10.10.2022
New Releases, Enhancements, Changes + Arize in the News!
What's New
File Importer: Microsoft Azure Support
Send data through Azure blob storage to automatically extract data from Azure to Arize. Seamlessly log model inferences in a few simple steps: connect your Azure storage account, set up containers, and enter your secret key to begin monitoring and troubleshooting with Arize.
Learn how to set up an import job from Azure here.
File Importer: Embeddings Parquet Support
Ingest embedding features through the file importer for a simple setup flow within the Arize UI. Specify the name of your embedding feature(s), the associated vector column name, and optional raw data column names in your parquet file. From there, your embeddings are ready for drift analysis, UMAP, and further analysis.
Stay tuned for embeddings Avro support. Coming soon!
Enhancements
Color by Feature
Color the UMAP point cloud by feature and view prediction details to uncover a more detailed view of your embedding drift.
Notes for Individual Monitors
Annotate any monitor for enhanced organization and increased transparency. Use notes to attach an explanation of your monitor's details, add an alert resolution ledger, link to run-books in the event of an alert, and more.
In The News
Ever want a quick-and-intuitive explanation of PR AUC? Then this blog is for you — animate the calculation process, learn when it's useful, and tease out the difference between PR-AUC and ROC-AUC!
Using Arize and OpenAI together can help organizations better build, monitor, and troubleshoot unstructured models in production, reducing costs and maximizing performance. Follow along with a code example that covers how to use OpenAI and Arize AI together for an NLP use case in this informative and helpful post.
We recently launched a year-round bug bounty program facilitated by Yogosha! Find security issues and get rewarded by joining the Yogosha Strick Force. Read more on Arize CISO, Remi Cattiau's, thoughts when partnering with Yogosha.
Last updated
