Researchers have introduced ExplAIner, a new declarative query language designed to uniformly specify, combine, and analyze various notions for explaining Boolean classification models. This language extends the FOIL framework to address limitations in expressing optimality-based queries and improves evaluation complexity. ExplAIner can express a wide array of explanation types, including abductive, contrastive, and feature-based queries, with its evaluation problem belonging to the Boolean hierarchy. An optimization-oriented fragment, Opt-FOIL, is also presented for computing minimal explanations, with its evaluation problem in FP^NP. AI
IMPACT Provides a unified framework for analyzing and specifying explanations for Boolean models, potentially simplifying XAI research.
RANK_REASON The cluster contains an academic paper detailing a new query language for explaining classification models.
AI-generated summary · Google Gemini · from 2 sources. How we write summaries →
