About

I’m a fifth year PhD candidate in computer science at Harvard University, co-advised by Finale Doshi-Velez (at the Data to Actionable Knowledge lab) and Susan Murphy. My research involves leveraging insights from behavior science to improve how artificial agents assist human decision-makers. My daily toolkit includes (but is not limited to) reinforcement learning, user studies, and probabilistic models. I am grateful for the support of the NSF Graduate Research Fellowship.

A large part of my work is inspired by digital interventions, where we use reinforcement learning to personalize behavioral interventions. For example, planning when to send “reminders to move” on a fitness app. Here, I ask how models from behavior science can provide the inductive bias our algorithm needs to quickly learn the best intervention for each person. Another part of my work involves explainable AI, where we investigate whether different downstream tasks require different explanation properties. Here, I ask how human models can be used to simulate user studies, which help us narrow in on the most promising property-task relationships prior to verification with real humans.

During my CS undergrad at the University of Kentucky, I had the pleasure of working with Simone Silvestri on algorithms for smart home applications. I also participated in some wonderful REU programs at DePaul University (mentored by Daniela Raicu) and USC (mentored by Stefanos Nikolaidis).

Research

Publications

• Reinforcement Learning Interventions on Boundedly Rational Human Agents in Frictionful Tasks. Eura Shin, Siddharth Swaroop, Weiwei Pan, Susan Murphy, and Finale Doshi-Velez. International Conference on Autonomous Agents and Multiagent Systems 2024.
• AMBER: An Entropy Maximizing Environment Design Algorithm for Inverse Reinforcement Learning. Paul Nitschke, Lars Lien Ankile, Eura Shin, Siddharth Swaroop, Finale Doshi-Velez, Weiwei Pan. ICML Workshop on Models of Human Feedback for AI Alignment 2024. (Mentorship role).
• Online Model Selection by Learning how Compositional Kernels Evolve. Eura Shin, Predag Klasnja, Susan Murphy, and Finale Doshi-Velez. Transactions on Machine Learning Research 2023.
• Discovering User Types: Mapping User Traits by Task-Specific Behaviors in Reinforcement Learning Lars L. Ankile, Brian S. Ham, Kevin Mao, Eura Shin, Siddharth Swaroop, Finale Doshi-Velez, Weiwei Pan. ICML Workshop on AI and HCI 2023. Best paper nomination. (Mentorship role).
• Modeling Mobile Health Users as Reinforcement Learning Agents. Eura Shin, Siddharth Swaroop, Weiwei Pan, Susan Murphy, and Finale Doshi-Velez. AAAI Workshop on AI for Behavior Change 2023. Contributed talk.
• Online Structural Kernel Selection for Mobile Health. Eura Shin, Predag Klasnja, Susan Murphy, and Finale Doshi-Velez. ICML Workshop in Interpretable Machine Learning for Healthcare 2021.
• Design and Evaluation of a Hair Combing System Using a General-Purpose Robotic Arm. Nathaniel Dennler, Eura Shin, Maja Matarić, Stefanos Nikolaidis. International Conference on Intelligent Robots and Systems (IROS) 2021.
• Untangling the seasonal dynamics of plant-pollinator communities. Bernat Bramon Mora, Eura Shin, Paul J CaraDonna, Daniel B Stouffer. Nature Communications 2020.

Invited Talks

• “Modeling mHealth Users for AI Interventions”. Biodesign Lab, Harvard University, July 2023.
• Invited Talk at “AI for health behavior change” workshop. Persuasive Technology conference, Eindhoven, April 2023.

Teaching and Mentorship

I am a teaching fellow and course developer for CS290A & CS290B: Effective Research Practices & Academic Culture (Fall 2022 - Spring 2023). We are currently working on making the teaching materials available online. In Spring 2022, I served as a teaching fellow for STAT234: Sequential Decision Making.

I serve as a direct research mentor for a number of undergraduate and master’s students in the intersection of reinforcement learning and human-computer interaction.

I’ve developed the material for a number of guest lectures on Machine Learning, Reinforcement Learning, and Research Skills:

• How to conduct a literature search. CS290 Fall 2022. Detailed course plan here and guide here.
• Hieararchical Reinforcement Learning. STAT234 Spring 2022. Slides here.

Outreach

Leadership I was one of the founding member of the University of Kentucky ACM-W chapter. During this time, I created a series of introductory CS workshops for high-school students from counties in Kentucky without a computer science class/program.

Talks and workshops

• I was a mentor for the Women in Data Science (WiDS) Cambridge datathon workshop in Spring 2020.
• I gave a joint talk titled “What is Machine Learning?” at Harvard’s WECode 2023.