Emerging technologies in artificial intelligence: quantum rule-based systems

Abstract

This article tries to establish synergies between two areas of research and development that are apparently disconnected: artificial intelligence (AI) and quantum computing (QC). The article begins with a brief description of the origins of AI and the type of problems that AI tries to solve. Next, we focus on a specific type of AI program, the so-called rule-based system (RBS). In this regard, one of the biggest problems with RBS is the pattern-matching that is related to the process of rule selection. This makes RBS very sensitive to the size of the problem and to changing environments, which is at the core of a very significant increase in computational cost. In this context we wonder if the use of QC techniques could help solve the problem. We have already presented ideas and concepts in the quantum world that could contribute to improving the performance of RBS. We place special emphasis on the principle of coherent superposition, on the intrinsically parallel character of QC, on the probabilistic nature of QC, and we describe some quantum operators that, of course, must be reversible. Next, we build a quantum architecture equivalent to a conventional inferential circuit and analyze the results. The article continues with the construction of a quantum rule-based system that solves a hypothetical problem in the field of medicine. Finally, the proposed approach is discussed, and preliminary conclusions are established.