Documentation Index
Fetch the complete documentation index at: https://arize-ax.mintlify.dev/docs/llms.txt
Use this file to discover all available pages before exploring further.
What’s New
September 26, 2022Alerting Integration: Direct Monitor Setup
Set up an alerting integration directly from the ‘Edit Monitor’ page to streamline your monitoring workflow. This enables increased flexibility to configure an integration specific to your alerting needs.Pick from our native integrations with Pagerduty, Slack, or OpsGenie.Do you use an alerting system not represented yet? Contact support@arize.com or reach out to us on Slack for options.
- Streamline your monitoring workflow per model or individual monitor
- Reduce time to resolution with custom model dimensions and enhanced metadata
- Stay organized within your system and across multiple stakeholders
Use our public GraphQL API to programmatically configure your alerting integration within your infrastructure. Learn more here.
Tags on Fairness
September 12, 2022 Filter by tags in the ‘Fairness’ tab to catch sensitive attributes passed as tags before they contribute to algorithmic harm. Follow the same troubleshooting flow for features with tags to see how your model’s metadata affects your fairness metric. Learn how to use our bias tracing tool here.
Drift Metric Selector
VisualizePSI, KL Divergence, and JS Distance in the ‘Drift’ tab and within each feature in the ‘Feature Drift’ card. The drift metric selector allows you to better understand your model’s behavior across different drift metrics for simplified troubleshooting.
Enhancements
September 26, 2022Drift Monitor: Delay Evaluation
Accurately evaluate drift on delayed data by configuring your ‘Delay Evaluation By’ setting. Navigate to the ‘Edit Monitor’ page under ‘Custom Settings’ to pick from a wide breadth of time windows relevant to your data. Learn more about an evaluation window here.
Drift Metric: KS Statistic
September 12, 2022 KS test statistic is a drift measurement that quantifies the maximum distance between two cumulative distribution functions. KS test is an efficient and general way to measure if two distributions significantly differ from one another.PSI and rank ordering tests focus more on how a population may have shifted between development and validation periods, while KS statistic is used to assess the predictive capability and performance of a model.
In The News
September 26, 2022Arize’s ML Observability Platform Is Now Available on Google Cloud Marketplace
Arize is now available on Google Cloud Marketplace. This availability marks the expansion of the company’s partnership with Google Cloud, which will help Arize deliver its platform – tracking billions of model predictions daily – to more customers globally.Shipping NLP Sentiment Classification Models With Confidence
Learn how to use Hugging Face and Arize to ship NLP sentiment classification models with confidence. Dive into how to ingest embedding data on Arize and how to look at embedding drift.
Six Takeaways From Our Event On the Evolution of the Data Stack
Recently, Arize hosted Monte Carlo’s co-founder and CTO in a session on “The Evolution of the Data Stack.” Miss the event? Here are six top takeaways…
The Death of Central ML Is Greatly Exaggerated
Debates on the ideal team structure for ML organizations are heating up, from Tecton’s recent piece arguing that “centralized machine learning teams fail” to Meta’s recent embrace of a decentralized approach. This article urges teams not to throw in the towel on centralized machine learning (ML) teams yet — offering a blueprint for when to embrace the approach and how to get it right.
[Arize Raises $38M Series B!]
September 12, 2022(https://arize.com/blog/arize-ais-next-era-of-growth/) We announced Arize’s $38 million Series B led by TCV with participation from existing investors – a record investment in the ML observability space!
Arize Receives Certifications Validating Health Information Security for HIPAA Compliance
Arize AI received certifications from an independent auditor validating that the company’s health information security program is fairly represented and includes the essential elements of HIPAA’s Security Rule and the HITECH Act!* *Learn why validating and complying with these requirements are table stakes in U.S. healthcare in this blog by Jim Groff, Compliance Officer at Arize.
Interview: Sid Roy, Devron
Sid Roy, Machine Learning Engineer at Devron, discusses how federated learning can unlock innovation while preserving privacy in this interview with Amber Roberts, Arize MLE.