Soomin Shin

I'm Soomin Shin, a Ph.D. candidate in Electrical and Computer Engineering at the University of Waterloo, advised by Professor Kerstin Dautenhahn. I am a member of the Social and Intelligent Robotics Research Lab, where I explore how Human-Robot Interaction (HRI), socially assistive robotics, and large language models (LLMs) can come together to create intelligent, user-centered systems for real-world therapy and education.

My ultimate research goal is bringing embodied agents into our daily lives so they can be used as easily as laptops and smartphones without any technical help from experts. I lead a project developing therapeutic robots that integrate therapy-focused games in collaboration with SLPs from Kickstart Therapy. The purpose of this project is integrating robots into the clinical setting so SLPs can use our system as they would use tablets during their regular sessions.

On top of this project, I am currently working on integrating LLMs into our system to enhance adaptability! The current system can only provide fixed interactions, e.g., fixed robot speech, gestures, and even fixed game interactions. However, thanks to LLMs, we can now make the robot conduct new interactions during real-time engagement that have never been done before. This would make our system more sustainable since it can evolve with the user as time goes by.

Also, prior to my Ph.D., I conducted research on contextual affect modeling at Korea University, comparing human emotion perception with CNN-based models—work that was presented at ACM ICMI. I'm passionate about building AI systems that are not only intelligent, but also empathetic and collaborative.

LLM-driven Therapy Robot System

A scalable robot system that can autonomously adjust its system under therapist guidance in real-world settings.

My role: project leader

Co-designed Robot System for Clinical Settings [1,2]

A robot system that can be utilized by non-technical users in real-world clinical settings.

My role: project leader

Robotic Assistance System for Hospital Isolation Wards [3]

Developing a system of heterogeneous robots to reduce the workload and infection risk for medical staff in hospital isolation wards.

My role: user study design and analysis

Human-Robot Trust and Control [5]

An investigation into whether people prefer robots that act proactively or those that follow specific verbal commands.

My role: project leader

Multi-Agent Robot Trust Studies [4]

Exploring how perceived hierarchies between robots affect human trust levels.

My role: project leader

Contextual Emotion Recognition [6]

Comparison study on human perception vs. CNN-based contextual emotion recognition model.

My role: project leader

Altruistic Behavior in Virtual Agents

Investigating if virtual agents can learn altruistic behaviors (i.e., food-sharing) using reinforcement learning in a simulated environment.

Empathy Toward Robots

Eliciting empathetic responses from humans toward robots using Vector robot.