From Features Engineering to Scenarios Engineering for Trustworthy AI: I&I, C&C, and V&V
Xuan Li, Peijun Ye, Juanjuan Li, Zhongmin Liu, Longbing Cao, Fei–Yue Wang
Abstract
Artificial intelligence (AI)’s rapid development has produced a variety of state-of-the-art models and methods that rely on network architectures and features engineering. However, some AI approaches achieve high accurate results only at the expense of interpretability and reliability. These problems may easily lead to bad experiences, lower trust levels, and systematic or even catastrophic risks. This article introduces the theoretical framework of scenarios engineering for building trustworthy AI techniques. We propose six key dimensions, including intelligence and index, calibration and certification, and verification and validation to achieve more robust and trusting AI, and address issues for future research directions and applications along this direction.