Classical algorithm for simulating experimental Gaussian boson sampling
Talkat Near-Term Quantum Computers: Fault Tolerance + Benchmarking + Quantum Advantage + Quantum Algorithms · 2024-04-22..04-26·
summary
A new classical algorithm simulates the largest reported Gaussian boson sampling experiments by exploiting high photon loss (>50%), using a semidefinite program to decompose into hard and easy layers and minimize the hard layer’s photon number, then simulating that reduced circuit with tensor networks (MPS). The optimization finds only 3–7% of photons in the hard layer, allowing generation of 10 million samples in about an hour on hundreds of GPUs, and the classical sampler matches or beats the experiment on high‑order correlations up to fifth order.
AI-written summary · about
chapters (9)
- 0:00Introduction and motivation
- 1:25Gaussian states overview
- 7:27Gaussian boson sampling
- 10:10Impact of photon loss
- 16:28Current quantum advantage
- 19:13Proposed classical algorithm
- 30:21Comparison with thermal states
- 32:20Verification and benchmarks
- 38:33Conclusion and future work
related talks
01h 10mLearning from Quantum Experiments
01h 15mImproved Machine Learning Algorithm for Predicting Ground State Properties
01h 06mTutorial: Methods from Statistical Physics II
39mNetworking for Quantum: how simulations help us to design the future
01h 11mStatistical Physics and Computation in High Dimension
01h 10mAn Introduction to Hamiltonian Monte Carlo Method for Sampling
35mPhase Retrieval In High Dimensions: Statistical And Computational Phase Transitions
01h 00mPhysics at the Large Hadron Collider: A New Window on Matter, Spacetime and the Universe II