D3MC: A Reinforcement Learning Based Data-Driven Dyna Model Compression.

Artificial intelligence (AI)-driven medical devices have created a new excitement in healthcare sector. While deeper and wider neural networks are designed for complex healthcare applications, model compression can be an effective way to deploy networks on medical devices that often have hardware and speed constraints. Most state-of-the-art model compression techniques require a resource centric manual process that explores a large model architecture space to find a trade-off solution between model size and accuracy. Recently, reinforcement learning (RL) approaches are proposed to automate such a hand-crafted process. However, most RL model compression algorithms are model-free that require longer time with no assumptions of the model. On the contrary, model-based (MB) approaches are data driven; have faster convergence but are sensitive to the bias in the model. In this paper, we develop data-driven dyna model compression (D3MC) algorithm that integrates model-based and model-free RL approaches. We evaluate our algorithm on a variety of imaging data from dermoscopy to X-ray on different popular and public model architectures. Compared to model-free RL approaches, our approach achieves faster convergence; exhibits better generalization across different data sets; and preserves comparable model performance.