Continual Learning via Local Module Composition

11/15/2021
by   Oleksiy Ostapenko, et al.
5

Modularity is a compelling solution to continual learning (CL), the problem of modeling sequences of related tasks. Learning and then composing modules to solve different tasks provides an abstraction to address the principal challenges of CL including catastrophic forgetting, backward and forward transfer across tasks, and sub-linear model growth. We introduce local module composition (LMC), an approach to modular CL where each module is provided a local structural component that estimates a module's relevance to the input. Dynamic module composition is performed layer-wise based on local relevance scores. We demonstrate that agnosticity to task identities (IDs) arises from (local) structural learning that is module-specific as opposed to the task- and/or model-specific as in previous works, making LMC applicable to more CL settings compared to previous works. In addition, LMC also tracks statistics about the input distribution and adds new modules when outlier samples are detected. In the first set of experiments, LMC performs favorably compared to existing methods on the recent Continual Transfer-learning Benchmark without requiring task identities. In another study, we show that the locality of structural learning allows LMC to interpolate to related but unseen tasks (OOD), as well as to compose modular networks trained independently on different task sequences into a third modular network without any fine-tuning. Finally, in search for limitations of LMC we study it on more challenging sequences of 30 and 100 tasks, demonstrating that local module selection becomes much more challenging in presence of a large number of candidate modules. In this setting best performing LMC spawns much fewer modules compared to an oracle based baseline, however, it reaches a lower overall accuracy. The codebase is available under https://github.com/oleksost/LMC.

READ FULL TEXT

page 8

page 23

page 25

page 26

page 29

research
04/04/2023

I2I: Initializing Adapters with Improvised Knowledge

Adapters present a promising solution to the catastrophic forgetting pro...
research
06/11/2023

A Probabilistic Framework for Modular Continual Learning

Modular approaches, which use a different composition of modules for eac...
research
10/18/2022

Exclusive Supermask Subnetwork Training for Continual Learning

Continual Learning (CL) methods mainly focus on avoiding catastrophic fo...
research
12/23/2020

Efficient Continual Learning with Modular Networks and Task-Driven Priors

Existing literature in Continual Learning (CL) has focused on overcoming...
research
03/20/2022

Continual Sequence Generation with Adaptive Compositional Modules

Continual learning is essential for real-world deployment when there is ...
research
11/20/2017

Modular Continual Learning in a Unified Visual Environment

A core aspect of human intelligence is the ability to learn new tasks qu...
research
04/06/2022

Georeferencing of Photovoltaic Modules from Aerial Infrared Videos using Structure-from-Motion

To identify abnormal photovoltaic (PV) modules in large-scale PV plants ...

Please sign up or login with your details

Forgot password? Click here to reset