Reconfigurable linear optical processors can be used to perform linear transformations and are instrumental in effectively computing matrix–vector multiplications required in each neural network layer. In this paper, we characterize and compare two thermally tuned photonic integrated processors realized in silicon-on-insulator and silicon nitride platforms suited for extracting feature maps in convolutional neural networks. The reduction in bit resolution when crossing the processor is mainly due to optical losses, in the range 2.3–3.3 for the silicon-on-insulator chip and in the range 1.3–2.4 for the silicon nitride chip. However, the lower extinction ratio of Mach–Zehnder elements in the latter platform limits their expressivity (i.e., the capacity to implement any transformation) to 75%, compared to 97% of the former. Finally, the silicon-on-insulator processor outperforms the silicon nitride one in terms of footprint and energy efficiency.

We report the first monolithic InP photonic integrated comb generator made by cascading optical modulators. The device is extremely compact and tunable both in repetition frequency and wavelength, making real world applications possible. A DBR laser, a Mach-Zehnder intensity modulator and two phase modulators are monolithically integrated on a 4.5×2.5 mm $^2$ chip that also includes a booster SOA at the output. Modulators integrated on the circuit have a 3-dB bandwidth of 7 GHz and can generate up to 28 comb lines within a 5 dB power range when properly electrically driven (here demonstrated in the range 4-5 GHz). Operation up to 10 GHz is also reported.

We report broadband, all-optical wavelength conversion over 100 nm span, in full S- and C-band, with positive conversion efficiency with low optical input power exploiting dual pump Four-Wave-Mixing in a Quantum Dot Semiconductor Optical Amplifier (QD-SOA). We also demonstrate by Error Vector Magnitude analysis the full transparency of the conversion scheme for coherent modulation formats (QPSK, 8-PSK, 16-QAM, OFDM-16QAM) in the whole C-band.

We report an experimental study of the gain recovery of an InAs-InGaAsP-InP columnar-quantum-dot semiconductor optical amplifier around 1550 nm. We find a three-step recovery process that gives around 85% gain recovery in only 10 ps and 100% in around 30 ps in case of a quite complete gain saturation. We also give a possible physical explanation for the peculiar gain dynamics.

We study the limiting-amplification capability of a saturated Semiconductor Optical Amplifier (SOA) followed by an optical band-pass filter. We experimentally demonstrate that this simple optical circuit can be effectively exploited to realize a low-power optical limiter for amplitude-modulated pulse trains at multi-GHz repetition rate. We report very large amplitude-modulation-reduction factors for the case of 20 and 40 GHz pulse trains that are super-imposed with modulating frequencies ranging from 100kHz to several GHz.

While the use of graphic processing units fueled the success of artificial intelligence models, their future evolution will likely require overcoming the speed and energy efficiency limitations of current implementations with the use of specialized neuromorphic hardware. In this scenario, neuromorphic photonic processors have recently proved to be a feasible solution. In this paper, we first discuss basic analog photonic processing elements based on Mach-Zehnder modulators and assess their effective bit resolution. Then we evaluate how different photonic integration technologies affect the performance and the scalability of analog optical processors, in order to provide a clearer path toward real-world implementations of such engines. To this aim, we focus our analysis on the silicon on insulator (SOI), lithium niobate on insulator (LNOI), and indium phosphide (InP) platforms. In particular, we have numerically evaluated the performance of the Photonic Electronic Multiply- Accumulate Neuron (PEMAN) and its tensorial version, both based on Mach-Zehnder modulators, with the three technologies in terms of resolution, energy efficiency, and footprint efficiency. LNOI modulators achieve the best resolution at high speed, with 4.3 bits at 56 GMAC/s for the single PEMAN and 3.6 bits at 224 GMAC/S for the tensorial version. The energy consumption in InP and LNOI platforms is the lowest, accounting for just 13.2 pJ/MAC and 4.6 pJ/MAC for the single and tensorial PEMAN, respectively. Nonetheless, SOI devices outperform the others in terms of footprint efficiency, reaching 18.6 GMAC/s/mm $^{2}$ in the single-neuron version and 29.6 GMAC/s/mm $^{2}$ in the tensorial version.