Semi-supervised Learning of Hierarchical Representations of Molecules Using Neural Message Passing

Shohei Hido

VP of Research and Development

This is a guest post in an interview style with Hai Nguyen, a former intern 2017 summer at Preferred Networks, whose research has been accepted at one of the NIPS 2017 workshops. After finishing PFN internship, he joined Kyoto University as a Ph.d student.
“Semi-supervised Learning of Hierarchical Representations of Molecules Using Neural Message Passing,” Hai Nguyen, Shin-ichi Maeda, and Kenta Oono; NIPS Workshop on Machine Learning for Molecules and Materials, 2017. (Link, arXiv)

– Please briefly introduce your NIPS paper and its value?
We proposed an unsupervised algorithm for learning representation of molecules and extended it to semi-supervised tasks to achieve better performance than supervised tasks only using labeled molecules. This preliminary work is valuable in the sense that it can be used to deal with a challenge of collecting molecules with labels while a vast number of unlabeled samples are available.
– This paper is based on your work at PFN internship. How did you choose this topic first?
In fact my previous research was not related to biology or chemistry. Since I wanted to challenge myself with other applications of machine learning and other fields, I decided to choose this topic, deep learning for molecular fingerprint prediction. Currently, I am also doing my Ph.d on bioinformatics.
– Please tell us more about your internship. Where and when did you find it?
Actually I heard of PFN around 3 years ago when I attended IBISML (a machine learning conference in Japan). I am interested in this company and looked around some information about it on the internet and accidentally knew about the internship. I decided to submit my CV to PFN right after. At there, I did my internship like doing my research in a laboratory with talented and friendly mentors. I could learn a lot from them, including knowledge and experience.
– What do you think is the unique benefit of PFN internship after all?
I think the best benefit is that there are many great researchers there that interns can learn how to do research and knowledge via their valuable discussions and supervision. Additionally, for an international student like me, I can understand more about working environment in Japanese company. They are professional and disciplined. By the way, I would like to express my gratitude to Maeda-san and Oono-san who helped and encouraged me a lot during the internship.

Preferred Networks is actively hiring from around the world not only interns and AI residency program students, but also full-time employees (job page). Also, call for application to summer internship for students in japan is still open. We are looking forward to receiving your application.

  • Twitter
  • Facebook

Archive List