YSYingtian Shi

Latest News

  1. Our preprint Rondo: Unsupervised Discovery of Recurring Temporal Structure, a framework that discovers reusable units and recurring compositions in unlabeled time series, updating its discoveries as new data arrives, is now available on arXiv!
  2. Our paper TRACE, on contextual activity recognition through multimodal sensing and temporal reasoning, has been accepted to ACM IMWUT!
  3. Our paper EduGage, on multimodal sensing for momentary engagement assessment in self-guided learning, has been accepted to ACM IMWUT!
  4. I attended the ACM AI Leadership Summit 2026 in Atlanta.
  5. We released the preprint of TRACE: Temporal Reasoning over Context and Evidence for Activity Recognition in Smart Homes, a framework that combines ambient and wearable sensing with LLM-based temporal reasoning to understand everyday activities in context.
  6. We released the preprint of EduGage: Methods and Dataset for Sensor-Based Momentary Assessment of Engagement in Self-Guided Video Learning, an approach that uses physiological and behavioral signals to estimate learners’ momentary engagement during self-guided video learning.
  7. Our paper Towards More Equitable Ulcer Recognition Models: A Dataset of Naturalistic Foot Images from People of Color Living with Diabetes, introducing naturalistic foot images from people of color living with diabetes, was published at IEEE BHI 2025.
  8. Our preprint LLMartini: Seamless and Interactive Leveraging of Multiple LLMs through Comparison and Composition, an interactive system for comparing and composing responses from multiple LLMs, is now available on arXiv!
  9. I joined AI-CARING at Georgia Tech as a Graduate Research Assistant, working on context-aware AI and multimodal activity understanding.
  10. I returned to the PI Lab at Tsinghua University as a Research Assistant, working on interactive multi-LLM systems.
  11. Our papers GenieWizard and COMETIC were published at ACM CHI 2025!
  12. Two of our papers, GenieWizard: Multimodal App Feature Discovery with Large Language Models and Enhancing Smartphone Eye Tracking with Cursor-Based Interactive Implicit Calibration, have been accepted to ACM CHI 2025! GenieWizard helps developers discover missing multimodal app features through LLM-simulated user interactions. COMETIC improves smartphone eye tracking through implicit calibration during cursor-based interaction.
  13. Our paper CalibRead: Unobtrusive Eye Tracking Calibration from Natural Reading Behavior was published in ACM IMWUT. CalibRead calibrates eye tracking from natural reading behavior, reducing the need for explicit calibration tasks.
  14. Our paper CalibRead, on eye tracking calibration from natural reading behavior, has been accepted to ACM IMWUT!
  15. Our preprint CasualGaze: Towards Modeling and Recognizing Casual Gaze Behavior for Efficient Gaze-based Object Selection, a gaze-based selection technique that models natural, casual glances, is now available on arXiv!
  16. Our preprint SonarWatch: Field sensing technique for smartwatches based on ultrasound and motion, a smartwatch sensing technique combining ultrasound and motion sensing, is now available on arXiv!
  17. Our preprint Bridging the gap between natural user expression with complex automation programming in smart homes, an LLM-based system for expressing and generating complex smart-home automations, is now available on arXiv!
  18. I started my Ph.D. in Computer Science at Georgia Tech, advised by Prof. Thomas Plötz!