In recent decades, the demand for computational power has surged, particularly with the rapid expansion of artificial intelligence (AI). As we navigate the post-Moore's law era, the limitations of traditional electrical digital computing, including process bottlenecks and power consumption issues, are propelling the search for alternative computing paradigms. Among various emerging technologies, integrated photonics stands out as a promising solution for next-generation high-performance computing, thanks to the inherent advantages of light, such as low latency, high bandwidth, and unique multiplexing techniques. Furthermore, the progress in photonic integrated circuits (PICs), which are equipped with abundant photoelectronic components, positions photonic-electronic integrated circuits as a viable solution for high-performance computing and hardware AI accelerators. In this review, we survey recent advancements in both PIC-based digital and analog computing for AI, exploring the principal benefits and obstacles of implementation. Additionally, we propose a comprehensive analysis of photonic AI from the perspectives of hardware implementation, accelerator architecture, and software-hardware co-design. In the end, acknowledging the existing challenges, we underscore potential strategies for overcoming these issues and offer insights into the future drivers for optical computing.

An Optical modulator based on a hybrid optical waveguide made up of Silicon and Indium Tin Oxide (ITO) is proposed. The hybrid waveguide is created by the coupling of a leaky optical mode guided in silicon with ITO. The optical modulation is realized with the heterojunction between p-type silicon and n-type ITO. The presence of ITO causes electrical tuning in permittivity and in imaginary part of the refractive index which provide a way to modulate the intensity of the guide optical mode. The device with Si-ITO heterojunction is fabricated to show intensity modulation at 1.55μm wavelength. To achieve an efficient optical modulation the carrier concentration of ITO is optimized by depositing ITO on p-type Si wafer at different oxygen partial pressures using Ion assisted electro-beam deposition. The fabricated device shows an extinction ratio of 7 dB for 1.7mm long device at low voltage of -5 volts.

In this paper, we demonstrated the design and experimental results of the near-infrared lab-on-a-chip optical biosensor platform that monolithically integrates the MRR and the on-chip spectrometer on the silicon-on-insulator (SOI) wafer, which can eliminate the external optical spectrum analyzer for scanning the wavelength spectrum. The symmetric add-drop MRR biosensor is designed to have a free spectral range (FSR) of ∼19 nm and a bulk sensitivity of ∼73 nm/RIU; then the drop-port output resonance peaks are reconstructed from the integrated spatial-heterodyne Fourier transform spectrometer (SHFTS) with the spectral resolution of ∼3.1 nm and the bandwidth of ∼50 nm, which results in the limit of detection of 0.042 RIU.

At present, pesticides are widely used in the cultivation of crops. Glyphosate is widely used in many pesticides. Glyphosate ingestion can cause a series of health problems. Therefore, this paper proposes to use localized surface plasmon resonance (LSPR) technology to develop a WaveFlex biosensor (plasma wave-based optical fiber sensor) to detect glyphosate concentration in pesticides. The evanescent field is improved by using the fusion of seven-core fiber and single-mode fiber and the tapering of the sensing area to improve the sensing performance. The gold nanoparticles (AuNPs) are used to excite the LSPR effect. Multi-walled carbon nanotubes (MWCNTs) and cerium oxide nanorods (CeO₂-NRs) are used to increase the surface area and promote the adhesion of the enzyme. The sensitivity of the sensor is 137.7 pm/µM in the range of 0-60 µM glyphosate concentration, and the limit of detection (LoD) is 1.94 µM, which has good performance in compared to the existing biosensors. Subsequently, the sensor was tested for reusability, reproducibility, selectivity, stability, and excellent results were obtained. Finally, the sensor is tested on real samples, and the results show that it can be applied in practical applications. The test findings demonstrate that the sensor has a great deal of potential for use in glyphosate content detection in food samples.

An electrically writable resistive memory with optical readout based on silicon nanophotonic structure is proposed. Hybridization of optical and surface plasmonic modes in the device enables nanoscale optical confinement to efficiently detect the resistive memory effect in a 13 nm thick SiO2 layer sandwiched between p-type silicon and gold. Electrical write and optical readout capabilities of the proposed device are experimentally demonstrated with well-defined optical and electrical hysteresis curves at a wavelength of 1550 nm. The p-type silicon carries multifold benefits-it provides low propagation loss and a defect-free interface resulting from thermally (locally) grown oxide; the combination of p-silicon, SiO2, and gold results in a self-rectifying operation to enable the realization of a memory stack. An on-off extinction ratio of 10 dB is demonstrated for a 5 mm long device. The proposed device shows an inherent stochastic property where the set (writing) voltage reduces in each set-reset cycle, which can be used for optical readout of synaptic weight for neuromorphic computations.