Harnessing Disordered-Ensemble Quantum Dynamics for Machine Learning

The authors describe an alternative to digital quantum computation that uses natural quantum dynamics for information processing. $Q\phantom{\rule{0}{0ex}}u\phantom{\rule{0}{0ex}}a\phantom{\rule{0}{0ex}}n\phantom{\rule{0}{0ex}}t\phantom{\rule{0}{0ex}}u\phantom{\rule{0}{0ex}}m$ $r\phantom{\rule{0}{0ex}}e\phantom{\rule{0}{0ex}}s\phantom{\rule{0}{0ex}}e\phantom{\rule{0}{0ex}}r\phantom{\rule{0}{0ex}}v\phantom{\rule{0}{0ex}}o\phantom{\rule{0}{0ex}}i\phantom{\rule{0}{0ex}}r$ $c\phantom{\rule{0}{0ex}}o\phantom{\rule{0}{0ex}}m\phantom{\rule{0}{0ex}}p\phantom{\rule{0}{0ex}}u\phantom{\rule{0}{0ex}}t\phantom{\rule{0}{0ex}}i\phantom{\rule{0}{0ex}}n\phantom{\rule{0}{0ex}}g$ does not require fine tuning of parameters, is robust against noise, and is based on existing devices. Simulations suggest that with this approach, a system of just 5 to 7 qubits is as powerful as a recurrent neural network with hundreds of nodes. This framework for artificial intelligence powered by quantum physics enables $t\phantom{\rule{0}{0ex}}e\phantom{\rule{0}{0ex}}m\phantom{\rule{0}{0ex}}p\phantom{\rule{0}{0ex}}o\phantom{\rule{0}{0ex}}r\phantom{\rule{0}{0ex}}a\phantom{\rule{0}{0ex}}l$ machine-learning tasks, such as natural language processing and predicting the stock market.