Development of quantum computation based artificial intelligence algorithms

Quantum computing is a computational paradigm based on quantum mechanics, which has proven effective today and has the potential to solve very difficult problems compared to classical computing. Artificial intelligence, on the other hand, aims to develop systems with human capabilities. One of the barriers to improving artificial intelligence systems is the high computing power they require. In the thesis study, the use of the advantages of quantum computing in artificial intelligence methods was investigated. The main scientific contributions obtained as a result of the research are the improvement of quantum-inspired optimization algorithms, quantum data encoding optimization, parallel processing of images in a quantum environment, and the development of quantum convolution/pooling methods for deep learning. It has been observed that the proposed approach for quantum-inspired genetic algorithms achieves a better solution by about 12% compared to the original method. The proposed quantum data coding method for the optimized encoding of large data such as images has been tested on a 4x4 sample image with 4-bit resolution. It has been observed that the proposed method produces approximately 57% better results in terms of cost115 compared to the ESOP method. Value-based quantum convolution and pooling methods developed for deep learning applications can be run on the whole image without the need for classical computers with the help of the proposed parallel image processing framework. It has been observed that the method produces exactly the same results as the classical convolution operation. Then, the variational quantum convolution method is proposed for modeling the classical convolution process using analog coded inputs. The parameters of the created Ansatz circuit were optimized with the differential evolution algorithm, and the classical convolution process was successfully modeled within 0.001 mean squared error. As a result, optimized method suggestions have been made for the use of quantum computing in various artificial intelligence algorithms. It has been seen that the advantages of quantum computers can be used in artificial intelligence methods without the need for classical computers.