Comparing Models in Production with Multi-Armed Bandits and Reinforcement Learning [Level 300]

June 9, 2021
Using the popular Hugging Face Transformers open source library for BERT, I will train and deploy multiple natural language understanding (NLU) models and compare them in live production using reinforcement learning to dynamically shift traffic to the winning model. Next, I will describe the differences between A/B and multi-armed bandit tests including exploration-exploitation, reward-maximization, and regret-minimization. Last, I will dive deep into the details of scaling a multi-armed bandit architecture on AWS using a real-time, stream-based text classifier with TensorFlow, PyTorch, and BERT on 150+ million reviews from the Amazon Customer Reviews Dataset.

Speaker: Chris Fregly, AWS Senior Advocate, AI/ML
Previous Video
Orchestrate and Automate Machine Learning Workflows with SageMaker Pipelines [Level 300]
Orchestrate and Automate Machine Learning Workflows with SageMaker Pipelines [Level 300]

In this session, you'll see how to create, automate, and manage end-to-end ML workflows using Amazon SageMa...

Next Video
Reduce Training Time and Cost with SageMaker Debugger  [Level 300]
Reduce Training Time and Cost with SageMaker Debugger [Level 300]

In this session, we walk through how to use the real-time training metrics and set up alerts so you can red...