Thuật ngữ AI
Từ điển đầy đủ về Trí tuệ nhân tạo
Accumulated Local Effects (ALE) Plot
An alternative to PDP that calculates the effects of a feature on local prediction differences, avoiding biases from correlations between variables and offering a more reliable interpretation.
Counterfactual
An explanation in the form of a hypothetical scenario showing the minimal changes to make to the input features to change the model's prediction to a desired outcome, facilitating causal interpretation.
Global Interpretability
The ability to understand and explain the overall behavior of a model across its entire data distribution, identifying general trends and the relationships learned by the algorithm.
Local Interpretability
The ability to explain a model's prediction for a specific data instance, detailing how the features of that instance individually contributed to the final decision.
Explainable AI (XAI) Dashboard
A visualization interface integrating various explainability tools (PDP, SHAP, LIME, etc.) to allow analysts and stakeholders to monitor, diagnose, and understand the behavior of models in production.