I’m Harsha. I’m a Machine Learning PhD student in Department of Computational Science and Engineering at Georgia Institute of Technology . I am affiliated with AdityaLab and am advised by Dr. B Aditya Prakash. I gratuated from Indian Institute of Technology Mardas and am fortunate to have worked with Dr. Balaraman Ravindran and Dr. Sutanu Chakraborti.
My research interests broadly revolves around the fields of Generative Modeling, Epidemic Forecasting, AI for public health, Reinforcement Learning and Deep Learning. My current research interests are:
- Generative modeling for multi-source and time-series forecasting with focus on uncertainty quantification and calibration.
- Spatio-temporal modeling with deep learning methods that are robust to noisy data
- Epidemic modeling and forecasting
- Reinforcement Learning in real-world including Network problems like graph exploration, influence maximization.
|Mar 1, 2023||PROFHIT: Probabilistic Robust Forecasting for Hierarchical Time-series on state-of-art hierarchical probabilistic forecasting is accepted at KDD 2023!|
|Feb 1, 2022||Our paper CAMul: Calibrated and Accurate Multi-view Time-Series Forecasting on State-of-Art multi-modal calibrated forecasting is accepted at WWW 2022!|
|Feb 1, 2022||Back2Future: Leveraging Backfill Dynamics for Improving Real-time Predictions in Future on leveraging revision data to imporve real-time forecasting and model evaluation is accepted at ICLR 2022!|
|Oct 1, 2021||Our paper When in Doubt: Neural Non-Parametric Uncertainty Quantification for Epidemic Forecasting on State-of-Art calibrated and explainable Epidemic forecasting is accepted at NeurIPS 2021!|
|Jan 1, 2021||Starting PhD in Machine Learning at Georgia Tech at AdityaLab advised by Dr B Aditya Prakash!|
PROFHIT: Probabilistic Robust Forecasting for Hierarchical Time-seriesKDD 2023
CAMul: Calibrated and Accurate Multi-view Time-Series ForecastingACM Web Conference (WWW) 2022
Back2Future: Leveraging Backfill Dynamics for Improving Real-time Predictions in FuturePreprint 2021
When in Doubt: Neural Non-Parametric Uncertainty Quantification for Epidemic ForecastingThirty-fifth Conference on Neural Information Processing Systems (NeurIPS) 2021
Reinforcement Learning for Unified Allocation and Patrolling in Signaling Games with Uncertainty20th International Conference on Autonomous Agents and Multiagent Systems (AAMAS) 2020
Influence maximization in unknown social networks: Learning Policies for Effective Graph Sampling19th International Conference on Autonomous Agents and Multiagent Systems (AAMAS), Nominated for Best Paper Award<\b> 2019