Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Plastic Learning with Deep Fourier Features

About

Deep neural networks can struggle to learn continually in the face of non-stationarity. This phenomenon is known as loss of plasticity. In this paper, we identify underlying principles that lead to plastic algorithms. In particular, we provide theoretical results showing that linear function approximation, as well as a special case of deep linear networks, do not suffer from loss of plasticity. We then propose deep Fourier features, which are the concatenation of a sine and cosine in every layer, and we show that this combination provides a dynamic balance between the trainability obtained through linearity and the effectiveness obtained through the nonlinearity of neural networks. Deep networks composed entirely of deep Fourier features are highly trainable and sustain their trainability over the course of learning. Our empirical results show that continual learning performance can be drastically improved by replacing ReLU activations with deep Fourier features. These results hold for different continual learning scenarios (e.g., label noise, class incremental learning, pixel permutations) on all major supervised learning datasets used for continual learning research, such as CIFAR10, CIFAR100, and tiny-ImageNet.

Alex Lewandowski, Dale Schuurmans, Marlos C. Machado• 2024

Related benchmarks

TaskDatasetResultRank
Continual Supervised LearningCIFAR 5+1
Total Average Online Task Accuracy72.29
49
Continual Supervised LearningCIFAR Random Label
Total Average Online Task Accuracy96.24
49
Continual Supervised LearningContinual ImageNet
Total Average Online Task Accuracy76.03
49
Continual LearningPermuted MNIST--
32
Continual LearningMNIST Shuffled Labels
Accuracy (ACC)92.61
22
Plasticity MeasurementLocomotion Tasks Aggregate (Ant, HalfCheetah, Humanoid) (train)
Plasticity Score (IQM)17.66
17
Continual Reinforcement LearningMiniGrid
Mean SR38.8
10
Showing 7 of 7 rows

Other info

Follow for update