Machine learning operators, such as neural networks, are universal function approximators—albeit, in practice, their generalization ability depends on the quality of the training data and the algorithm designer’s wisdom in choosing a particular operator form, i.e. how well it matches the function at hand. Scientific machine learning is a class of methods that constrain the neural network operator by forcing its output to match time-series data from a partially known dynamical model, e.g. an ordinary or partial vector differential equation. In this talk, we make the case for regularizing optical image measurements using this approach. Applications are expected to be in processes with high-complexity constitutive relationships, such as pharmaceutical and cell manufacturing, plant biology, and ecology.
We discuss the use of machine learning in computational imaging for manufacturing process inspection and control. In a recent article we described a physics-enhanced auto-correlation based estimator (Peace) for quantitative speckle. We derived an explicit forward relationship between the Particle Size Distribution (PSD) and the speckle autocorrelation for particle sizes significantly larger than the wavelength (x100 to approximately x1,000). We subsequently trained a machine learning kernel to invert the autocorrelation and obtain the PSD, using the explicit forward model to reduce the number of experimentally acquired examples. In this talk, we present an expanded discussion of Peace and its properties, including spatial and temporal sampling and accuracy, and more general applications.
The development of low-loss optical phase change materials (O-PCMs) promises to enable a plethora of nonvolatile integrated photonic applications. However, the relatively large optical constants change between different states of calls for a set of new design rationales. Here we report a non-perturbative design that enables low-loss device operation beyond the traditional figure-of-merit limit. The basic design rationale is to engineer the light propagation path through the OPCMs when it is in the low-loss amorphous state, and divert light away from the lossy crystalline state leveraging the large mode modification induced by the O-PCM phase transition. Following this approach, we demonstrate broadband photonic switches with significantly enhanced performances compared to current state-of-the-art.
Optical phase change materials (O-PCMs) are a unique class of materials which exhibit extraordinarily large optical property change (e.g. refractive index change > 1) when undergoing a solid-state phase transition. Traditional O-PCMs suffer from large optical losses even in their dielectric states, which fundamentally limits the performance of optical devices based on the materials. To resolve the issue, we have recently demonstrated a new O-PCM Ge-Sb-Se-Te (GSST) with broadband low loss characteristics. In this talk, we will review an array of reconfigurable photonic devices enabled by the low-loss O-PCM, including nonvolatile waveguide switches with unprecedented low-loss and high-contrast performance, free-space light modulators, bi-stable reconfigurable metasurfaces, and transient couplers facilitating waferscale device probing and characterizations.
The dramatic optical property change of optical phase change materials (O-PCMs) between their amorphous and crystalline states potentially allows the realization of reconfigurable photonics devices with low power consumption, such as optical switches and routers, reconfigurable meta-optics, displays, and photonic memories. However, conventional O-PCMs, such as VO2 and Ge2Sb2Te5, are inherently plagued by their excessive optical losses even in dielectric states, limiting their optical performance and hence application space. In this talk, we present the development of a new group of O-PCMs and their implementations in novel photonic devices. Ge-Sb-Se-Te (GSST), obtained by partially substituting Te with Se in traditional GST alloys, feature unprecedented broadband optical transparency covering the telecommunication bands to LWIR. Capitalizing on the dramatically-enhanced optical performance, novel non-volatile, reconfigurable on-chip photonics devices and architectures are demonstrated. GSST-integrated Si photonics based on the material innovation and novel “non-perturbative” designs exhibit significantly improved switching performance over state-of-the-art GST-based approaches. The technology is further scalable to realize non-blocking matrix switches with arbitrary network complexity, paving the path towards high performance reconfigurable photonics chips.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.