Vindula Jayawardana

Email : vindula [AT] mit [DOT] edu

I am a PhD student in computer science at MIT. I am fortunate to be advised by Prof. Cathy Wu in the Laboratory for Information and Decision Systems (LIDS). Earlier, I completed my bachelors at University of Moratuwa in Sri Lanka.

Before graduate school, I was a part of Mobility, Algorithms, and Society Lab at Cornell University and Data Science, Engineering & Analytics Research Hub of University of Moratuwa. I have also spent time at WSO2, Autochatic, PickMe, and Trancite24.

I am interested in advancing the understanding of learning for control for large-scale cyber-physical systems and how such learning could enable advanced decision-making. In light of this, my research includes 1). making learning suitable for real-world problems by tackling the complexity, diversity, and scale of real world, and 2). leveraging learning to understand complex dynamical systems and their interplay with each other. I take sustainable cities as my domain focus with applications in learning for autonomous vehicles and tackling climate change. Currently, I am also leading the project Greenwave .

Project Highlights: : AI-driven Eco-driving for Tackling Climate Change

Google Scholar | CV | LinkedIn | Github | Blog

Recent Updates

Our paper "The Impact of Task Underspecification in Evaluating Deep Reinforcement Learning" was accepted at NeurIPS 2022.
Our paper "Learning Surrogates for Diverse Emission Models" was accepted at Climate Change AI at NeurIPS 2022.
Our recent work has been featured in National Public Radio (NPR) in The Loh Down on Science Podcast.
I am excited to receive 2022 Harold L. Hazen Teaching Award from EECS MIT.
Our paper "The Braess Paradox in Dynamic Traffic" was accepted at ITSC 2022.
Our recent work on eco-driving was featured as a spotlight on MIT front page.
Our paper "Reinforcement Learning for Eco-Lagrangian Control at Intersections" was accepted at Robotics for Climate Change at ICRA 2022.
I presented our work at the MIT CEE Annual Research Day.
Our paper "Learning Eco-Driving Strategies at Signalized Intersections" was accepted at ECC 2022.
I gave an invited talk at Univerity of Moratuwa.
Our paper "Mixed Autonomous Supervision in Traffic Signal Control" was accepted at ITSC 2021.
Our paper "Fleet management for ride-pooling with meeting points at scale: a case study in the five boroughs of New York City" is now on arXiv.
I presented a poster at MIT-IBM Watson AI Lab Virtual Open House 2021.

Articles in Review

Model-free Learning of Multi-objective Corridor Clearance in Mixed Autonomy
Vindula Jayawardana*, Dajiang Suo*, Cathy Wu (* = equal contribution)

Mitigating The Braess’s Paradox in a Closed System Using Reinforcement Learning
Yuzhu Huang*, Dingyi Zhuang*, Vindula Jayawardana, Jinhua Zhao, Dajiang Suo, Cathy Wu

What is a Typical Signalized Intersection in a City?
Ao Qu*, Anirudh Valiveru*, Catherine Tang, Vindula Jayawardana, Baptiste Freydt, Cathy Wu

Selected Recent Publications

For a complete list of publications, please visit my Google Scholar page.

	The Impact of Task Underspecification in Evaluating Deep Reinforcement Learning Vindula Jayawardana, Catherine Tang, Sirui Li, Dajiang Suo, Cathy Wu Advances in Neural Information Processing Systems (NeurIPS), 2022 , To appear [Website] [Paper] [Cite] × @INPROCEEDINGS{jayawardana2022mdpfamily, author={Jayawardana, Vindula and Tang, Catherine and Li, Sirui and Suo, Dajiang and Wu, Cathy}, booktitle={Advances in Neural Information Processing Systems (NeurIPS)}, title={The Impact of Task Underspecification in Evaluating Deep Reinforcement Learning}, year={2022}}
	Learning Eco-Driving Strategies at Signalized Intersections Vindula Jayawardana, Cathy Wu European Control Conference (ECC), 2022 International Conference on Robotics and Automation (ICRA) 2022. Robotics for Climate Change workshop, 2022 (Spotlight Talk) Press: On the road to cleaner, greener, and faster driving – MIT News Home page feature. Press: Perceptron: Risky teleoperation, Rocket League simulation and zoologist multiplication – Tech Crunch. Podcast: The Loh Down on Science Podcast - NPR. [Website] [Paper] [ICRA Presentation] [ICRA Poster] [ECC Video] [Cite] × @inproceedings{jayawardana2022ecolearning, title={Reinforcement Learning for Eco-Lagrangian Control at Intersections}, author={Vindula Jayawardana and Cathy Wu}, booktitle={20th European Control Conference (ECC)}, year={2022}, }
	The Braess Paradox in Dynamic Traffic Dingyi Zhuang, Yuzhu Huang, Vindula Jayawardana, Jinhua Zhao, Dajiang Suo, Cathy Wu IEEE Intelligent Transportation Systems Conference (ITSC), 2022 [Paper] [Cite] × @INPROCEEDINGS{braess_paradox_zhuang, author={Zhuang, Dingyi and Huang, Yuzhu and Jayawardana, Vindula and Zhao, Jinhua and Suo, Dajiang and Wu, Cathy}, booktitle={2021 IEEE International Intelligent Transportation Systems Conference (ITSC)}, title={The Braess Paradox in Dynamic Traffic}, year={2021}}
	Mixed Autonomous Supervision in Traffic Signal Control Vindula Jayawardana, Anna Landler, Cathy Wu IEEE Intelligent Transportation Systems Conference (ITSC), 2021 [Paper] [Cite] × @INPROCEEDINGS{9565053, author={Jayawardana, Vindula and Landler, Anna and Wu, Cathy}, booktitle={2021 IEEE International Intelligent Transportation Systems Conference (ITSC)}, title={Mixed Autonomous Supervision in Traffic Signal Control}, year={2021}, volume={}, number={}, pages={1767-1773}, doi={10.1109/ITSC48978.2021.9565053}}
	Fleet management for ride-pooling with meeting points at scale: a case study in the five boroughs of New York City Motahare Mounesan, Vindula Jayawardana, Yaocheng Wu, Samitha Samaranayake, Huy T. Vo To appear, 2021 [Paper] [Cite] × @misc{mounesan2021fleet, title={Fleet management for ride-pooling with meeting points at scale: a case study in the five boroughs of New York City}, author={Motahare Mounesan and Vindula Jayawardana and Yaocheng Wu and Samitha Samaranayake and Huy T. Vo}, year={2021}, eprint={2105.00994}, archivePrefix={arXiv}, primaryClass={eess.SY} }