# Publications by Thomas Barrett

## Fully reconfigurable coherent optical vector–matrix multiplication

Optics Letters The Optical Society **45** (2020) 5752-5752

Optics is a promising platform in which to help realise the next generation of fast, parallel and energy-efficient computation. We demonstrate a reconfigurable free-space optical multiplier that is capable of over 3000 computations in parallel, using spatial light modulators with a pixel resolution of only 340x340. This enables vector-matrix multiplication and parallel vector-vector multiplication with vector size of up to 56. Our design is the first to simultaneously support optical implementation of reconfigurable, large-size and real-valued linear algebraic operations. Such an optical multiplier can serve as a building block of special-purpose optical processors such as optical neural networks and optical Ising machines.

## Exploratory combinatorial optimization with reinforcement learning

Proceedings of the AAAI Conference on Artificial Intelligence Association for the Advancement of Artificial Intelligence **34** (2020)

Many real-world problems can be reduced to combinatorial optimization on a graph, where the subset or ordering of vertices that maximize some objective function must be found. With such tasks often NP-hard and analytically intractable, reinforcement learning (RL) has shown promise as a framework with which efficient heuristic methods to tackle these problems can be learned. Previous works construct the solution subset incrementally, adding one element at a time, however, the irreversible nature of this approach prevents the agent from revising its earlier decisions, which may be necessary given the complexity of the optimization task. We instead propose that the agent should seek to continuously improve the solution by learning to explore at test time. Our approach of exploratory combinatorial optimization (ECO-DQN) is, in principle, applicable to any combinatorial problem that can be defined on a graph. Experimentally, we show our method to produce state-of-the-art RL performance on the Maximum Cut problem. Moreover, because ECO-DQN can start from any arbitrary configuration, it can be combined with other search methods to further improve performance, which we demonstrate using a simple random search.

## Pushing Purcell enhancement beyond its limits

NEW JOURNAL OF PHYSICS **22** (2020) ARTN 063013

## Multimode interferometry for entangling atoms in quantum networks

Quantum Science and Technology IOP Publishing (2019)

## Polarisation Oscillations in Birefringent Emitter-Cavity Systems

Physical Review Letters American Physical Society (2019)

We present the effects of resonator birefringence on the cavity-enhanced interfacing of quantum states of light and matter, including the first observation of single photons with a time-dependent polarisation state that evolves within their coherence time. A theoretical model is introduced and experimentally verified by the modified polarisation of temporally-long single photons emitted from a $^{87}$Rb atom coupled to a high-finesse optical cavity by a vacuum-stimulated Raman adiabatic passage (V-STIRAP) process. Further theoretical investigation shows how a change in cavity birefringence can both impact the atom-cavity coupling and engender starkly different polarisation behaviour in the emitted photons. With polarisation a key resource for encoding quantum states of light and modern micron-scale cavities particularly prone to birefringence, the consideration of these effects is vital to the faithful realisation of efficient and coherent emitter-photon interfaces for distributed quantum networking and communications.

## Nonlinear Zeeman effects in the cavity-enhanced emission of polarised photons

New Journal of Physics IOP Publishing **20** (2018) 073030

We theoretically and experimentally investigate nonlinear Zeeman (NLZ) effects within a polarised single-photon source that uses a single 87Rb atom strongly coupled to a high finesse optical cavity. The breakdown of the atomic hyperfine structure in the ${{\rm{D}}}_{2}$ transition manifold for intermediate strength magnetic fields is shown to result in asymmetric and, ultimately, inhibited operation of the polarised atom–photon interface. The coherence of the system is considered using Hong–Ou–Mandel interference of the emitted photons. This informs the next steps to be taken and the modelling of future implementations, based on feasible cavity designs operated in regimes minimising NLZ effects, is presented and shown to provide improved performance.

## Learning Group Structure and Disentangled Representations of Dynamical Environments

ArXiv (0)

Learning disentangled representations is a key step towards effectively discovering and modelling the underlying structure of environments. In the natural sciences, physics has found great success by describing the universe in terms of symmetry preserving transformations. Inspired by this formalism, we propose a framework, built upon the theory of group representation, for learning representations of a dynamical environment structured around the transformations that generate its evolution. Experimentally, we learn the structure of explicitly symmetric environments without supervision from observational data generated by sequential interactions. We further introduce an intuitive disentanglement regularisation to ensure the interpretability of the learnt representations. We show that our method enables accurate long-horizon predictions, and demonstrate a correlation between the quality of predictions and disentanglement in the latent space.

## Backpropagation through nonlinear units for all-optical training of neural networks

ArXiv (0)

Backpropagation through nonlinear neurons is an outstanding challenge to the field of optical neural networks and the major conceptual barrier to all-optical training schemes. Each neuron is required to exhibit a directionally dependent response to propagating optical signals, with the backwards response conditioned on the forward signal, which is highly non-trivial to implement optically. We propose a practical and surprisingly simple solution that uses saturable absorption to provide the network nonlinearity. We find that the backward propagating gradients required to train the network can be approximated in a pump-probe scheme that requires only passive optical elements. Simulations show that, with readily obtainable optical depths, our approach can achieve equivalent performance to state-of-the-art computational networks on image classification benchmarks, even in deep networks with multiple sequential gradient approximations. This scheme is compatible with leading optical neural network proposals and therefore provides a feasible path towards end-to-end optical training.