Algorithm Design for Online Meta-Learning with Task Boundary Detection

by   Daouda Sow, et al.

Online meta-learning has recently emerged as a marriage between batch meta-learning and online learning, for achieving the capability of quick adaptation on new tasks in a lifelong manner. However, most existing approaches focus on the restrictive setting where the distribution of the online tasks remains fixed with known task boundaries. In this work, we relax these assumptions and propose a novel algorithm for task-agnostic online meta-learning in non-stationary environments. More specifically, we first propose two simple but effective detection mechanisms of task switches and distribution shift based on empirical observations, which serve as a key building block for more elegant online model updates in our algorithm: the task switch detection mechanism allows reusing of the best model available for the current task at hand, and the distribution shift detection mechanism differentiates the meta model update in order to preserve the knowledge for in-distribution tasks and quickly learn the new knowledge for out-of-distribution tasks. In particular, our online meta model updates are based only on the current data, which eliminates the need of storing previous data as required in most existing methods. We further show that a sublinear task-averaged regret can be achieved for our algorithm under mild conditions. Empirical studies on three different benchmarks clearly demonstrate the significant advantage of our algorithm over related baseline approaches.


page 1

page 2

page 3

page 4


Continuous Meta-Learning without Tasks

Meta-learning is a promising strategy for learning to efficiently learn ...

Fully Online Meta-Learning Without Task Boundaries

While deep networks can learn complex functions such as classifiers, det...

Online Meta-Learning

A central capability of intelligent systems is the ability to continuous...

Non-stationary Bandits and Meta-Learning with a Small Set of Optimal Arms

We study a sequential decision problem where the learner faces a sequenc...

Distribution Embedding Networks for Meta-Learning with Heterogeneous Covariate Spaces

We propose Distribution Embedding Networks (DEN) for classification with...

Fast and Slow Learning of Recurrent Independent Mechanisms

Decomposing knowledge into interchangeable pieces promises a generalizat...

A Channel Coding Benchmark for Meta-Learning

Meta-learning provides a popular and effective family of methods for dat...

Please sign up or login with your details

Forgot password? Click here to reset