Search

Kolmogorov-Arnold Networks (KAN) models were recently proposed and claimed to provide improved parameter scaling and interpretability compared to conventional multilayer perceptron (MLP) models. Inspired by the KAN architecture, we propose the Photonic KAN -- an integrated all-optical neuromorphic platform leveraging highly parametric optical nonlinear transfer functions along KAN edges. In this work, we implement such nonlinearities in the form of cascaded ring-assisted Mach-Zehnder Interferometer (MZI) devices. This innovative design has the potential to address key limitations of current photonic neural networks. In our test cases, the Photonic KAN showcases enhanced parameter scaling and interpretability compared to existing photonic neural networks. The photonic KAN achieves approximately 65$\times$ reduction in energy consumption and area, alongside a 50$\times$ reduction in latency compared to previous MZI-based photonic accelerators with similar performance for function fitting task. This breakthrough presents a promising new avenue for expanding the scalability and efficiency of neuromorphic hardware platforms., Comment: 11 pages, 7 figures, 1 table

Operator learning has become a powerful tool in machine learning for modeling complex physical systems governed by partial differential equations (PDEs). Although Deep Operator Networks (DeepONet) show promise, they require extensive data acquisition. Physics-informed DeepONets (PI-DeepONet) mitigate data scarcity but suffer from inefficient training processes. We introduce Separable Operator Networks (SepONet), a novel framework that significantly enhances the efficiency of physics-informed operator learning. SepONet uses independent trunk networks to learn basis functions separately for different coordinate axes, enabling faster and more memory-efficient training via forward-mode automatic differentiation. We provide a universal approximation theorem for SepONet proving that it generalizes to arbitrary operator learning problems, and then validate its performance through comprehensive benchmarking against PI-DeepONet. Our results demonstrate SepONet's superior performance across various nonlinear and inseparable PDEs, with SepONet's advantages increasing with problem complexity, dimension, and scale. For 1D time-dependent PDEs, SepONet achieves up to $112\times$ faster training and $82\times$ reduction in GPU memory usage compared to PI-DeepONet, while maintaining comparable accuracy. For the 2D time-dependent nonlinear diffusion equation, SepONet efficiently handles the complexity, achieving a 6.44\% mean relative $\ell_{2}$ test error, while PI-DeepONet fails due to memory constraints. This work paves the way for extreme-scale learning of continuous mappings between infinite-dimensional function spaces. Open source code is available at \url{https://github.com/HewlettPackard/separable-operator-networks}., Comment: SepONet version 2. This revised version polishes writing and open sources code. The initial version was submitted to arXiv on July 15, 2024

Implementing on-chip non-volatile optical memories has long been an actively pursued goal, promising significant enhancements in the capability and energy efficiency of photonic integrated circuits. Here, a novel optical memory has been demonstrated exclusively using the semiconductor primary material, silicon. By manipulating the optoelectronic effect of this device, we introduce a hysteresis effect at the silicon-silicon oxide interface, which in turn demonstrates multi-level, non-volatile optical data storage with robust retention and endurance. This new silicon optical memory provides a distinctively simple and accessible route to realize optical data storage in standard silicon foundry processes.

Shape optimization approaches to inverse design offer low-dimensional, physically-guided parameterizations of structures by representing them as combinations of shape primitives. However, on discretized rectilinear simulation grids, computing the gradient of a user objective via the adjoint variables method requires a sum reduction of the forward/adjoint field solutions and the Jacobian of the simulation material distribution with respect to the structural shape parameters. These shape parameters often perturb large or global parts of the simulation grid resulting in many non-zero Jacobian entries, which are typically computed by finite-difference in practice. Consequently, the gradient calculation can be non-trivial. In this work we propose to accelerate the gradient calculation by invoking automatic differentiation (AutoDiff) in instantiations of structural material distributions. In doing so, we develop extensible differentiable mappings from shape parameters to shape primitives and differentiable effective logic operations (denoted AutoDiffGeo). These AutoDiffGeo definitions may introduce some additional discretization error into the field solutions because they relax notions of sub-pixel smoothing along shape boundaries. However, we show that some mappings (e.g. simple cuboids) can achieve zero error with respect to volumetric averaging strategies. We demonstrate AutoDiff enhanced shape optimization using three integrated photonic examples: a multi-etch blazed grating coupler, a non-adiabatic waveguide transition taper, and a polarization-splitting grating coupler. We find accelerations of the gradient calculation by AutoDiff relative to finite-difference often exceed 50x, resulting in total wall time accelerations of 4x or more on the same hardware with little or no compromise to final device performance. Our code is available open source at https://github.com/smhooten/emopt, Comment: 29 pages, 15 figures

The 2008 Corona effort was inspired by a pressing need for more of everything, as demanded by the salient problems of the day. Dennard scaling was no longer in effect. A lot of computer architecture research was in the doldrums. Papers often showed incremental subsystem performance improvements, but at incommensurate cost and complexity. The many-core era was moving rapidly, and the approach with many simpler cores was at odds with the better and more complex subsystem publications of the day. Core counts were doubling every 18 months, while per-pin bandwidth was expected to double, at best, over the next decade. Memory bandwidth and capacity had to increase to keep pace with ever more powerful multi-core processors. With increasing core counts per die, inter-core communication bandwidth and latency became more important. At the same time, the area and power of electrical networks-on-chip were increasingly problematic: To be reliably received, any signal that traverses a wire spanning a full reticle-sized die would need significant equalization, re-timing, and multiple clock cycles. This additional time, area, and power was the crux of the concern, and things looked to get worse in the future. Silicon nanophotonics was of particular interest and seemed to be improving rapidly. This led us to consider taking advantage of 3D packaging, where one die in the 3D stack would be a photonic network layer. Our focus was on a system that could be built about a decade out. Thus, we tried to predict how the technologies and the system performance requirements would converge in about 2018. Corona was the result this exercise; now, 15 years later, it's interesting to look back at the effort., Comment: 2 pages. Proceedings of ISCA-50: 50 years of the International Symposia on Computer Architecture (selected papers) June 17-21 Orlando, Florida

We present a novel design of optical micro-cavity where the optical energy resides primarily in free space, therefore is readily accessible to foreign objects such as atoms, molecules, mechanical resonators, etc. We describe the physics of these resonators, and propose a design method based on stochastic optimization. Cavity designs with diffraction-limited mode volumes and quality factors in the range of $10^4$--$10^6$ are presented. With a purely planar geometry, the cavity can be easily integrated on-chip using conventional micro- and nano- fabrication processes.

Sub-wavelength dielectric gratings (SWG) have emerged recently as a promising alternative to distributed-Bragg-reflection (DBR) dielectric stacks for broadband, high-reflectivity filtering applications. A SWG structure composed of a single dielectric layer with the appropriate patterning can sometimes perform as well as thirty or forty dielectric DBR layers, while providing new functionalities such as polarization control and near-field amplification. In this paper, we introduce a remarkable property of grating mirrors that cannot be realized by their DBR counterpart: we show that a non-periodic patterning of the grating surface can give full control over the phase front of reflected light while maintaining a high reflectivity. This new feature of dielectric gratings could have a substantial impact on a number of applications that depend on low-cost, compact optical components, from laser cavities to CD/DVD read/write heads., Comment: submitted to Nature Photonics

Coherent population trapping at zero magnetic field was observed for nitrogen-vacancy centers in diamond under optical excitation. This was measured as a reduction in photoluminescence when the detuning between two excitation lasers matched the 2.88 GHz crystal-field splitting of the color center ground states. This behavior is highly sensitive to strain, which modifies the excited states, and was unexpected following recent experiments demonstrating optical readout of single nitrogen-vacancy electron spins based on cycling transitions. These results demonstrate for the first time that three-level Lambda configurations suitable for proposed quantum information applications can be realized simultaneously for all four orientations of nitrogen-vacancy centers at zero magnetic field., Comment: 8 pages, 5 figures, new figure added

We demonstrate a simple, robust, and ultrabright parametric down-conversion source of polarization-entangled photons based on a polarization Sagnac interferometer. Bidirectional pumping in type-II phase-matched periodically poled KTiOPO4 yields a measured flux of 5 000 polarization-entangled pairs/s per mW of pump power in a 1-nm bandwidth at 96.8% quantum-interference visibility. The common-path arrangement of the Sagnac interferometer eliminates the need for phase stabilization for the biphoton output state., Comment: Submitted to PRA

We present a simple and low-cost implementation of a fast electronic coincidence detector based on PECL logic with a TTL-compatible interface. The detector has negligible dead time and the coincidence window is adjustable with a minimum width of 1 ns. Coincidence measurements of two independent sources of Bose-Einstein distributed photocounts are presented using different coincidence window widths., Comment: 4 pages, 6 figures

We demonstrate a new class of frequency-entangled states generated via spontaneous parametric down-conversion under extended phase matching conditions. Biphoton entanglement with coincident signal and idler frequencies is observed over a broad bandwidth in periodically poled KTiOPO$_4$. We demonstrate high visibility in Hong-Ou-Mandel interferometric measurements under pulsed pumping without spectral filtering, which indicates excellent frequency indistinguishability between the down-converted photons. The coincident-frequency entanglement source is useful for quantum information processing and quantum measurement applications., Comment: 4 pages, 3 figures, submitted to PRL

A SWAP operation between different types of qubits of single photons is essential for manipulating hyperentangled photons for a variety of applications. We have implemented an efficient SWAP gate for the momentum and polarization degrees of freedom of single photons. The SWAP gate was utilized in a single-photon two-qubit quantum logic circuit to deterministically transfer momentum entanglement between a pair of down-converted photons to polarization entanglement. The polarization entanglement thus obtained violates Bell's inequality by more than 150 standard deviations., Comment: Changes in the body of the paper, one reference added, typos corrected

We demonstrate a robust implementation of a deterministic linear-optical Controlled-NOT (CNOT) gate for single-photon two-qubit quantum logic. A polarization Sagnac interferometer with an embedded 45$^{\circ}$-oriented dove prism is used to enable the polarization control qubit to act on the momentum (spatial) target qubit of the same photon. The CNOT gate requires no active stabilization because the two spatial modes share a common path, and it is used to entangle the polarization and momentum qubits., Comment: 10 pages, 4 figures. Typos corrected, referee comments and corrections