Single-Net Continual Learning with Progressive Segmented Training (PST)
arXiv:1905.11550
Abstract
There is an increasing need of continual learning in dynamic systems, such as the self-driving vehicle, the surveillance drone, and the robotic system. Such a system requires learning from the data stream, training the model to preserve previous information and adapt to a new task, and generating a single-headed vector for future inference. Different from previous approaches with dynamic structures, this work focuses on a single network and model segmentation to prevent catastrophic forgetting. Leveraging the redundant capacity of a single network, model parameters for each task are separated into two groups: one important group which is frozen to preserve current knowledge, and secondary group to be saved (not pruned) for a future learning. A fixed-size memory containing a small amount of previously seen data is further adopted to assist the training. Without additional regularization, the simple yet effective approach of PST successfully incorporates multiple tasks and achieves the state-of-the-art accuracy in the single-head evaluation on CIFAR-10 and CIFAR-100 datasets. Moreover, the segmented training significantly improves computation efficiency in continual learning.
References in corpus (18)
- Very Deep Convolutional Networks for Large-Scale Image Recognition
- Overcoming catastrophic forgetting in neural networks
- Quantized Neural Networks: Training Neural Networks with Low Precision Weights and Activations
- Pruning Convolutional Neural Networks for Resource Efficient Inference
- Riemannian Walk for Incremental Learning: Understanding Forgetting and Intransigence
- Compressing Deep Convolutional Networks using Vector Quantization
- Continual Learning Through Synaptic Intelligence
- Rethinking the Value of Network Pruning
- Pruning Filters for Efficient ConvNets
- Variational Continual Learning
- Overcoming Catastrophic Forgetting by Incremental Moment Matching
- Overcoming catastrophic forgetting with hard attention to the task
- Reinforced Continual Learning
- Lifelong Learning with Dynamically Expandable Networks
- Towards Robust Evaluations of Continual Learning
- Gradient based sample selection for online continual learning
- Class-incremental Learning via Deep Model Consolidation
- Regularize, Expand and Compress: Multi-task based Lifelong Learning via NonExpansive AutoML