Alternating Bi-Objective Optimization for Explainable Neuro-Fuzzy Systems
This addresses the problem of balancing accuracy and explainability in AI systems for domains requiring interpretable models, though it is incremental relative to existing MOO and scalarization methods.
The paper tackles the accuracy-explainability trade-off in neuro-fuzzy systems by proposing X-ANFIS, an alternating bi-objective gradient-based optimization scheme that uses Cauchy membership functions and a differentiable explainability objective. In approximately 5,000 experiments on nine UCI regression datasets, it consistently achieved target distinguishability while maintaining competitive predictive accuracy and recovering solutions beyond the convex hull of the MOO Pareto front.
Fuzzy systems show strong potential in explainable AI due to their rule-based architecture and linguistic variables. Existing approaches navigate the accuracy-explainability trade-off either through evolutionary multi-objective optimization (MOO), which is computationally expensive, or gradient-based scalarization, which cannot recover non-convex Pareto regions. We propose X-ANFIS, an alternating bi-objective gradient-based optimization scheme for explainable adaptive neuro-fuzzy inference systems. Cauchy membership functions are used for stable training under semantically controlled initializations, and a differentiable explainability objective is introduced and decoupled from the performance objective through alternating gradient passes. Validated in approximately 5,000 experiments on nine UCI regression datasets, X-ANFIS consistently achieves target distinguishability while maintaining competitive predictive accuracy, recovering solutions beyond the convex hull of the MOO Pareto front.