Megha Thukral
756 W Peachtree St NW
CODA Building
Atlanta, Georgia, USA
Hey! I am a 2nd-year ML PhD student at School of Interactive Computing, Georgia Institute of Technology. My current research is focused on developing generalizable machine learning models for sensor-based Human Activity Recognition. I work with different sensor modalities including but not limited to wearables , smart home ambient sensors, and a combination of both. I am currently funded by NSF AI Caring project , focussing on HAR models tragetted for elderly care. I am grateful to be advised by Prof. Thomas Ploetz.
news
| May 16, 2025 | 🎉⌚ Excited to share that I am interning with Samsung Research America in Digital Health Lab building ML/AI models for large-scale wearable datasets with Biomarkers team !! |
|---|---|
| May 1, 2025 | 🎉📄 Excited to share our new work “AgentSense: Virtual Sensor Data Generation Using LLM Agents in Simulated Home Environments” where we leverage LLMs and VirtualHome to simulate human behavior and generate ambient sensor data. Read the Preprint |
| Jan 11, 2025 | 🎉📄 Our paper Layout-Agnostic Human Activity Recognition in Smart Homes through Textual Descriptions Of Sensor Triggers (TDOST), has been accepted to ACM IMWUT. Looking forward to Ubicomp 2025! |
| Oct 20, 2024 | 🎉📄 Our paper Cross-Domain HAR: Few-Shot Transfer Learning for Human Activity Recognition has been accepted to ACM Transactions on Intelligent Systems and Technology |
| Aug 6, 2024 | Successfully deployed a working HAR system using our Layout Agnostic TDOST approach in a real smart home – Georgia Tech Aware Home. Take a look at the video of demo for NSF AI Caring project! |