Investigating and improving near-term quantum algorithms for machine learning and optimization

The QUTAC working group Quantum Machine Learning and Optimization is investigating various approaches on how to leverage quantum computing for industrial use cases in these fields. We present extensions of the well-known Quantum Approximate Optimization Algorithm (QAOA) to reduce resource requirements on NISQ devices and benchmark their performance on the knapsack problem, which describes various use cases such as supply chain optimization or allocation of production tasks. Another possibility to tackle such optimization problems with near-term quantum devices is to employ generative neural networks. We show results from a first study using quantum-inspired techniques such as tensor networks, which can be extended to quantum algorithms in the future. Moreover, ideas to improve full quantum generative adversarial networks (qGAN) are discussed.