Reinforcement Learning Coach + AWS SageMaker = Accelerated Deep Reinforcement Learning

At this year’s AWS re:Invent conference,  Amazon announced the integration of Intel’s Reinforcement Learning (RL) library, Reinforcement Learning Coach, with the Amazon SageMaker* platform for machine learning. SageMaker is an AI service that enables developers to build and train machine learning models for predictive or analytical applications in AWS’s cloud environment.

Reinforcement Learning Coach is an open source research framework for training and evaluating reinforcement learning agents. Coach contains implementations for more than 20 of today’s leading RL algorithms, combined with various games and robotics training environments. It enables efficient training of reinforcement learning agents by harnessing the power of multi-core CPU processing. The integration of Coach with Sagemaker will enable developers and researchers to experiment with and deploy RL models while taking advantage of AWS’ scalable computing infrastructure.

Reinforcement Learning Coach provides baseline implementations with reproducibility benchmarks for the latest RL algorithms. For developers who want to build new algorithms, Coach can also be used as a sandbox for simplifying development by leveraging its components in the framework. With its set of APIs and key components for reinforcement learning, Coach enables developers to easily reuse these components to build new algorithms on top of existing ones.  The framework also provides visualization and debugging tools that facilitate research and data science use case development. And now, this framework is integrated with SageMaker for a seamless experience when running distributed training and simulation models.

Reinforcement Learning on SageMaker

For developers,  the integration of Reinforcement Learning Coach with Amazon SageMaker is a recipe for success. Now they can take advantage of AWS C5 instances based on Intel® Xeon® Scalable Processors to run advanced compute-heavy workloads – like reinforcement learning models.

SageMaker/C5 instances have been fully optimized to quickly and easily build and deploy machine learning (ML) models. Amazon SageMaker is pre-configured with the latest Intel-optimized versions of TensorFlow* and Apache MXNet* for maximum performance on C5 instances running on Intel Xeon Platinum Processors, so developers can quickly train deep learning models.

In addition, developers working on sequential decision-making problems can choose from many algorithms within SageMaker that use Coach. Together with SageMaker, Reinforcement Learning Coach offers various techniques for visualizing the training process and understanding the underlying mechanisms of the agents. All the algorithms are implemented using C5 and offer a scalable solution for Industrial RL applications.

We are very excited that this collaboration between Intel and Amazon will further our mission of making machine learning more accessible. Now developers, researchers, and students will have access to world-class hardware and tools from Intel running on Amazon’s machine learning platform to quickly and easily build intelligent applications using deep RL.

Intel technologies’ features and benefits depend on system configuration and may require enabled hardware, software or service activation. Performance varies depending on system configuration. No computer system can be absolutely secure. Check with your system manufacturer or retailer or learn more at  
Intel, the Intel logo, and Xeon are trademarks of Intel Corporation or its subsidiaries in the U.S. and/or other countries.
 *Other names and brands may be claimed as the property of others.
 © Intel Corporation