LEARN: A Unified Framework for Multi-Task Domain Adapt Few-Shot Learning

TL;DR

This paper introduces LEARN, a comprehensive, modular framework that unifies domain adaptation and few-shot learning across multiple computer vision tasks, supporting various configurations including SSL pre-training and incremental learning.

Contribution

It presents the first unified, flexible framework combining domain adaptation and few-shot learning for classification, detection, and video tasks, with on-the-fly setup and extensive benchmarking.

Findings

Supports multiple tasks and algorithms

Enables incremental n-shot learning

Provides extensive benchmarks across datasets

Abstract

Both few-shot learning and domain adaptation sub-fields in Computer Vision have seen significant recent progress in terms of the availability of state-of-the-art algorithms and datasets. Frameworks have been developed for each sub-field; however, building a common system or framework that combines both is something that has not been explored. As part of our research, we present the first unified framework that combines domain adaptation for the few-shot learning setting across 3 different tasks - image classification, object detection and video classification. Our framework is highly modular with the capability to support few-shot learning with/without the inclusion of domain adaptation depending on the algorithm. Furthermore, the most important configurable feature of our framework is the on-the-fly setup for incremental $n$-shot tasks with the optional capability to configure the system to scale to a traditional many-shot task. With more focus on Self-Supervised Learning (SSL) for current few-shot learning approaches, our system also supports multiple SSL pre-training configurations. To test our framework's capabilities, we provide benchmarks on a wide range of algorithms and datasets across different task and problem settings. The code is open source has been made publicly available here: https://gitlab.kitware.com/darpa_learn/learn