Tomas Lopes | Searcharxiv

arXiv2026

Photonic Extreme Learning Machines using Event-Based detection

Photonic extreme learning machines use random optical propagation, detection nonlinearity, and a trained linear readout for energy-efficient and scalable optical computing. However, conventional intensity readout with CCD or CMOS cameras constrain the dimensionality of the hidden representation space. Here, we experimentally replace intensity detection with event-based camera, whose thresholded log-intensity response provides alternative pixel-wise hidden representations: first-event time, binary activation, and event count. In nonlinear two spiral classification task we obtain accuracies of $93\pm 3\%$ with strong intrinsic generalization, and comparable ridge and pseudo-inverse performance indicating sample-limited effective dimensionality. Regression results reveal sensitivity to systematic optical-intensity drift, identifying stability requirements for future event-based PELMs. These results establish event-based detection as a route toward richer photonic hidden representations while clarifying current limitations.

physics.optics↗

arXiv2026

Asynchronous Event-Based Spectroscopy for Microsecond-Resolved Spectral Reconstruction

Many physical and chemical processes of interest evolve on timescales that push the limits of conventional spectroscopic instrumentation. Indeed, the temporal resolution of standard spectrometers is often insufficient to track these dynamics, which is connected to the fact that most systems rely on frame-based sensors, imposing fundamental constraints on acquisition speed, sensitivity, and data efficiency, frequently limiting practical operation to the kHz regime. In this work, we present an approach to circumvent this limitation by developing an event-based spectrometer to enable spectral reconstruction with microsecond temporal resolution by leveraging a Czerny-Turner configuration combined with asynchronous and event-driven sensing. A dedicated signal processing pipeline converts the resulting stream of binary events into calibrated spectra through temporal accumulation, geometric correction, and vertical spatial integration of the spectral line, covering a 234nm bandwidth in the visible range with a spectral resolution of approximately 0.18nm per pixel. Performance characterization under temporally modulated illumination demonstrates that the event-based spectrometer can reconstruct spectra at probing rates of up to tens of kilohertz, far exceeding the practical limits of a conventional frame-based spectrometer operated in parallel, while accurately preserving spectral peak positions and relative spectral features. Finally, to further illustrate its potential applications, the system is validated in a microfluidic experiment integrated into an inverted microscope, where spectral changes induced by an absorbing dye are tracked with higher temporal fidelity and resolution compared with the frame-based approach. These results establish event-based spectroscopy as a promising paradigm for real-time, high-temporal-resolution spectral measurements in dynamic and low-light applications.

physics.optics↗