Skip to main content
​
2021-11-02

​
What’s New

Page TitlesIntuitive page titles to stay organized while working with multiple tabs!

​
Enhancements

Python SDK 3.1.3A new set of client-side validations and options for automatic data cleaning and an improved troubleshooting workflow for data sent to the Arize platform. With Python SDK 3.1.3, improve clarity on how to navigate errors amongst newly ingested data.Default ActualsIncreased support for models without actuals. Treat predictions without actuals as having implicit negative class actuals when calculating certain performance metrics such as: AUC, PR-AU, and Log Loss with new default actuals.

​
In the News

Continuous Monitoring, Continuous Improvements for ML Models Using Neptune AI and Arize AIWe’ve partnered with Neptune AI to consistently deploy — and maintain — higher-quality models in production!** Connect Arize’s ML observability platform with Neptune’s metadata store to streamline your MLOps workflow. Learn how to leverage the power of both platforms to enable better ML Ops: **Read More ****
Best Practices In ML Observability for Monitoring, Mitigating, and Preventing FraudFraud is a burden on the global economy annually to the tune of $5 trillion and threats are evolving and taking new forms. Learn how data science and ML teams can develop an observability strategy to catch, monitor, and troubleshoot problems with their fraud models in production.** Read more**
Arize AI Selected by Gartner as Cool Vendor in Enterprise AI Operationalization and Engineering ReportArize AI Selected by Gartner as Cool Vendor in Enterprise AI Operationalization and Engineering ReportWe’ve been named an Enterprise Cool Vendor by Gartner who believes Arize should be on the radar of “enterprises seeking to maximize ROI and have visibility into how the model impacts your business bottom line.” Check out Gartner’s website to get a copy of the full report.Read our release: