Cambridge, Massachusetts, United States
Computer Science PhD researcher passionate about applying machine learning and robotics to human-centered real-world applications
I am working with Prof. Ying Xu on human-centered AI/ML with a focus on its application in education. My research focuses on developing personalized, human-centered AI/ML agents and robots that adapt to users' individual needs to support learning, health, and overall quality of life.
I was advised by Prof. Maja Matarić. - Applying supervised/unsupervised domain adaptation to design personalized machine learning models outperforming non-personalized baselines for affective recognition - Developed multimodal models with statistical learning and deep learning methods to detect engagement and cognitive affective states for children with ASD such that a robot tutor can offer preemptive support and feedback
- Integrated ROS into Unity 3D’s C# environment along with software engineering using the Mixed Reality Toolkit for the Microsoft Hololens - Trained both normal and individualized Bayesian Knowledge Tracing Models to study the learning curves for individual participant with autism after studying with our socially assistive robot math tutor - Developed a multimodal data processing pipline to synchronize and synthesize a multimodal dataset collected from in-home SAR deployments
At Sara, we are building state-of-the-art AI speech tutor for speech articulation learning that empower speech therapists to extend their care beyond the clinic and help families access a cost-effective way to practice speech at home.
Worked on developing machine learning methods for fairness-aware data distillation for attribute classification. Our work on machine learning systems and methods for fairness-aware data distillation for attribute classification has been granted as a US patent: https://patents.google.com/patent/US12153554B2/en
Worked on developing Multi-attribute Selective Suppression, or MaSS, a general machine learning framework for performing precisely targeted data surgery to simultaneously suppress any selected set of attributes while preserving the rest for downstream machine learning tasks. Our work on machine learning systems and methods for removal of attributes from multi-modality and multi-attribute data has been granted as a US patent: https://patents.google.com/patent/US12153554B2/en