The promise of Quantum AI
Over the last years many promising quantum-algorithms have been proposed for a wide range of typical machine learning (ML) and artificial intelligence (AI) tasks including clustering, dimension reduction, continuous estimation, classification, anomaly detection, recommendation systems, data generation and general optimization problems. The behavior of quantum-computing devices is best described by linear algebra in very high-dimensional vector spaces, which is a promising starting-point for pursuing enhancement of classical ML/AI. Also, quantum processing units (QPUs) can generate quantumstates with probability distributions that are hard to sample from classically. This feature is believed to allow QPUs to recognize data-patterns that are hard to recognize classically. Many other classically intractable ML/AI problems could perhaps benefit from quantum-enhancements, including e.g. belief propagation, tensor factorization or submodular problems. Within the Quantum AI domain, Qu & Co's main focus is on developing quantum-enhanced AI/ML: hybrid methods combining QPU with CPU/GPU solutions.