In hardware implementation of reservoir computing, network structure is tied to individual setups, and therefore their prediction performances are limited depending on the hardware configuration. We develop a tunable reservoir computing system based on iterative-function system. Optical affine parameters in the iterative-function systems modulate reservoir properties and improve the prediction performance.

Cylindrical vector beams have an axially symmetric polarization distribution, which can present rich topological structures on the beam cross-section. Due to their unique polarization characteristics, cylindrical vector beams have been applied to various fields including optical trapping, super-resolution microscopy, and laser processing. In this paper, we review the basic properties of cylindrical vector beams and their applications in those areas.

Solvate ionic liquids (SILs) composed of glymes and Li salts are potential electrolyte materials for Li ion batteries. Competing characters of solvates are favorable for battery applications: stable solvate in the bulk leads to non-volatility, and its rapid desolvation near an electrode is necessary for high performance. For mechanistic understanding of such characters, potential-dependent interfacial structure of [Li(G4)][TFSA] on an Au electrode was investigated by using surface-enhanced infrared absorption spectroscopy (SEIRAS). We observed the partial desolvation of Li⁺ on the electrode at the potential much more positive than Li⁺ reduction onset and ascribed to the key of such competing characters.

The heart of topological photonics lies at topological invariants, which are deeply related to singular points. In this report, we take up three examples in topological photonics—photonic topological insulators, polarization singular states, and exceptional points in non-Hermitian photonic crystals, and show the definition of the topological invariant (charge), and its topological properties. Owing to the conservation of the topological invariant, one can control various optical properties of Hermitian and non-Hermitian photonic crystals simply by tuning their crystal symmetry. Such intriguing concepts and unique tunability may pave the way to novel photonic devices and processing.

Analog photonic processing is one of the attractive computation engine for machine learning. Here, I review recent progress of this field. I also introduce our work on scaling up analog photonic platforms including large-scale WDM-based matrix-vector processor, on-chip photonic linear processor and its application to reservoir computer, and its training by the hardware-oriented training. Our approach scales up the photonic analog processing towards the fundamental Nyquist limit.

This article reviews the fundamental understanding of the electromagnetic coupling between plasmons and molecular excitons. It introduces the concept of a plasmon, specifically a localized surface plasmon with a single dipole, as an optical resonator, i.e., a plasmonic resonator. This article then explores the coupling between the plasmonic resonator and the molecular excitons located within the plasmonic resonator. Furthermore, the concept of strong coupling is introduced as an illustrative example of this electromagnetic coupling.

Topological photonics brings the concept of mathematical topology into the field of optics, which allows us to systematically handle information derived from the topology of light, such as optical spin and orbital angular momentum. In this paper, we focus on topological edge states, the most important phenomenon in the field of topological photonics, and describe the specific behavior of an optical spin caused by the states.