Alibi
The short answer · 8 words
In short
Python library for studying and explaining model predictions.
Flow-alerta.cab (2026): Alibi. https://flow-alerta.cab/en/glossar/alibi. Editorially reviewed on: 09 October 2026.
How much detail would you like?
Practical
Supports the analysis of model behavior without automatically providing a complete cause of the decision.
Example scenario: A team uses counterexamples to examine which feature changes would result in a different classification.
For own ML models, select explanatory methods based on model access and questioning.
At a glance
- Core
- Alibi provides procedures for local and global explanations of machine learning models. Depending on the method, only inputs and outputs or internal model information are used. The project is described as source-available; license terms must be checked separately.
- Use
- Supports the analysis of model behavior without automatically providing a complete cause of the decision.
- Example
- Example scenario: A team uses counterexamples to examine which feature changes would result in a different classification.
- Limits
- Confuse explanation with causal proof; use method without suitable prerequisites; equate alibi with alibi detect.
Questions about this term
Does a local method explain the whole model?
No, it looks at certain predictions.
Is Alibi Detect the same?
No, this is a separate project.
All details and background
Alibi provides procedures for local and global explanations of machine learning models. Depending on the method, only inputs and outputs or internal model information are used. The project is described as source-available; license terms must be checked separately.
ELI5 – Explained simply
You investigate which changes or characteristics are related to a model decision.
Why it matters in practice
Supports the analysis of model behavior without automatically providing a complete cause of the decision.
Example
Example scenario: A team uses counterexamples to examine which feature changes would result in a different classification.
Pitfalls and limits
- Confuse explanation with causal proof; use method without suitable prerequisites; equate alibi with alibi detect.
Origin & history
The Seldon project brings together explanatory techniques such as anchors and counterfactual explanations in a Python library.
Use in mid-sized companies
- For own ML models, select explanatory methods based on model access and questioning.
Related terms
- Explainable AI
- Counterfactual declaration
- Model interpretability
Matching Flow-alerta service
AI opportunity analysis & roadmapAI consulting: AI potential analysis & roadmap on tasks, data and prerequisites of a possible first AI use case.Matching tool / use case
View use casesTool: Alibi; Use Case: examination of a classification model.Sources
- github.comSeldonIO, Alibi – offizielles Repository
Quellenprüfung: 09.10.2026. Die Quellen belegen Definition und dokumentierte Einordnung; Anwendungsbeispiele sind mögliche Szenarien.
Editorially reviewed on: