ExplainabilityTransparencyTrustworthinesscs.AI
Natural Language Interaction with Explainable AI Models
Arjun R Akula, Sinisa Todorovic, Joyce Y Chai, Song-Chun Zhu
March 13, 2019
CVPR Workshops
23 citations
2 influential
Key Takeaways
-
Natural language makes AI explanations more accessible: The system answers user questions in plain language rather than technical outputs, making it easier for non-experts to understand why an AI system made a particular prediction.
-
Explanations go beyond simple answers: The XAI model provides three types of reasoning—direct evidence (like detection scores), part-based inference (how detected components support conclusions), and contextual evidence (surrounding information)—giving users a more complete picture of the AI's decision-making process.
-
AI can explain its reasoning in multiple ways: By correlating different types of user questions with appropriate explanation types, the system demonstrates that there's no single "right" way to explain AI decisions; different users may need different types of evidence to understand and trust the system's predictions.
-
Explainability requires structured AI design: The two-component architecture (prediction model + explanation model) suggests that building trustworthy AI systems requires intentionally designing explainability into the system from the start, rather than treating it as an afterthought.
Abstract
This paper presents an explainable AI (XAI) system that provides explanations for its predictions. The system consists of two key components -- namely, the prediction And-Or graph (AOG) model for recognizing and localizing concepts of interest in input data, and the XAI model for providing explanations to the user about the AOG's predictions. In this work, we focus on the XAI model specified to interact with the user in natural language, whereas the AOG's predictions are considered given and represented by the corresponding parse graphs (pg's) of the AOG. Our XAI model takes pg's as input and provides answers to the user's questions using the following types of reasoning: direct evidence (e.g., detection scores), part-based inference (e.g., detected parts provide evidence for the concept asked), and other evidences from spatio-temporal context (e.g., constraints from the spatio-temporal surround). We identify several correlations between user's questions and the XAI answers using Youtube Action dataset.