Integrated Deployment KNIME Blog Articles

Why is deploying data science in production so hard? At first glance, putting data science in production seems trivial: Just run it on the production server or chosen device! But on closer examination, it becomes clear that what was built during data science creation is not what is being put into production.

What are the challenges you face when moving from creating a model to using it into prodution? Data sources and types change rapidly as do the methods for their analysis. This article provides an introduction to the Integrated Deployment approach and dives in deeper to look at the workflows involved in model creation and model deployment.

Find out how Integrated Deployment is used to train multiple models. How can we flexibly deploy the best one? How can we retrain and redeploy on a biggerdataset? This article looks at how Integrated Deployment can help accomplish these tasks.

Wouldn't it be great to have an Integrated Deployment strategy that automatically updates when more data are added or you want to change the prediction focus. AutoML is the topic here and we look at an AutoML component designed to flexibly automate training, validation, and deployment of different machine learning algorithms.

In this article, we explain how to build a guided analytics application around the AutoML Component to give the business user an easy process to automatically train machine learning models.

The innovative character of the integrated deployment technique changes the "moving into production" into a faithful "capturing" of the relevant data computations directly from model training. Here, we describe a range of  use cases from automatic generation of a production workflow through to building a black-box predictor.

This article discusses the importance of monitoring model performance in the data science context and demonstrates a KNIME component that has been built to monitor the ML model and re-evaluate its performance during production.