Reinforcement Learning and Adaptive Sampling for Optimized DNN Compilation

Abstract

Achieving faster execution with shorter compilation time can enable further diversity and innovation in neural networks. However, the current paradigm of executing neural networks either relies on hand-optimized libraries, traditional compilation heuristics, or very recently, simulated annealing and genetic algorithms. Our work takes a unique approach by formulating compiler optimizations for neural networks as a reinforcement learning problem, whose solution takes fewer steps to converge. This solution, dubbed ReLeASE, comes with a sampling algorithm that leverages clustering to focus the costly samples (real hardware measurements) on representative points, subsuming an entire subspace. Our adaptive sampling not only reduces the number of samples, but also improves the quality of samples for better exploration in shorter time. As such, experimentation with real hardware shows that reinforcement learning with adaptive sampling provides 4.45x speed up in optimization time over AutoTVM, while also improving inference time of the modern deep networks by 5.6%. Further experiments also confirm that our adaptive sampling can even improve AutoTVM's simulated annealing by 4.00x.

References in corpus (14)

• MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications
• Caffe: Convolutional Architecture for Fast Feature Embedding
• Neural Architecture Search with Reinforcement Learning
• Deep Speech 2: End-to-End Speech Recognition in English and Mandarin
• MXNet: A Flexible and Efficient Machine Learning Library for Heterogeneous Distributed Systems
• AMC: AutoML for Model Compression and Acceleration on Mobile Devices
• cuDNN: Efficient Primitives for Deep Learning
• Efficient Neural Architecture Search via Parameter Sharing
• Tensor Comprehensions: Framework-Agnostic High-Performance Machine Learning Abstractions
• Learning to Optimize Tensor Programs
• Meta-Gradient Reinforcement Learning
• DeepArchitect: Automatically Designing and Training Deep Architectures
• ReLeQ: A Reinforcement Learning Approach for Deep Quantization of Neural Networks
• Importance Weighted Active Learning

Cited by in corpus (7)

• A Survey of Machine Learning for Computer Architecture and Systems
• Mind Mappings: Enabling Efficient Algorithm-Accelerator Mapping Space Search
• ConfuciuX: Autonomous Hardware Resource Assignment for DNN Accelerators using Reinforcement Learning
• ProTuner: Tuning Programs with Monte Carlo Tree Search
• AIRCHITECT: Learning Custom Architecture Design and Mapping Space
• Continual Learning Approach for Improving the Data and Computation Mapping in Near-Memory Processing System
• Spatial Sharing of GPU for Autotuning DNN models