Minori Narita
Minori Narita
Embedded Files
I'm a 3rd-year Ph.D. student at University of Toronto. I feel fortunate that I am advised by Prof. Chris Beck for my Ph.D. thesis. I am passionate about doing research on artificial intelligence. My research interest lies in the intersection between combinatorial optimization, deep learning, and reinforcement learning.
Before starting PhD, I have focused on detecting incorrect state representations in the agent model using human feedback, supervised by Prof. Shlomo Zilberstein at the University of Massachusetts Amherst.
In my previous research projects, I proposed a data transformation user interface for non-programmers and game AI that learns from its failure moves. I have also worked on image anomaly detection for medical analysis. Before coming to UofT and UMass, I did Master's of Science in planetary science at The University of Tokyo, working with Prof. Takeshi Imamura.
When I'm not working on research, I'm playing basketball, drawing illustrations, or playing a street piano.
Education
Education
Ph.D. in Mechanical and Industrial Engineering, University of Toronto (September 2021-)
M.S. in Computer Science, UMass Amherst, May 2021
M.S. in Planetary Science, The University of Tokyo, March 2019
B.Eng in Computer Engineering, The University of Tokyo, March 2017
Contact Info:
minori.narita [[at]] mail.utoronto.ca
publications
publications
Pre-prints (full papers)
Pre-prints (full papers)
Learning from Failure: Introducing Failure Ratio in Reinforcement Learning
Minori Narita and Daiki Kimura.
Published (full paper)
Published (full paper)
Correlation of Venusian Mesoscale Cloud Morphology Between Images Acquired at Various Wavelengths
Minori Narita, Takeshi Imamura, Yeon Joo Lee, Sigeto Watanabe, Atsushi Yamazaki, Takehiko Satoh, Makoto Taguchi, Tetsuya Fukuhara, Manabu Yamada, Toru Kouyama, Naomoto Iwagami (2022). Journal of Geophysical Research (JGR): Planets.
Data-centric interaction for data transformation with Programming-by-Example
Minori Narita, Nolwenn Maudet, Yi Lu, and Takeo Igarashi. Intelligent User Interfaces (IUI 2021) (acceptance rate=25.0%). 04.2021. [Honorable Mention]
Published (workshops and symposiums)
Published (workshops and symposiums)
Identifying Missing Features in State Representation for Safe Decision-Making
Minori Narita, Sandhya Saisubramanian, Roderic A. Grupen, and Shlomo Zilberstein. The 38th International Conference on Machine Learning (ICML), Workshop on Human-AI Collaboration in Sequential Decision-Making (Human AI). 07.2021. [Spotlight presentation]
Learning from Failure: Introducing Failure Ratio in Reinforcement Learning
Minori Narita, Daiki Kimura. The 29th International Joint Conference on Artificial Intelligence and the 17th Pacific Rim International Conference on Artificial Intelligence (IJCAI-PRICAI 2020), Knowledge Based Reinforcement Learning Workshop. 01.2021.
Spatially-weighted Anomaly Detection with Regression Model
Daiki Kimura, Minori Narita, Asim Munawar, and Ryuki Tachibana. Meeting on Image Recognition and Understanding (MIRU) 2018.
Spatially-weighted Anomaly Detection
Minori Narita, Daiki Kimura, and Ryuki Tachibana. Symposium on Sensing via Image Informasion (SSII) 2018. Honorable Mention (Top 6%).
Others
Others
Statistical analysis of Venus’ cloud morphologies using multi-wavelength image data
Minori Narita. Master's thesis in planetary science at The University of Tokyo.
Work experiences
Work experiences
2021.7 - 2021.8 Bloomberg L. P., Data Technologies Team - NY, USA
2021.7 - 2021.8 Bloomberg L. P., Data Technologies Team - NY, USA
Machine learning engineering intern.
Developed a front-end image annotation tool using React and back-end object detection algorithms called FasterRCNN and Detection Transformers to enable users to easily develop computer vision models.
2020.7 - 2020.8 Bloomberg L. P., Tokyo office - Tokyo, Japan
2020.7 - 2020.8 Bloomberg L. P., Tokyo office - Tokyo, Japan
Software engineering intern.
Built a tool to display status and configuration of central storage in Bloomberg called TickerPlant using a reactive framework (vue.js). It offers programmers safe and reliable interaction with machines who investigate issues in storage.
2018.5 - 2019.8 Igarashi Lab, The University of Tokyo - Tokyo, Japan
2018.5 - 2019.8 Igarashi Lab, The University of Tokyo - Tokyo, Japan
Part-time research staff. Research topic - Data-centric Interaction for Data Transformation.
Developed a new interaction framework to enable non-programmers to perform data transformation efficiently.Users can achieve complex data transformation by providing input-output examples to the system.
2017.11 - 2018.1 IBM Research AI - Tokyo, Japan
2017.11 - 2018.1 IBM Research AI - Tokyo, Japan
Research intern. Research topic - Image Anomaly Detection.
Created a novel image anomaly detection algorithm that utilized regional information extracted from the convolutional neural networks. It improved accuracy by reducing the effects of ambient image noise. Anomalies are highlighted so users can check the basis for judgment.
2017.8 Japan Aerospace Exploration Agency (JAXA) - Ibaraki, Japan
2017.8 Japan Aerospace Exploration Agency (JAXA) - Ibaraki, Japan
Short-term research intern. Research topic - Deep Reinforcement Learning for Asteroid Exploration.
Designed a new space probe utilizing deep reinforcement learning to make the probe flexibly adapt to different asteroid environments.
Page updated
Google Sites
Report abuse