
  8. ◯Akifumi Okuno and Hidetoshi Shimodaira. Robust cross-domain matching: Analyzing multi-domain data vectors under mismatched associations. Machine Learning Summer School 2015, Kyoto Unviersity, August 2015.
(Domestic Conferences/Symposiums)
  1. ◯Akifumi Okuno, Hidetoshi Shimodaira. Robust Graph Embedding. The 21st Information-Based Induction Sciences Workshop. Tokyo, Japan. Poster presentation (in Japanese). November, 2018. (Scheduled)
  2. ◯Akifumi Okuno, Geewook Kim, Hidetoshi Shimodaira. Approximation theorems for Neural Network-based Graph Embedding and Its Extension. The 21st Information-Based Induction Sciences Workshop. Tokyo, Japan. Poster presentation (in Japanese). November, 2018. (Scheduled)
  3. ◯Takuma Tanaka, Akifumi Okuno, Hidetoshi Shimodaira. A study on degree-correction for neural network-based graph embedding. The 21st Information-Based Induction Sciences Workshop. Tokyo, Japan. Poster presentation (in Japanese). November, 2018. (Scheduled)
  4. ◯Akifumi Okuno, Hidetoshi Shimodaira. Neural Network-Based Multi-view Graph Embedding. Japanese Joint Statistical Meeting. Tokyo, Japan. Oral Presentation (in Japanese). Sep., 2018. [Abst]
  5. ◯Akifumi Okuno, Tetsuya Hada, and Hidetoshi Shimodaira. Maximum likelihood estimation of the link structure among multi-view data vectors via neural networks. The 20th Information-Based Induction Sciences Workshop. Tokyo, Japan. Poster presentation (in Japanese). November, 2017. [Poster]
  6. ◯Akifumi Okuno and Hidetoshi Shimodaira. Cross-view link prediction with attribute vectors and its information criterion. Japanese Joint Statistical Meeting. Nagoya, Japan. Oral Presentation (in English). Sep., 2017. [Abst]
  7. ◯Akifumi Okuno and Hidetoshi Shimodaira. Statistical consistency of Multi-view Matching Analysis with the many-to-many association. The 11th Japan Joint Statistical Meeting in Spring, Tokyo, Japan. Poster presentation (in Japanese). May, 2017. (Excellent Presentation Award)
  8. ◯Akifumi Okuno and Hidetoshi Shimodaira. Statistical consistency of Multi-view Matching Analysis with the many-to-many association. JST CRESTsymposium, Nagoya, Japan. Oral presentation (in Japanese). Feb., 2017.
  9. ◯Akifumi Okuno and Hidetoshi Shimodaira. Consistent Matching Analysis. Japanese Joint Statistical Meeting. Kanazawa, Japan. Oral presentation (in Japanese). September, 2016.
  10. ◯Akifumi Okuno and Hidetoshi Shimodaira. Iteratively-Reweighed Matching Correlation Analysis and its robustness. Data science symposium. Osaka, Japan. Oral presentation (in Japanese). January, 2016.
  11. ◯Akifumi Okuno and Hidetoshi Shimodaira. Robust matching correlation analysis and its application to image recognition. The 18th Information-Based Induction Sciences Workshop. Tsukuba, Japan. Poster presentation (in Japanese). November, 2015.
  12. ◯Akifumi Okuno, Kazuki Fukui, and Hidetoshi Shimodaira. Robustifying cross-domain matching correlation analysis and its application. Japanese Joint Statistical Meeting. Okayama, Japan. Oral presentation (in Japanese). September, 2016.
  13. ◯Kazuki Fukui, Akifumi Okuno, and Hidetoshi Shimodaira. Bidirectional retrieval between images and multipl,e tags using matching correlation analysis. The meeting on Image Recognition and Understanding. Osaka, Japan. Poster presentation (in Japanese). July, 2015.
  14. ◯Akifumi Okuno and Hidetoshi Shimodaira. Robust information criterion based on Density Power Divergence and its application. Japanese Joint Statistical Meeting. Tokyo, Japan. Oral presentation (in Japanese). September, 2014. [Abst]
(Misc.)
  1. ◯Akifumi Okuno. Approximation theorems for neural network-based graph embedding and its extensions. Kyushu university statistics sminar, Fukuoka, December, 2018. (Scheduled)
  2. ◯Akifumi Okuno. Neural Network-Based Graph Embedding: Its Properties and Extensions. Kyoto, November, 2018. (Scheduled)
  3. ◯Akifumi Okuno, Hidetoshi Shimodaira. Neural Network-Based Graph Embedding: Its Properties and Extensions. NTT CS-Lab. seminar, Kyoto, October, 2018. (Scheduled)
  4. ◯Akifumi Okuno, Hidetoshi Shimodaira. A probabilistic model for graph embedding and its properties. Young Statisticians' Group Summer Seminar. Gifu. Japan. Oral presentation (in Japanese). August, 2018.
  5. ◯Akifumi Okuno, Tetsuya Hada, and Hidetoshi Shimodaira. Maximum likelihood estimation of the link structure among multi-view data vectors via neural networks. Hiroshima university friday seminar, Hiroshima, January, 2018. [Abst]
  6. ◯Akifumi Okuno and Hidetoshi Shimodaira. On consistency of Matching correlation analysis with many-to-many association. Young Statisticians' Group Summer Seminar. Chiba, Japan. Oral presentation (in Japanese). August, 2016.
  7. ◯Akifumi Okuno and Hidetoshi Shimodaira. Robustifying multi-domain matching correlation analysis and its application. Young Statisticians' Group Summer Seminar. Yamaguchi, Japan. Oral presentation (in Japanese). August, 2015.