Title: Generating Approximate Ground State of Molecules Using Quantum Machine Learning<\/strong><\/h4>\n\n\n\nI will be discussing our paper, which has the same title as the title of the talk, on training hybrid classical-quantum neural networks to generate ground states of molecules on quantum devices. I will discuss the architecture used, rigorous sample bounds on training the model, lower bounds derived on the amount\/type of training data needed to train models of a similar form, and extensive numerics, where we show the effectiveness of our approach.<\/p>\n\n\n\n
<\/figure>\n\n\n\nAkhil Francis<\/strong><\/h3>\n\n\n\nGraduate Student, NC State University<\/p>\n\n\n\n
Bio<\/h4>\n\n\n\n
Akhil Francis is a physics Ph.D. student at North Carolina State University (NCSU) working under the supervision of Dr. Alexander Kemper. His research is primarily in using quantum computation to solve condensed matter physics problems, especially using noisy intermediate-scale quantum (NISQ) machines. He is also interested in benchmarking quantum machines as well as in employing machine learning tools to explore quantum computing data.<\/p>\n\n\n\n
Title: Subspace Diagonalization on Quantum Computers using Eigenvector Continuation<\/strong><\/h4>\n\n\n\nQuantum subspace diagonalization (QSD) methods are quantum-classical hybrid methods, commonly used to find ground and excited state energies by projecting the Hamiltonian to a smaller subspace. In applying these, the choice of subspace basis is critical from the perspectives of basis completeness and efficiency of implementation on quantum computers. In this work, we present Eigenvector Continuation (EC) as a QSD method, where low-energy states of the Hamiltonian at different points in parameter space are chosen as the subspace basis. This unique choice enables rapid evaluation of low-energy spectra, including ground and nearby excited states, with minimal hardware effort. As a particular advantage, EC is able to capture the spectrum across ground state crossovers corresponding to different symmetry sectors of the problem. We demonstrate this method for interacting spin models and molecules.<\/p>\n\n\n\n
<\/figure>\n\n\n\nDominik Hangleiter<\/strong><\/h3>\n\n\n\nHartree Postdoctoral Fellow, QuICS (University of Maryland & NIST)<\/p>\n\n\n\n
Bio<\/h4>\n\n\n\n
Dominik is a quantum scientist working at the interface between computer science, mathematics and physics, as well as the philosophy of quantum science. In his research, he explores the potential of analog and digital quantum computing devices from the theoretical perspective of computational complexity and classical simulation algorithms, and tries to understand the physical mechanisms underlying their speedups. In collaboration with experimental groups, he develops and applies resource-efficient characterization tools to help improve those devices. He also enjoys thinking about methodological aspects of science in general, and quantum science in particular from a philosophy of science point of view. Dominik studied physics, philosophy, and mathematics in Konstanz, Oxford and Munich, and received his Ph.D. from Freie Universit\u00e4t Berlin.<\/p>\n\n\n\n
Title: Is there an advantage of quantum learning?<\/strong><\/h4>\n\n\n\nQuantum machine learning has been hailed as one of the promising near-term applications of small quantum computers and much research is focused on devising quantum heuristics that might yield an advantage over classical learning algorithms. In this talk, I want to take a step back and ask: To what extent can we hope for a quantum advantage in machine learning? I will begin by introducing different models of learning. I will then focus on the following task: Given samples from some unknown discrete probability distribution, output an efficient algorithm for generating new samples from that distribution. I will first present a result showing that for very specific distributions the answer is a resounding \u201cYes!\u201d. Then, I will present some results on the learnability of the output distributions of quantum circuits\u2014a natural general setting in which one might hope for a quantum advantage.<\/p>\n\n\n\n
<\/figure>\n\n\n\n