Skip to main content
​
2021-06-25

​
What’s New

Projects - Organize and Manage your ModelsTired of manually sorting through tons of models to find the one you’re looking for? Want an easy way to organize and compare related models? You’re in luck! Look no further than our new Projects page. We’ve added the ability to group models together for easy navigation and comparative analysis. Some situations where projects are especially useful are when you have a bunch of models serving the same goal, one model instance deployed and for one store, similar models in different locations, or multiple types of users.

Projects View

Business Impact for Pre-production DatasetsWe have updated the Business Impact Tab to now support comparisons between production and pre-production curves! Simply select the pre-production dataset in the drop down menu to visualize your production and pre-production curves on the same chart. Toggle any of the curves on or off by clicking their respective label on the chart to visualize just one curve at a time.

Compare business impact of different datasets

​
Enhancements

Streamlined Monitor StatusWe’ve improved our monitor statuses to be more intuitive and expansive. Hovering over them provides you with a brief description of how your model is doing. From there you can directly click into the monitor or drift page to immediately begin digging into the issue.

New statuses for monitors

New Additions to Drift TabsWith our new drift visualization at the top of the Drift tab, you are able to easily cross-reference drift with metrics such as model traffic and accuracy. Additionally, we’ve added a feature drift summary with metrics such as Prediction Drift Impact, Feature Importance, and PSI. Analyzing drift in your models has never been easier!

Drift Page

On our last release, we added the ability to visualize data quality metrics when clicking into features. Now we’ve added in this super useful drift visualization as well!

Feature Drift Tab

​
In the News

Blog Post: Take My Drift AwayIn this joint piece with Hua Ai, Data Science Manager at Delta Airlines, we dive into drift, why it’s important to keep track of, and how to troubleshoot and resolve the underlying issue when drift occurs. Read the article.

Drift - distribution changes over time

Arize is Growing!We’re thrilled to welcome our new team members to the Arize AI team! We are also hiring new roles, be sure to check out our Linkedin page for more updates. Meet the Team.
Arize Partners with UbiOpsWhile deployment of a production-worthy AI model poses a challenge to many, observability is another, deeper challenge that awaits a model in production. With Arize AI and UbiOps, data scientists and ML engineers can work together to develop a model, push it to production, and gain full visibility and control of its performance. Read the article.