Photonic platforms for universal quantum computing, quantum simulation, and neuromorphic computing

Quantum photonics offers a variety of ways to realize scalable computational architectures, be they for universal, fault tolerant quantum computing, quantum simulation, or neuromorphic computing. In this talk, I will review our work, at the University of Virginia, on all these topics as well as on machine-learning-enabled photonic quantum state generation. Particular emphasis will be placed on field-based, a.k.a. qumode, implementations which leverage the natural spectral, temporal, and spatial scalability of quantum optics while still allowing fault tolerant qubit encodings such as the Gottesman-Kitaev-Preskill one. The formalism of measurement-based quantum computing, e.g. based on cluster states, can be deployed in a very compact manner using the spectrally multiplexed entanglement pioneered by my group using massively multimode squeezing. We have also shown that the adjunction of photon-number-resolving measurements as the sole non-Gaussian (Wigner function) resource to the "cheaper" Gaussian squeezing and interferometric operations is sufficient to reach quantum computing universality and fault tolerance. Finally, I will pivot to neuromorphic computing where none of this is actually required... and present a new instance of linear photonics able to embody neuromorphic computing, in line with results by other groups in 2024. All these results are in principle transferable to integrated photonics on chip, which bodes well for the relevance of photonic machines for a variety of applications.