Other links:

Other links:

Event Calender

Loading Events

Physics Department: Paper presentations for the subject breadth evaluation| Souradeep Sengupta

Can Your Computer \"Learn\" Quantum Mechanics?

  • This event has passed.





Abstract: This talk will review applications of machine learning to quantum many-body physics, primarily centering around Carleo et al., Science 355, 602–606 (2017), “Solving the quantum many-body problem with artificial neural networks”, with additional references as and when required. I shall first briefly discuss the computational challenges of quantum many-body physics, as well as 20th century techniques for dealing with the curse of dimensionality in these problems, such as quantum Monte Carlo and tensor networks. I will spend a little time discussing how the scaling behaviour of entanglement entropy of the quantum system determines the efficiency and accuracy of the numerical approximation. Then I shall discuss what machine learning is, and how it brings in a new paradigm for attacking this problem. I shall discuss ideas like reinforcement learning, artificial neural networks and Restricted Boltzmann Machines in a way accessible to people with a physics background, as well as how neural networks help us deal with massive entanglement. I shall finally work through in detail the variational Neural Quantum State (NQS) framework Carleo and Troyer used for efficiently and accurately calculating the ground state and describing the unitary time evolution of two model quantum systems – the transverse field Ising model and the antiferromagnetic Heisenberg model. 


Study at Ashoka

Study at Ashoka