Continual Learning in Neural Networks

Abstract

Artificial neural networks have exceeded human-level performance in accomplishing several individual tasks (e.g. voice recognition, object recognition, and video games). However, such success remains modest compared to human intelligence that can learn and perform an unlimited number of tasks. Humans' ability of learning and accumulating knowledge over their lifetime is an essential aspect of their intelligence. Continual machine learning aims at a higher level of machine intelligence through providing the artificial agents with the ability to learn online from a non-stationary and never-ending stream of data. A key component of such a never-ending learning process is to overcome the catastrophic forgetting of previously seen data, a problem that neural networks are well known to suffer from. The work described in this thesis has been dedicated to the investigation of continual learning and solutions to mitigate the forgetting phenomena in neural networks. To approach the continual learning problem, we first assume a task incremental setting where tasks are received one at a time and data from previous tasks are not stored. Since the task incremental setting can't be assumed in all continual learning scenarios, we also study the more general online continual setting. We consider an infinite stream of data drawn from a non-stationary distribution with a supervisory or self-supervisory training signal. The proposed methods in this thesis have tackled important aspects of continual learning. They were evaluated on different benchmarks and over various learning sequences. Advances in the state of the art of continual learning have been shown and challenges for bringing continual learning into application were critically identified.

PhD Thesis, Supervisor: Tinne Tuytelaars

References in corpus (27)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Distilling the Knowledge in a Neural Network
• Overcoming catastrophic forgetting in neural networks
• ADADELTA: An Adaptive Learning Rate Method
• Playing Atari with Deep Reinforcement Learning
• DeCAF: A Deep Convolutional Activation Feature for Generic Visual Recognition
• Fine-Grained Visual Classification of Aircraft
• Riemannian Walk for Incremental Learning: Understanding Forgetting and Intransigence
• One weird trick for parallelizing convolutional neural networks
• An Empirical Investigation of Catastrophic Forgetting in Gradient-Based Neural Networks
• Hierarchical modularity in human brain functional networks
• PathNet: Evolution Channels Gradient Descent in Super Neural Networks
• A Survey on Multi-Task Learning
• Efficient Lifelong Learning with A-GEM
• Bayesian Compression for Deep Learning
• New insights and perspectives on the natural gradient method
• Learning Task Grouping and Overlap in Multi-task Learning
• Re-evaluating Continual Learning Scenarios: A Categorization and Case for Strong Baselines
• Towards Robust Evaluations of Continual Learning
• A PAC-Bayesian bound for Lifelong Learning
• Lifelong Generative Modeling
• On Quadratic Penalties in Elastic Weight Consolidation
• Regularizing CNNs with Locally Constrained Decorrelations
• Reducing Overfitting in Deep Networks by Decorrelating Representations
• Continual Classification Learning Using Generative Models
• Positivity theorems for solid-angle polynomials
• Online Deep Learning: Growing RBM on the fly

Cited by in corpus (6)

• Quantum Continual Learning Overcoming Catastrophic Forgetting
• Imbalanced Continual Learning with Partitioning Reservoir Sampling
• Monitoring multimode processes: a modified PCA algorithm with continual learning ability
• Continual Learning: Tackling Catastrophic Forgetting in Deep Neural Networks with Replay Processes
• Sparse Meta Networks for Sequential Adaptation and its Application to Adaptive Language Modelling
• IB-DRR: Incremental Learning with Information-Back Discrete Representation Replay