Continual and Few-Shot Learning

AI techniques have proven very successful in many critical applications such as object recognition, speech recognition, and others. However, these successes have relied on collecting enormous datasets and careful manual annotations. This process is expensive, time-consuming, and in many scenarios, enough data is not available. Transfer learning offers a solution to these problems by leveraging past data seen by a machine to solve future problems using only few annotated examples. This research focuses on challenges in transfer learning and aims at developing algorithms that can fundamentally learn from multiple heterogeneous tasks, moving beyond low-level task similarity to enable broader transfer across distinct tasks. Such algorithms will find general applicability in several areas, including computer vision and natural language processing, and will substantially reduce the dependence on large amounts of annotated data and consequently reduce costs and time for deployment and maintenance of AI systems.

Published Research

To learn more about Earth Intelligence Engine research and other AI Accelerator projects, view our published research here.

Are you up for a Challenge?

Modern deep learning approaches show promising results in meteorological applications like precipitation nowcasting, synthetic radar generation, front detection, and several others. The DAF-MIT AI Accelerator is working to rapidly develop new approaches to these challenges. The AIA is seeking collaborators from across the globe to leverage a curated, "machine learning ready" dataset. Learn more about the challenge here.