Towards Relatable Explainable AI with the Perceptual Process

Abstract

Machine learning models need to provide contrastive explanations, since people often seek to understand why a puzzling prediction occurred instead of some expected outcome. Current contrastive explanations are rudimentary comparisons between examples or raw features, which remain difficult to interpret, since they lack semantic meaning. We argue that explanations must be more relatable to other concepts, hypotheticals, and associations. Inspired by the perceptual process from cognitive psychology, we propose the XAI Perceptual Processing Framework and RexNet model for relatable explainable AI with Contrastive Saliency, Counterfactual Synthetic, and Contrastive Cues explanations. We investigated the application of vocal emotion recognition, and implemented a modular multi-task deep neural network to predict and explain emotions from speech. From think-aloud and controlled studies, we found that counterfactual explanations were useful and further enhanced with semantic cues, but not saliency explanations. This work provides insights into providing and evaluating relatable contrastive explainable AI for perception applications.

arXiv comment14 pages, 7 figures, 4 tables, accepted by chi2022

References in corpus (14)

• Axiomatic Attribution for Deep Networks
• Deep Speech 2: End-to-End Speech Recognition in English and Mandarin
• Explaining NonLinear Classification Decisions with Deep Taylor Decomposition
• Understanding Black-box Predictions via Influence Functions
• Explaining Machine Learning Classifiers through Diverse Counterfactual Explanations
• Interpretable classifiers using rules and Bayesian analysis: Building a better stroke prediction model
• Questioning the AI: Informing Design Practices for Explainable AI User Experiences
• Expanding Explainability: Towards Social Transparency in AI systems
• Interpretability Beyond Feature Attribution: Quantitative Testing with Concept Activation Vectors (TCAV)
• Deep Convolutional Neural Networks for Breast Cancer Histology Image Analysis
• Contrastive Explanation: A Structural-Model Approach
• Towards Automatic Concept-based Explanations
• Interpretable Representation Learning for Speech and Audio Signals Based on Relevance Weighting
• Proceedings of the 2018 ICML Workshop on Human Interpretability in Machine Learning (WHI 2018)

Cited by in corpus (9)

• Towards Human-centered Explainable AI: A Survey of User Studies for Model Explanations
• "Help Me Help the AI": Understanding How Explainability Can Support Human-AI Interaction
• Sensible AI: Re-imagining Interpretability and Explainability using Sensemaking Theory
• Interpretable Directed Diversity: Leveraging Model Explanations for Iterative Crowd Ideation
• The AI-DEC: A Card-based Design Method for User-centered AI Explanations
• User Experience with LLM-powered Conversational Recommendation Systems: A Case of Music Recommendation
• Towards Interpretability in Audio and Visual Affective Machine Learning: A Review
• XAIxArts Manifesto: Explainable AI for the Arts
• Towards Uncertainty Aware Task Delegation and Human-AI Collaborative Decision-Making