Publications

Publications

Theses

• • Bachelor: Deterministic Random Walks on Finite Graphs

  • Master: Convex Formulations for Learning from Crowds

  • PhD: Privacy-Preserving Crowdsourcing

Preprints

• 1. Hiroshi Kajino, Kohei Miyaguchi, Takayuki Osogami: "Biases in In Silico Evaluation of Molecular Optimization Methods and Bias-Reduced Evaluation Methodology", Jan. 28, 2022. [paper]

  2. Hiroshi Kajino: Molecular Hypergraph Grammar with its Application to Molecular Optimization, Sep. 8, 2018. [paper]

Books

1. Hiroshi Kajino, Kohei Miyaguchi, Takayuki Osogami, Ryo Iwaki, Akihumi Wachi: "Reinforcement Learning and Its Applications to Trustworthy Decision Making", Saiensu-sha, 2024. [in Japanese]

2. Hiroshi Kajino: "Molecular Optimization Using Machine Learning: From Theory to Practice", Ohm-sha, 2023. [in Japanese]

3. Masataro Asai, Hiroshi Kajino, Alex Fukunaga, Christian Muise: "Symbolic Reasoning in Latent Space: Classical Planning as an Example", Neuro-Symbolic Artificial Intelligence: The State of the Art, Chapter 2, IOS Press, 2022.

Journal Papers

• 1. Masataro Asai, Hiroshi Kajino, Alex Fukunaga, Christian Muise: Classical Planning in Deep Latent Space, Journal of Artificial Intelligence Research, Vol.74, pp.1599-1686, 2022. [paper]

  2. Sho Yokoi, Hiroshi Kajino, Hisashi Kashima: Link Prediction in Sparse Networks Using Incidence Matrix Factorization. Journal of Information Processing, Vol.25, pp.477-485, 2017. [paper]

  3. Shunsuke Kajimura, Yukino Baba, Hiroshi Kajino, Hisashi Kashima: Quality Control for Crowdsourced Enumeration, Journal of Japanese Society of Artificial Intelligence, Vol.xx, No.xx, pp.xx-xx, 2016. [in Japanese]

  4. Hiroshi Kajino, Hiromi Arai, Hisashi Kashima: Preserving Worker Privacy in Crowdsourcing, Data Mining and Knowledge Discovery, Vol.28, Issue 5, pp.1314-1335, 2014.  [paper] [slide] [errata]

  5. Hiroshi Kajino, Yuta Tsuboi, Issei Sato, Hisashi Kashima: Learning from Crowds and Experts, Journal of Japanese Society of Artificial Intelligence, Vol.28, No.3, pp.243-248, 2013. [in Japanese]

  6. Hiroshi Kajino and Hisashi Kashima: Convex Formulations of Learning from Crowds, Journal of Japanese Society of Artificial Intelligence, Vol.27, No.3, pp.133-142, 2012. [in Japanese] JSAI Best Paper Award

Conference/Workshop Papers (refereed)

• 1. Akihiro Kishimoto, Hiroshi Kajino, Masataka Hirose, Junta Fuchiwaki, Indra Priyadarsini, Lisa Hamada, Hajime Shinohara, Daiju Nakano, Seiji Takeda: "MHG-GNN: Combination of Molecular Hypergraph Grammar with Graph Neural Network", NeurIPS 2023 Workshop AI4Mat, 2023.

  2. Akihiro Kishimoto, Hiroshi Kajino, Hajime Shinohara, Daiju Nakano, Seiji Takeda: “Autoencoder based on Graph and Recurrent Neural Networks and Application to Property Prediction”, Materials Research Society (MRS) Fall Meeting, 2023.

  3. Hiroshi Kajino, Kohei Miyaguchi, Takayuki Osogami: "Biases in Evaluation of Molecular Optimization Methods and Bias Reduction Strategies", ICML-23, Hawaii, 2023.

  4. Hiroshi Kajino: "A Differentiable Point Process with Its Application to Spiking Neural Networks", ICML-21, Virtual, 2021.

  5. Miguel Garcia-Ortegon, Andreas Bender, Carl E Rasmussen, Hiroshi Kajino, Sergio Bacallado, “Combining variational autoencoder representations with structural descriptors improves prediction of docking scores”, Machine Learning for Molecules Workshop @ NeurIPS 2020.

  6. Hiroshi Kajino: "Molecular Hypergraph Grammar with Its Application to Molecular Optimization", ICML-19, Long Beach, CA, 2019. [code]

  7. Masataro Asai, Hiroshi Kajino: "Towards Stable Symbol Grounding with Zero-Suppressed State AutoEncoder", ICAPS-2019, Berkeley, CA, 2019.

  8. Kohei Miyaguchi, Hiroshi Kajino: "Cogra: Concept-drift-aware Stochastic Gradient Descent for Time-series Forecasting", In Proceedings of the Thirty-Third AAAI Conference on Artificial Intelligence (AAAI-19), pp.xx-xx, Hawaii, USA, 2019.

  9. Shohei Ohsawa, Kei Akuzawa, Tatsuya Matsushima, Gustavo Bezerra, Yusuke Iwasawa, Hiroshi Kajino, Seiya Takenaka, Yutaka Matsuo: Neuron as an Agent, Workshops of the 6th International Conference on Learning Representations (ICLR-18 workshop), Vancouver, BC, Canada, 2018 [paper]

  10. Takayuki Osogami, Hiroshi Kajino, Taro Sekiyama: Bidirectional learning for time-series models with hidden units, In Proceedings of the Thirty-fourth International Conference on Machine Learning (ICML-2017), pp.2711-2720, Sydney Australia, 2017. [paper]

  11. Hiroshi Kajino: A Functional Dynamic Boltzmann Machine, In Proceedings of the International Joint Conference on Artificial Intelligence 2017 (IJCAI-2017), pp.1987-1993, Melbourne, Australia, 2017. [paper]

      • Errata 1: W^{[\delta]} and U_l are in R^{N \times N}, not in R^{N \times M}.

      • Errata 2: K(P, X^{[t]}) -> K_{\sigma^2}(P, X^{[t]}), the first term of the gradient of L^{[t]}(\Theta) w.r.t. b (although, this will not be problematic almost surely...).

  12. Sho Yokoi, Hiroshi Kajino, Hisashi Kashima: Link Prediction by Incidence Matrix Factorization, In Proceedings of the 22nd European Conference on Artificial Intelligence (ECAI), pp.xx-xx, The Hague, Holland, 2016.

  13. Shunsuke Kajimura, Yukino Baba, Hiroshi Kajino, Hisashi Kashima: Quality Control for Crowdsourced POI Collection, in Proceedings of the 19th Pacific-Asia Conference on Knowledge Discovery and Data Mining, pp.255-267, Ho Chi Minh City, Viet Nam, 2015. Long paper (27/405). [paper]

  14. Hiroshi Kajino, Akihiro Kishimoto, Adi Botea, Elizabeth Daly, Spyros Kotoulas: Active Learning for Multi-relational Data Construction, In Proceedings of the 24th International World Wide Web Conference (WWW 2015), pp.560-569, Florence, Italy, 2015. Acceptance rate = 131/929 = 14.1%. [paper] [slide]

      • Errata: The degree of freedom of a rotation matrix is d(d-1)/2, not (d-1).

  15. Hiroshi Kajino, Yukino Baba, Hisashi Kashima: Instance-privacy Preserving Crowdsourcing, In Proceedings of the Second AAAI Conference on Human Computation and Crowdsourcing (HCOMP-2014), pp. 96-103, Pittsburgh, USA, 2014. Acceptance rate = 26/80 = 32%. [paper] [slide] [code]

  16. Hiroshi Kajino, Yuta Tsuboi, Hisashi Kashima: Clustering Crowds, In Proceedings of the 27th AAAI Conference on Artificial Intelligence (AAAI), pp.1120-1127, Bellevue, Washington, USA, 2013. Acceptance rate = 203/690 = 29%. (paper) (slide) (poster)

  17. Hiroshi Kajino, Yuta Tsuboi, Issei Sato, Hisashi Kashima: Learning from Crowds and Experts, In Proceedings of the 4th Human Computation Workshop, pp.107-113, Toronto, Ontario, Canada, 2012. (paper) (poster)

  18. Hiroshi Kajino, Yuta Tsuboi, Hisashi Kashima: A Convex Formulation for Learning from Crowds, In Proceedings of the 26th AAAI Conference on Artificial Intelligence (AAAI-12), pp.73-79, Toronto, Ontario, Canada, 2012. Acceptance rate = 294/1129 = 26% (paper) (poster) (spotlight)

Conference/Workshop Papers (non-refereed)

• 1. Hiroshi Kajino, Yukino Baba, Hisashi Kashima: Instance-privacy Preserving Crowdsourcing, The 28th Annual Conference of the Japanese Society for Artificial Intelligence, 2014. [in Japanese] JSAI Annual Conference Award

  2. Hiroshi Kajino, Yuta Tsuboi, and Hisashi Kashima: Clustering Crowds, The 27th Annual Conference of the Japanese Society for Artificial Intelligence, 2013. [in Japanese]

  3. Hiroshi Kajino, Yuta Tsuboi, Issei Sato, and Hisashi Kashima: Learning from Crowds and Experts, The 26th Annual Conference of the Japanese Society for Artificial Intelligence, 2012. [in Japanese]

  4. Hiroshi Kajino and Hisashi Kashima: A Convex Formulation of Learning from Crowds, In Proceedings of the 14th Information-Based Induction Science Workshop, pp.231-236, 2011. [in Japanese] Encouragement Award at IBIS 2011

Review Articles

• 1. Hisashi Kashima, Hiroshi Kajino: Crowdsourcing and Machine Learning, Journal of the Japanese Society for Artificial Intelligence, Vol.27, No.4, pp.381-388, 2012. [in Japanese]

Presentations & Talks

1. Hiroshi Kajino: Graph generation using a graph grammar, IBIS 2019 (invited), 2019. [slide]

2. Hiroshi Kajino, Akihiro Kishimoto, Adi Botea, Elizabeth Daly, Spyros Kotoulas: Active Learning for Multi-relational Data Construction, IBIS 2015 (invited talk), 2015 [in Japanse] [slide]

3. Shunsuke Kajimura, Yukino Baba, Hiroshi Kajino, Hisashi Kashima: Quality Control for Crowdsourced Enumeration Tasks, HCOMP2014 Work-in-Progress.

4. Hiroshi Kajino: Convex Formulations for Learning from Crowds, Sugiyama Lab. Guest Talk, 2013.

5. Hiroshi Kajino: A Convex Formulation for Learning from Crowds and Its Extensions, Machine Learning Summer School 2012, Kyoto.

6. Hiroshi Kajino, Akisato Kimura and Katsuhiko Ishiguro: A Multimodal Topic Model for Video Categorization , The 14th Information-Based Induction Science Workshop, 2011. [in Japanese]

Awards

1. 2018/07: JSAI Annual Conference Award

2. 2017/07: JSAI Annual Conference Award

3. 2015/07: JSAI Annual Conference Award

4. 2014/10: Microsoft Research Asia Fellowship Award

5. 2014/07: JSAI Annual Conference Award

6. 2013/06: JSAI Best Paper Award

7. 2013/03: DEIM2013 Presentation Award

8. 2011/11: Encouragement Award at IBIS 2011