alibi-detect

What problem does it solve?

This Skill helps identify potential issues in AI model performance by detecting outliers, adversarial attacks, and shifts in data or concepts over time.

Core Features & Use Cases

• Outlier Detection: Identify unusual data points that deviate from the norm.
• Adversarial Instance Detection: Spot inputs designed to fool the AI model.
• Data/Concept Drift Detection: Monitor for changes in the input data distribution or the underlying concept the model is trying to predict.
• Use Case: A financial institution can use this skill to monitor its fraud detection model for signs that fraudsters are adapting their techniques (adversarial instances) or that the nature of fraudulent transactions is changing (concept drift).

Quick Start

Use the alibi-detect skill to check for outliers in the provided dataset.