17544314. TRAINING OF QUANTUM BOLTZMANN MACHINES BY QUANTUM IMAGINARY-TIME EVOLUTION simplified abstract (INTERNATIONAL BUSINESS MACHINES CORPORATION)

From WikiPatents
Jump to navigation Jump to search

TRAINING OF QUANTUM BOLTZMANN MACHINES BY QUANTUM IMAGINARY-TIME EVOLUTION

Organization Name

INTERNATIONAL BUSINESS MACHINES CORPORATION

Inventor(s)

Antonio Mezzacapo of Westchester NY (US)

Mario Motta of San Jose CA (US)

TRAINING OF QUANTUM BOLTZMANN MACHINES BY QUANTUM IMAGINARY-TIME EVOLUTION - A simplified explanation of the abstract

This abstract first appeared for US patent application 17544314 titled 'TRAINING OF QUANTUM BOLTZMANN MACHINES BY QUANTUM IMAGINARY-TIME EVOLUTION

Simplified Explanation

The patent application describes a system and methods for training quantum Boltzmann machines using quantum imaginary-time evolution. This involves evaluating a Kullback-Leibler divergence gradient through a sampling procedure using quantum imaginary-time evolution and a Hadamard circuit.

  • The system includes computer executable components stored in memory.
  • The evaluation component calculates the Kullback-Leibler divergence gradient.
  • The sampling procedure generates samples using quantum imaginary-time evolution and a Hadamard circuit.

Potential Applications

  • Quantum computing
  • Machine learning
  • Artificial intelligence research

Problems Solved

  • Training quantum Boltzmann machines efficiently
  • Improving the accuracy of Kullback-Leibler divergence gradient evaluation

Benefits

  • Faster and more accurate training of quantum Boltzmann machines
  • Enhanced performance in machine learning and artificial intelligence tasks


Original Abstract Submitted

Systems, computer-implemented methods, and computer program products to facilitate training of quantum Boltzmann machines by quantum imaginary-time evolution. According to an embodiment, a system can comprise computer executable components stored in memory. The computer executable components comprise an evaluation component that evaluates a Kullback-Leibler divergence gradient by a sampling procedure, where samples are generated by quantum imaginary-time evolution and a Hadamard circuit.

Retrieved from "http://wikipatents.org/index.php?title=17544314._TRAINING_OF_QUANTUM_BOLTZMANN_MACHINES_BY_QUANTUM_IMAGINARY-TIME_EVOLUTION_simplified_abstract_(INTERNATIONAL_BUSINESS_MACHINES_CORPORATION)&oldid=25748"