Prof. Keng Hock, Mark Goh

School of Business, National University of Singapore, Singapore

Research Area: Supply Chain Management, Purchasing, Quality And Healthcare, Service Operations; Scorecarding And Supplier Performance Management Systems; Reverse Logistics; Scm

Prof. Weishan Zhang

China University of Petroleum (East China), China

Research Area: Big data intelligent processing, artificial intelligence, autonomous computing

Title: Software Architecture for Responsible Artificial Intelligence Systems: Thoughts and Practices

Abstract: Existing work on responsible AI ignores system-level architecture design for implementing responsible AI, and there lacks solutions on software engineering for responsible AI. Therefore, we propose a comprehensive approach for Responsible AI based Software Architecture, using an example of Digitalization of Industry Drawings (RAISA-DID), as a reference on software engineering for responsible AI to other industry domains. We analyze requirements of RAISA-DID and deign it using a quality-driven approach. A comprehensive and effective AI algorithms stack is proposed that can recognize all elements thoroughly in industry drawings, supported by federated learning, computer vision, natural language processing, and other AI algorithms. Finally, we use piping and instrumentation diagram recognition as an application case to evaluate the proposed approach. Comprehensive evaluations on different software qualities demonstrate that the solutions we proposed are efficient and effective for practical industry usage.