Research Topics

Our research focuses on developing advanced data analysis methods such as machine learning and data mining, and their applications to important real-world problems in various fields including marketing, healthcare, and industrial systems. Our research interest also includes human-computer cooperative problem solving for hard problems that cannot be solved only by computers.


1. Advanced Machine Learning Methods

Although machine learning has achieved significant developments in recent years, we still face situations in the real world where existing techniques are not applicable .

We generalize such situations to find out novel machine learning problems, and develop solutions for them.
Example include predictive modeling of complex structured data such as sequences, trees and graphs, that are general representations of natural language texts, chemical compounds, and social and biological networks.

graphs[Selected Publications]

  • Hisashi Kashima, Teruo Koyanagi. Kernels for Semi-Structured Data. In Proc. 19th International Conference on Machine Learning (ICML), pp.291-298, 2002.
  • Hisashi Kashima, Koji Tsuda, Akihiro Inokuchi. Marginalized Kernels Between Labeled Graphs. In Proc. 20th International Conference on Machine Learning (ICML), pp.321-328, 2003.
  • Hisashi Kashima, Naoki Abe. A Parameterized Probabilistic Model of Network Evolution for Supervised Link Prediction. In Proc. 6th IEEE International Conference on Data Mining (ICDM), pp.340-349, 2006.
  • Hisashi Kashima, Tsuyoshi Kato, Yoshihiro Yamanishi, Masashi Sugiyama, Koji Tsuda. Link Propagation: A Fast Semi-supervised Learning Algorithm for Link Prediction. In Proc. 2009 SIAM Conference on Data Mining (SDM), pp. 1099-1110, 2009.
  • Atsuhiro Narita, Kohei Hayashi, Ryota Tomioka, Hisashi Kashima. Tensor Factorization Using Auxiliary Information. Data Mining and Knowledge Discovery. Vol.25, No.2, pp.298-324, 2012.
  • Nozomi Nori, Danushka Bollegala, Hisashi Kashima. Multinomial Relation Prediction in Social Data: A Dimension Reduction Approach. In Proc. 26th AAAI Conference on Artificial Intelligence (AAAI), pp.115-121, 2012.
  • Ryusuke Takahama, Toshihiro Kamishima, Hisashi Kashima. Progressive Comparison for Ranking Estimation. In Proc. 25th International Joint Conference on Artificial Intelligence (IJCAI), pp.3882-3888, 2016.

2. Novel Machine Learning Applications

While machine learning has been successfully applied to various fields including marketing, healthcare, and industrial systems, there still remains a number of unexplored fields where machine learning can make significant contributions.

Collaborating with partners in industries and governments, we investigate new applications of machine learning to make great impacts in real-world contexts.

Our target domains include:
Web marketing, transportation, industrial systems, education, bio-/chemo-informatics, healthcare, material science, human resource management, and patent analysis.

applications

[Selected Publications]

  • Hisashi Kashima, Yoshihiro Yamanishi, Tsuyoshi Kato, Masashi Sugiyama, Koji Tsuda. Simultaneous Inference of Biological Networks of Multiple Species from Genome-wide Data and Evolutionary Information: A Semi-supervised Approach. Bioinformatics, Vol.25, No.22, pp.2962-2968, 2009.
  • Junichiro Mori, Yuya Kajikawa, Hisashi Kashima, Ichiro Sakata. Machine Learning Approach for Finding Business Partners and Building Reciprocal Relationships. Expert Systems With Applications, Vol.39, No.12, pp.10402-10407, 2012.
  • Yukino Baba, Hisashi Kashima, Yasunobu Nohara, Eiko Kai, Partha Ghosh, Rafiqul Islam, Ashir Ahmed, Masahiro Kuroda, Sozo Inoue, Tatsuo Hiramatsu, Michio Kimura, Shuji Shimizu, Kunihisa Kobayashi, Koji Tsuda, Masashi Sugiyama, Mathieu Blondel, Naonori Ueda, Masaru Kitsuregawa, Naoki Nakashima. Predictive Approaches for Low-cost Preventive Medicine Program in Developing Countries. In Proc. 21st ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD), 2015.
  • Nozomi Nori, Hisashi Kashima, Kazuto Yamashita, Hiroshi Ikai, Yuichi Imanaka. Simultaneous Modeling of Multiple Diseases for Mortality Prediction in Acute Hospital Care. In Proc. 21st ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) , pp.855-864, 2015.
  • Nozomi Nori, Hisashi Kashima, Kazuto Yamashita, Susumu Kunisawa, Yuichi Imanaka. Learning Implicit Tasks for Patient-Specific Risk Modeling in ICU. In Proc. 31st AAAI Conference on Artificial Intelligence (AAAI), 2017.
  • Yuji Horiguchi, Yukino Baba, Hisashi Kashima, Masahito Suzuki, Hiroki Kayahara, Jun Maeno. Predicting Fuel Consumption and Flight Delays for Low-cost Airlines. In Proc. 29th Conference on Innovative Applications of Artificial Intelligence (IAAI), 2017.

3. Human Computation / Human-in-the-loop AI

Despite of the recent significant advances of artificial intelligence technologies and the spreading idea of AI being a possible threat to mankind, the reality is that the ability of artificial intelligence is still behind that of mankind in terms of flexibility and creativity, especially for abstract, open-ended, and context-dependent tasks.

Human computation / Human-in-the-loop AI is a relatively new research area aiming at solving hard problems by combining artificial intelligence and human intelligence. We explore the world of human computation and human-in-the-loop AI, a new frontier of artificial intelligence research.

humancomputation

[Selected Publications]

  • Hiroshi Kajino, Yuta Tsuboi, Hisashi Kashima: A Convex Formulation for Learning from Crowds, In Proc. 26th AAAI Conference on Artificial Intelligence (AAAI), pp.73-79, 2012.
  • Yukino Baba, Hisashi Kashima. Statistical Quality Estimation for General Crowdsourcing Tasks, In Proc. 19th ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD), pp.554-562, 2013.
  • Hiroshi Kajino, Yukino Baba, Hisashi Kashima. Instance-privacy Preserving Crowdsourcing. In Proc. 2nd Conference on Human Computation and Crowdsourcing (HCOMP), pp.96-103,
  • Hiroshi Kajino, Hiromi Arai, Hisashi Kashima. Preserving Worker Privacy in Crowdsourcing. Data Mining and Knowledge Discovery, Vol.27, No.5-6, pp.1314-1335, 2014.
  • Yukino Baba, Hisashi Kashima, Kei Kinoshita, Goushi Yamaguchi, Yosuke Akiyoshi. Leveraging Non-expert Crowdsourcing Workers for Improper Task Detection in Crowdsourcing Marketplaces. Expert Systems with Applications, Vol.41, No.6, pp.2678-2687, 2014.
  • Yukino Baba, Kei Kinoshita, Hisashi Kashima. Participation Recommendation System for Crowdsourcing Contests. Expert Systems with Applications, Vol.58, pp.174–183, 2016.
  • Naoki Otani, Yukino Baba, Hisashi Kashima. Quality Control of Crowdsourced Classication Using Hierarchical Class Structures. Expert Systems with Applications, Vol.58, pp.155–163, 2016.
  • Satoshi Oyama, Yukino Baba, Ikki Ohmukai, Hiroaki Dokoshi, Hisashi Kashima. Crowdsourcing Chart Digitizer: Task Design and Quality Control for Making Legacy Open Data Machine-Readable. International Journal of Data Science and Analytics, 2016.
  • Takeru Sunahase, Yukino Baba, Hisashi Kashima. Pairwise HITS: Quality Estimation from Pairwise Comparisons in Creator-Evaluator Crowdsourcing Process. In Proc. 31st AAAI Conference on Artificial Intelligence (AAAI), 2017.
  • Jiyi Li, Yukino Baba, Hisashi Kashima.
    Hyper Questions: Unsupervised Targeting of a Few Experts in Crowdsourcing.
    In Proceeding of the 26th ACM International Conference on Information and Knowledge Management (CIKM), 2017.