Miles Stoudenmire: Introduction to Tensor Networks for Machine Learning.
Miles Stoudenmire (Flatiron Institute)
Talk given at CMAC2020
https://research.kent.ac.uk/pqm/condensed-matter-physics-in-the-cities-2020/schedule/
Abstract:
Tensor networks are a tool to store and compute with tensors having many indices, which would otherwise be exponentially costly. While in physics the prototypical use of tensor networks is for approximating many-body wavefunctions, it has been appreciated over the last decade that tensor networks such as matrix product states (MPS) are much more general tools. One emerging application of tensor networks is for machine learning, where they can be used for a variety of tasks such as supervised learning, generative modeling, and compression of neural network parameters. Compared to other machine learning approaches, tensor networks offer advantages such as scalability, more sophisticated training algorithms, and theoretical clarity. After reviewing the basics of tensor networks, I will introduce the setting of machine learning and explain how tensor networks can be used for machine learning tasks, showing exemplary results for both classical (non-physics) applications and physics applications.
Видео Miles Stoudenmire: Introduction to Tensor Networks for Machine Learning. канала Condensed Matter Physics In the City
Talk given at CMAC2020
https://research.kent.ac.uk/pqm/condensed-matter-physics-in-the-cities-2020/schedule/
Abstract:
Tensor networks are a tool to store and compute with tensors having many indices, which would otherwise be exponentially costly. While in physics the prototypical use of tensor networks is for approximating many-body wavefunctions, it has been appreciated over the last decade that tensor networks such as matrix product states (MPS) are much more general tools. One emerging application of tensor networks is for machine learning, where they can be used for a variety of tasks such as supervised learning, generative modeling, and compression of neural network parameters. Compared to other machine learning approaches, tensor networks offer advantages such as scalability, more sophisticated training algorithms, and theoretical clarity. After reviewing the basics of tensor networks, I will introduce the setting of machine learning and explain how tensor networks can be used for machine learning tasks, showing exemplary results for both classical (non-physics) applications and physics applications.
Видео Miles Stoudenmire: Introduction to Tensor Networks for Machine Learning. канала Condensed Matter Physics In the City
Показать
Комментарии отсутствуют
Информация о видео
18 июня 2020 г. 6:47:36
01:14:01
Другие видео канала
Far-from-equilibrium dynamics of systems with conservation lawsAshley Cook 2 “Topological skyrmion phases of matter II "Pascal Simon “New insights from electronic transport in superconducting bound-states”Elio Koenig “Topological order and quantum materials”Topological responses from gapped Weyl points in 2D altermagnetsEwan Scott (UCL) "Hard axis ordering in Kondo magnets "Masaki Oshikawa (ISSP Tokyo) Lecture 1.Andrei Bernevig “Mapping TBG into a Topological Heavy Fermion”New Paths for Topological Flat Bands & Roton Condensate in Fractional Quantum Anomalous Hall SystemsMiles Stoudenmire (Flatiron Institute)Je-Geun Park (Seoul National University)“Emerging new opportunities of van der Waals magnets”Tony Carrington (Bristol University) “Magneto-transport in cuprate superconductors”Synthesis as the heart of New Materials PhysicsRoser Valenti (Frankfurt):Steve Simon (Oxford U)“Kekule spirals in twisted bilayer graphene”Non-equilibrium THz response of time-reversal symmetry-breaking superconductorsAmalia Coldea “Fermi surfaces and quasiparticle behaviour of iron chalcogenide superconductors”Supersolidity in the ultracold: When atoms behave as crystal and superfluid at the same timeQuantum critical line encloses intensely magnetic field–resilient superconductivity in UTe2Masaki Oshikawa (ISSP, Tokyo)