Tiny Time Mixers (TTMs): Fast Pre-trained Models for Enhanced Zero/Few-Shot Forecasting of Multivariate Time Series
- Vijay E
- Arindam Jati
- et al.
- 2024
- NeurIPS 2024
Vijay Ekambaram is a Senior Research Leader and Master Inventor at IBM Research-India, specializing in building Foundation Models for Time-Series and advancing generative AI models. He is the lead author of several popular time-series foundation models, including TTM and PatchTSMixer, that are featured on HuggingFace with several million downloads and highlights in media articles such as Forbes, Techopedia, etc.
Vijay Ekambaram has over 270+ USPTO-granted inventions, making him among India’s Top 10 inventors and the Top 1,000 globally for the most patents awarded to an individual. In recognition of his contributions, he received the Worldwide Prolific Inventor award in 2023 and is listed among the top three Indian inventors on Wikipedia's Prolific Inventor List. He is also a lead author of several publications presented in top-tier AI conferences such as NeurIPS, KDD, AAAI, and BigData.
Vijay holds an MTech in Computer Science from IIT Madras, where he was a gold medalist and also received the best thesis award for his academic excellence. He currently leads a team of researchers at IRL, driving innovation in time-series foundation models and building scalable Deep-learning products.