Kaoru MINOSHIMA (美濃島 薫)

Abstract

Optical reservoir computing (ORC) promises fast, energy-efficient temporal inference by harnessing the rich transient dynamics of photonic systems. Yet most ORC demonstrations still depend on fiber delay lines or camera-based spatial multiplexing, which caps the clock rate at a few tens of MSa/s and complicates monolithic integration. Here we introduce a frequency-multiplexed ORC whose nodes are the individual modes of a dissipative Kerr-soliton microcomb generated in a high-Q Si₃N₄ microresonator. The input signal is encoded as a rapid detuning modulation of the pump laser, so the intracavity dynamics of the microcomb provide both the high-dimensional nonlinear mapping and tens of nanoseconds of memory, while output weighting is realized optically with standard microring arrays. Numerical modeling with 60 comb modes provides a normalized mean-square error (NMSE) of 0.015 on the Santa Fe chaotic time-series task at 50 MSa/s and more than a tenfold reduction in symbol-error rate for nonlinear equalization (NLEQ) at 100 MSa/s. A proof-of-concept experiment using 37 measured modes also confirms the concept on the Santa Fe chaotic time-series and NLEQ benchmarks. Because both the microcomb and weighting network are fabricated by a complementary metal-oxide semiconductor (CMOS)-compatible process, the architecture offers a clear path toward compact, energy-efficient photonic processors operating at greater than 1 GSa/s, directly addressing the scalability and speed challenges of nanophotonic ORC.

Photonic extreme learning machines (ELMs) offer rapid training, low power consumption, and inherent parallelism compared to conventional electronic systems. Time-domain photonic ELMs employing Rayleigh backscattering in an optical fiber for high-dimensional mapping enable simplified architectures and ultrafast processing. In this work, we introduce a novel phase-sensitive optical-time-domain reflectometry method that encodes input data in pulse phase and employs heterodyne detection to eliminate the need for optical amplification of the Rayleigh backscattered signal. In the proof-of-concept experiments, we showcase competitive classification accuracies of 94.93% and 91.56% on the Iris and Vowel datasets, respectively, highlighting the effectiveness and practical advantages of our approach.

Quantum correlations of ultrafast quantum optical states in the time–frequency domain manifest prominently on femtosecond time scales, an elusive challenging regime. We develop a new method to probe the time–frequency mode structure of ultrafast two-photon states utilizing dual-comb-based asynchronously overlapping pulses with single-photon frequency upconversion through femtosecond optical sampling. This method enables quick, highly precise joint temporal intensity measurements over a broad temporal dynamic range with only a single linear scanning stage. We demonstrate full characterization of time–frequency entangled two-photon states in both time and frequency domains. We quantified the time–frequency mode structure in both time and frequency domains through Schmidt decomposition, resulting in temporal and spectral purities of 87.7% and 76.9%, respectively, for heralded single-photons without any spectral filtering. Furthermore, through the simple control of pump spectral bandwidth, we demonstrate the ability to tailor time–frequency entanglement, verifying the two-dimensional Fourier duality. This method paves the way for a substantial upgrade to ultrafast quantum optical techniques, enabling a rapid, detailed, and multi-dimensional characterization of entangled states with potential applications in quantum metrology and photon-deficient fluorescence imaging.