Search

The development of bare fiber or air-gapped microlens-fiber coupled Integral Field Units (IFUs) for astronomical applications requires careful treatment of the fiber end-faces (terminations). Previous studies suggest that minimization of fiber end face irregularity leads to better optical performance in terms of the diminishing effect of focal ratio degradation. Polishing has typically been performed using commercial rotary polishers with multiple gradually decreasing grit sizes. These polishers generally lack the ability to carefully adjust angular position and polishing force. Control of these parameters vastly help in getting a repeatable and controllable polish over a variety of glass/epoxy/metal matrices that make up integral filed units and fiber slits. A polishing arm is developed to polish the fiber terminations (IFU, mini-bundles and v-grooves) of the NIR Fiber System for the RSS spectrograph at SALT. The polishing arm angular adjustments ensure the correct position and orientation of each termination on the polishing surface during the polish. Various studies have indicated that the fiber focal ratio also degrades if the fiber end face comes under excessive stress. The polishing arm is fitted with a load cell to enable control of the polishing force. We have explored the minimal applicable end stress by applying different loads while polishing. The arm is modular to hold a variety of fiber termination styles. The polishing arm is also designed to access a fiber inspection microscope without removing the fiber termination from the arm. This enables inspection of the finish quality at various stages through polishing process.

Recurrent neural networks (RNNs) with rich feature vectors of past values can provide accurate point forecasts for series that exhibit complex serial dependence. We propose two approaches to constructing deep time series probabilistic models based on a variant of RNN called an echo state network (ESN). The first is where the output layer of the ESN has stochastic disturbances and a shrinkage prior for additional regularization. The second approach employs the implicit copula of an ESN with Gaussian disturbances, which is a deep copula process on the feature space. Combining this copula with a non-parametrically estimated marginal distribution produces a deep distributional time series model. The resulting probabilistic forecasts are deep functions of the feature vector and also marginally calibrated. In both approaches, Bayesian Markov chain Monte Carlo methods are used to estimate the models and compute forecasts. The proposed models are suitable for the complex task of forecasting intraday electricity prices. Using data from the Australian National Electricity Market, we show that our deep time series models provide accurate short term probabilistic price forecasts, with the copula model dominating. Moreover, the models provide a flexible framework for incorporating probabilistic forecasts of electricity demand as additional features, which increases upper tail forecast accuracy from the copula model significantly.

We present 'Pix2Prof', a deep learning model that can eliminate any manual steps taken when extracting galaxy profiles. We argue that a galaxy profile of any sort is conceptually similar to a natural language image caption. This idea allows us to leverage image captioning methods from the field of natural language processing, and so we design Pix2Prof as a float sequence 'captioning' model suitable for galaxy profile inference. We demonstrate the technique by approximating a galaxy surface brightness (SB) profile fitting method that contains several manual steps. Pix2Prof processes $\sim$1 image per second on an Intel Xeon E5 2650 v3 CPU, improving on the speed of the manual interactive method by more than two orders of magnitude. Crucially, Pix2Prof requires no manual interaction, and since galaxy profile estimation is an embarrassingly parallel problem, we can further increase the throughput by running many Pix2Prof instances simultaneously. In perspective, Pix2Prof would take under an hour to infer profiles for $10^5$ galaxies on a single NVIDIA DGX-2 system. A single human expert would take approximately two years to complete the same task. Automated methodology such as this will accelerate the analysis of the next generation of large area sky surveys expected to yield hundreds of millions of targets. In such instances, all manual approaches -- even those involving a large number of experts -- will be impractical., Comment: Accepted for publication in MNRAS. 10 pages, and 8 figures. Code: https://github.com/Smith42/pix2prof

We present AstroVaDEr, a variational autoencoder designed to perform unsupervised clustering and synthetic image generation using astronomical imaging catalogues. The model is a convolutional neural network that learns to embed images into a low dimensional latent space, and simultaneously optimises a Gaussian Mixture Model (GMM) on the embedded vectors to cluster the training data. By utilising variational inference, we are able to use the learned GMM as a statistical prior on the latent space to facilitate random sampling and generation of synthetic images. We demonstrate AstroVaDEr's capabilities by training it on gray-scaled \textit{gri} images from the Sloan Digital Sky Survey, using a sample of galaxies that are classified by Galaxy Zoo 2. An unsupervised clustering model is found which separates galaxies based on learned morphological features such as axis ratio, surface brightness profile, orientation and the presence of companions. We use the learned mixture model to generate synthetic images of galaxies based on the morphological profiles of the Gaussian components. AstroVaDEr succeeds in producing a morphological classification scheme from unlabelled data, but unexpectedly places high importance on the presence of companion objects---demonstrating the importance of human interpretation. The network is scalable and flexible, allowing for larger datasets to be classified, or different kinds of imaging data. We also demonstrate the generative properties of the model, which allow for realistic synthetic images of galaxies to be sampled from the learned classification scheme. These can be used to create synthetic image catalogs or to perform image processing tasks such as deblending., Comment: Accepted in MNRAS. 23 Pages, 16 Figures. GitHub: https://github.com/AshleySpindler/AstroVaDEr-Public

A generic fiber positioning strategy and a fabrication path are presented for microlens-fiber-coupled integral field units. It is assumed that microlens-produced micro-images are carried to the spectrograph input through step-index,multi-mode fiber, but our results apply to micro-pupil reimaging applications as well. Considered are the performance trades between the filling percentage of the fiber core with the micro-image versus throughput and observing efficiency.A merit function is defined as the product of the transmission efficiency and the etendue loss. For a hexagonal packing of spatial elements, the merit function has been found to be maximized to 94% of an ideal fiber IFU merit value (which has zero transmission loss and does not increase the etendue) with a microlens-fiber alignment (centering) tolerance of 1 um RMS. The maximum acceptable relative tilt between the fiber and the microlens face has been analyzed through optical modeling and found to be 0.3 degree RMS for input f-ratio slower than f/3.5 but it is much more relaxed for faster beams. Several options of fabricating fiber holders have been compared to identify cost-effective solutions that deliver the desired fiber positioning accuracy. Femto-second laser-drilling methods deliver holes arrayed on plates with a position and diameter accuracy of 1.5 um RMS, and with an aspect ratio of 1:10. A commercial vendor produces plates with thickness of 5 mm, but with similar (1 um RMS) positioning accuracy. Both of these techniques are found to be moderately expensive. A purely photo-lithographic technique performed at WCAM (a facility at the University of Wisconsin, Madison), in tandem with deep reactive ion etching, has been used to produce a repeatable recipe with 100% yield. Photo-lithography is more precise (0.5 um RMS) in terms of hole positioning and similar diameter accuracy (1 um RMS)., Comment: 26 Pages, 18 Figures, 1 Table

This report is an outcome of the workshop "AI for Nuclear Physics" held at Thomas Jefferson National Accelerator Facility on March 4-6, 2020. The workshop brought together 184 scientists to explore opportunities for Nuclear Physics in the area of Artificial Intelligence. The workshop consisted of plenary talks, as well as six working groups. The report includes the workshop deliberations and additional contributions to describe prospects for using AI across Nuclear Physics research., Comment: This version includes reference updates, improved figures and minor clarifications in the text

Deliberation among individuals online plays a key role in shaping the opinions that drive votes, purchases, donations and other critical offline behavior. Yet, the determinants of opinion-change via persuasion in deliberation online remain largely unexplored. Our research examines the persuasive power of $\textit{ethos}$ -- an individual's "reputation" -- using a 7-year panel of over a million debates from an argumentation platform containing explicit indicators of successful persuasion. We identify the causal effect of reputation on persuasion by constructing an instrument for reputation from a measure of past debate competition, and by controlling for unstructured argument text using neural models of language in the double machine-learning framework. We find that an individual's reputation significantly impacts their persuasion rate above and beyond the validity, strength and presentation of their arguments. In our setting, we find that having 10 additional reputation points causes a 31% increase in the probability of successful persuasion over the platform average. We also find that the impact of reputation is moderated by characteristics of the argument content, in a manner consistent with a theoretical model that attributes the persuasive power of reputation to heuristic information-processing under cognitive overload. We discuss managerial implications for platforms that facilitate deliberative decision-making for public and private organizations online.

Many environments currently employ machine learning models for data processing and analytics that were built using a limited number of training data points. Once deployed, the models are exposed to significant amounts of previously-unseen data, not all of which is representative of the original, limited training data. However, updating these deployed models can be difficult due to logistical, bandwidth, time, hardware, and/or data sensitivity constraints. We propose a framework, Self-Updating Models with Error Remediation (SUMER), in which a deployed model updates itself as new data becomes available. SUMER uses techniques from semi-supervised learning and noise remediation to iteratively retrain a deployed model using intelligently-chosen predictions from the model as the labels for new training iterations. A key component of SUMER is the notion of error remediation as self-labeled data can be susceptible to the propagation of errors. We investigate the use of SUMER across various data sets and iterations. We find that self-updating models (SUMs) generally perform better than models that do not attempt to self-update when presented with additional previously-unseen data. This performance gap is accentuated in cases where there is only limited amounts of initial training data. We also find that the performance of SUMER is generally better than the performance of SUMs, demonstrating a benefit in applying error remediation. Consequently, SUMER can autonomously enhance the operational capabilities of existing data processing systems by intelligently updating models in dynamic environments., Comment: 17 pages, 13 figures, published in the proceedings of the Artificial Intelligence and Machine Learning for Multi-Domain Operations Applications II conference in the SPIE Defense + Commercial Sensing, 2020 symposium

Models with a large number of latent variables are often used to fully utilize the information in big or complex data. However, they can be difficult to estimate using standard approaches, and variational inference methods are a popular alternative. Key to the success of these is the selection of an approximation to the target density that is accurate, tractable and fast to calibrate using optimization methods. Most existing choices can be inaccurate or slow to calibrate when there are many latent variables. Here, we propose a family of tractable variational approximations that are more accurate and faster to calibrate for this case. It combines a parsimonious parametric approximation for the parameter posterior, with the exact conditional posterior of the latent variables. We derive a simplified expression for the re-parameterization gradient of the variational lower bound, which is the main ingredient of efficient optimization algorithms used to implement variational estimation. To do so only requires the ability to generate exactly or approximately from the conditional posterior of the latent variables, rather than to compute its density. We illustrate using two complex contemporary econometric examples. The first is a nonlinear multivariate state space model for U.S. macroeconomic variables. The second is a random coefficients tobit model applied to two million sales by 20,000 individuals in a large consumer panel from a marketing study. In both cases, we show that our approximating family is considerably more accurate than mean field or structured Gaussian approximations, and faster than Markov chain Monte Carlo. Last, we show how to implement data sub-sampling in variational inference for our approximation, which can lead to a further reduction in computation time. MATLAB code implementing the method for our examples is included in supplementary material., Comment: Macroeconomic example was replaced by the bigger and more challenging time varying parameter vector autoregression model with stochastic volatility. Microeconomic example was extended to 20,000 individuals and variational subsampling is also implemented for this example. Small microeconomics example now uses 1000 individuals

Despite the potential of Machine learning (ML) to learn the behavior of malware, detect novel malware samples, and significantly improve information security (InfoSec) we see few, if any, high-impact ML techniques in deployed systems, notwithstanding multiple reported successes in open literature. We hypothesize that the failure of ML in making high-impacts in InfoSec are rooted in a disconnect between the two communities as evidenced by a semantic gap---a difference in how executables are described (e.g. the data and features extracted from the data). Specifically, current datasets and representations used by ML are not suitable for learning the behaviors of an executable and differ significantly from those used by the InfoSec community. In this paper, we survey existing datasets used for classifying malware by ML algorithms and the features that are extracted from the data. We observe that: 1) the current set of extracted features are primarily syntactic, not behavioral, 2) datasets generally contain extreme exemplars producing a dataset in which it is easy to discriminate classes, and 3) the datasets provide significantly different representations of the data encountered in real-world systems. For ML to make more of an impact in the InfoSec community requires a change in the data (including the features and labels) that is used to bridge the current semantic gap. As a first step in enabling more behavioral analyses, we label existing malware datasets with behavioral features using open-source threat reports associated with malware families. This behavioral labeling alters the analysis from identifying intent (e.g. good vs bad) or malware family membership to an analysis of which behaviors are exhibited by an executable. We offer the annotations with the hope of inspiring future improvements in the data that will further bridge the semantic gap between the ML and InfoSec communities., Comment: 14 pages, 2 Figures, 6 tables

Recently, hollow thermoplastic microspheres, such as Expancel made by Nouryon, have emerged as an innovative filler material for use in polymer-matrix composites. The resulting all-polymer syntactic foam takes on excellent damage tolerance properties, strong recoverability under large strains, and favourable energy dissipation characteristics. Despite finding increasing usage in various industries and applications, including in coatings, films, sealants, packaging, composites for microfluidics, medical ultrasonics and cementious composites, there is a near-complete absence of statistical geometrical information for Expancel microspheres. Further, their mechanical properties have not yet been reported. In this work we characterise the geometrical quantities of two classes of Expancel thermoplastic microspheres using X-ray computed tomography, focused ion beam and electron microscopy. We also observe the spatial distribution of microspheres within a polyurethane-matrix syntactic foam. We show that the volume-weighted polydisperse shell diameter in both classes of microsphere follows a normal distribution. Interestingly, polydispersity of the shell wall thickness is not observed and in particular the shell thickness is not correlated to the shell diameter. We employ the measured geometrical information in analytical micromechanical techniques in the small strain regime to determine, for the first time, estimates of the Young's modulus and Poisson's ratio of the microsphere shell material. Our results contribute to potential future improvements in the design and fabrication of syntactic foams that employ thermoplastic microspheres. Given the breadth of fields which utilise thermoplastic microspheres, we anticipate that our results, together with the methods used, will be of use in a much broader context in future materials research., Comment: 19 page, 7 figure, 3 tables

Most stars form in binaries, and both stars may grow by accreting material from a circumbinary disc onto their personal discs. We suspect that in many cases a wide molecular wind will envelope a collimated atomic jet emanating from close to an orbiting young star. This so-called Circumbinary Scenario is explored here in order to find common identifiable properties. The dynamical set up is studied with three dimensional simulations with chemistry and cooling included. We extract the properties on scales of order 100\,AU and compare to the Co-Orbital Scenario in which the wind and jet sources are in orbit. We find that the rapid orbital motion generates a wide ionised sheath around the jet core with a large opening angle at the base. This is independent of the presence of the surrounding molecular outflow. However, the atomic jet is recollimated beyond ~ 55 AU when the molecular outflow restricts the motion of the ambient medium which, in turn, confines the jet. These physical properties are relayed to the optical H-alpha imaging, providing a means of distinguishing between models. The high excitation sheath and recollimation region can be explored on these scales through the next generation of instruments. However, in general, the amount and location of the ionised material, whether in the knots or the sheath, will depend on several parameters including the orbital period, axis alignment and pulse amplitude., Comment: 12 pages, 13 figures, submitted to MNRAS