DeepAI AI Chat
Log In Sign Up

Closed-Loop GAN for continual Learning

by   Amanda Rios, et al.
University of Southern California

Sequential learning of tasks using gradient descent leads to an unremitting decline in the accuracy of tasks for which training data is no longer available, termed catastrophic forgetting. Generative models have been explored as a means to approximate the distribution of old tasks and bypass storage of real data. Here we propose a cumulative closed-loop generator and embedded classifier using an AC-GAN architecture provided with external regularization by a small buffer. We evaluate incremental learning using a notoriously hard paradigm, single headed learning, in which each task is a disjoint subset of classes in the overall dataset, and performance is evaluated on all previous classes. First, we show that the variability contained in a small percentage of a dataset (memory buffer) accounts for a significant portion of the reported accuracy, both in multi-task and continual learning settings. Second, we show that using a generator to continuously output new images while training provides an up-sampling of the buffer, which prevents catastrophic forgetting and yields superior performance when compared to a fixed buffer. We achieve an average accuracy for all classes of 92.26 after 5 tasks using GAN sampling with a buffer of only 0.17 dataset size. We compare to a network with regularization (EWC) which shows a deteriorated average performance of 29.19 baseline of no regularization (plain gradient descent) performs at 99.84 (MNIST) and 99.79 tasks. Our method has very low long-term memory cost, the buffer, as well as negligible intermediate memory storage.


Integral Continual Learning Along the Tangent Vector Field of Tasks

We propose a continual learning method which incorporates information fr...

Experience Replay for Continual Learning

Continual learning is the problem of learning new tasks or knowledge whi...

Overcoming catastrophic forgetting problem by weight consolidation and long-term memory

Sequential learning of multiple tasks in artificial neural networks usin...

Online Continual Learning in Image Classification: An Empirical Survey

Online continual learning for image classification studies the problem o...

Does Continual Learning Equally Forget All Parameters?

Distribution shift (e.g., task or domain shift) in continual learning (C...

Transfer without Forgetting

This work investigates the entanglement between Continual Learning (CL) ...

Incremental Learning of Structured Memory via Closed-Loop Transcription

This work proposes a minimal computational model for learning a structur...