SIGCSE '18: Proceedings of the 49th ACM Technical Symposium on Computer Science Education
Quantum Artificial Intelligence for Natural Language Processing Applications
February 1, 2018
Natural Language Processing and Semantic Web include several NP complete/hard problems that are intractable for classical computing machines. Even though distributed computing has provided remarkable advances (more precisely in dealing with big data), non-decomposable NP problems are still intractable in many real-world applications. And, from a quantum computing perspective, solving complex problems with universal quantum gates requires the developing of quantum algorithms. Considering commercializing quantum annealing machines by D-Wave, achieving global optimum for discrete optimization problems has been realized. In this study, a novel approach has been introduced to convert symbolic AI problems into quadratic unconstrained binary optimization (QUBO) form. More narrowly, this method represents the classification of text documents (fragments) as optimizing a QUBO function. After embedding the training corpus into a QUBO function, D-Wave quantum annealer is used to classify new observations by finding the minimum energy level of the system.