Large-scale photonic computing with nonlinear disordered media.

Neural networks find widespread use in scientific and technological applications, yet their implementations in conventional computers have encountered bottlenecks due to ever-expanding computational needs. Photonic computing is a promising neuromorphic platform with potential advantages of massive parallelism, ultralow latency and reduced energy consumption but mostly for computing linear operations. Here we demonstrate a large-scale, high-performance nonlinear photonic neural system based on a disordered polycrystalline slab composed of lithium niobate nanocrystals. Mediated by random quasi-phase-matching and multiple scattering, linear and nonlinear optical speckle features are generated as the interplay between the simultaneous linear random scattering and the second-harmonic generation, defining a complex neural network in which the second-order nonlinearity acts as internal nonlinear activation functions. Benchmarked against linear random projection, such nonlinear mapping embedded with rich physical computational operations shows improved performance across a large collection of machine learning tasks in image classification, regression and graph classification. Demonstrating up to 27,648 input and 3,500 nonlinear output nodes, the combination of optical nonlinearity and random scattering serves as a scalable computing engine for diverse applications.

Nanoimprinting Solution-Derived Barium Titanate for Electro-Optic Metasurfaces.

Electro-optic metasurfaces have demonstrated significant potential in enhancing the modulation speed and efficiency for fast and large-scale free-space optical devices. Barium titanate has a strong electro-optic Pockels coefficient, but its availability in thin-film form is restricted due to costly growth processes or low thickness. Here, we fabricated active metasurfaces using an etch-free bottom-up process with sol-gel-based polycrystalline barium titanate with a large electro-optic coefficient similar to bulk lithium niobate. We achieve strong hybrid Mie/surface lattice resonances with a quality-factor of 200 at 633 nm wavelength, enhancing the light-matter interaction and therefore the Pockels effect. The metasurface transmission is electro-optically modulated with up to 5 MHz driving frequency at low voltages of less than 1 V thanks to resonant enhancement of the modulation amplitude by 2 orders of magnitude. This successful demonstration of electro-optic modulation in nanoimprinted barium titanate structures paves the way for low-cost and large-scale free-space modulators or tunable metalenses.

Encoding information in the mutual coherence of spatially separated light beams.

Coherence has been used as a resource for optical communications since its earliest days. It is widely used for the multiplexing of data, but not for the encoding of data. Here we introduce a coding scheme, which we call mutual coherence coding, to encode information in the mutual coherence of spatially separated light beams. We describe its implementation and analyze its performance by deriving the relevant figures of merit (signal-to-noise ratio, maximum bit-rate, and spectral efficiency) with respect to the number of transmitted beams. Mutual coherence coding yields a quadratic scaling of the number of transmitted signals with the number of employed light beams, which might have benefits for cryptography and data security.