2021 Call for Code Awards: Live from New York, with SNL’s Colin Jost! Learn more

Hyperparameter Tuning with Bayesian Optimization

Bayesian Optimization is widely used in Machine Learning for tuning the hyperparameters such that they result in the best performance for a given problem. It is efficient and effective in helping us find near to an optimal solution for a machine learning model in few iterations.

In the container landscape, for an application deployed to the cloud, it is important that we set optimal values for resource parameters such as limits and requests. This can be done with the help of Bayesian Optimization. The resource limit and request would be the hyperparameters and the throughput or response time of the application would be the value that is being optimized.

In this session, we will cover the main components of Bayesian Optimization, look at why Bayesian Optimization is preferred over some other techniques used for hyperparameter tuning, how we can use it to optimise our containers in the cloud and the different ML libraries available that help us perform Bayesian Optimization.