I’m a research engineer at Lily MedTech Inc., a medical AI startup based in Tokyo. I’m working on research and development of computer vision techniques for a novel ultrasound computed tomography (USCT) device.
I obtained a PhD in computer science majoring artificial intelligence from Tokyo Institute of Technology in 2021.
In 2018 summer, I did a research internship at Preferred Networks Inc. I was involved in designing deep generative models for generating molecular graphs. This work resulted in GraphNVP, the first normalizing-flow based model for generating attributed graphs.
Kaushalya Madhawa and Tsuyoshi Murata, “MetAL: Active Semi-Supervised Learning on Graphs via Meta Learning”, Asian Conference on Machine Learning (ACML), 2020 [Arxiv]
Kaushalya Madhawa and Tsuyoshi Murata, “Active Learning on Graphs via Meta Learning”, Graph Representation and Beyond (GRL+) Workshop, International Conference on Machine Learning (ICML), 2020
Kaushalya Madhawa, Katushiko Ishiguro, Kosuke Nakago, and Motoki Abe, “GraphNVP: An Invertible Flow Model for Generating Molecular Graphs”, Arxiv preprint, 2019 [Arxiv]
Kaushalya Madhawa and Tsuyoshi Murata, “Exploring Partially Observed Networks with Non-parametric Bandits”, Conference on Complex Networks and Their Applications, Cambridge, UK, 2018 [springer.com] [Arxiv]
Kaushalya Madhawa, Choong Jun Jin, Arie Wahyu Wijayanto, and Tsuyoshi Murata, “Robustness of Compressed Convolutional Neural Networks”, Workshop on Big Data for CyberSecurity (BigCyber-2018), IEEE International Conference on Big Data, Seattle, Washington, 2018
P.K.K.Madhawa and A.S. Athukorale, A Robust Algorithm for Determining the Newsworthiness of Microblogs”, International Conference on Advances in ICT for Emerging Regions (ICTer), Colombo, Sri Lanka, 2015