Geometric Matrix Completion With Deep Conditional Random Fields.

The problem of completing high-dimensional matrices from a limited set of observations arises in many big data applications, especially recommender systems. The existing matrix completion models generally follow either a memory- or a model-based approach, whereas geometric matrix completion (GMC) models combine the best from both approaches. Existing deep-learning-based geometric models yield good performance, but, in order to operate, they require a fixed structure graph capturing the relationships among the users and items. This graph is typically constructed by evaluating a pre-defined similarity metric on the available observations or by using side information, e.g., user profiles. In contrast, Markov-random-fields-based models do not require a fixed structure graph but rely on handcrafted features to make predictions. When no side information is available and the number of available observations becomes very low, existing solutions are pushed to their limits. In this article, we propose a GMC approach that addresses these challenges. We consider matrix completion as a structured prediction problem in a conditional random field (CRF), which is characterized by a maximum a posteriori (MAP) inference, and we propose a deep model that predicts the missing entries by solving the MAP inference problem. The proposed model simultaneously learns the similarities among matrix entries, computes the CRF potentials, and solves the inference problem. Its training is performed in an end-to-end manner, with a method to supervise the learning of entry similarities. Comprehensive experiments demonstrate the superior performance of the proposed model compared to various state-of-the-art models on popular benchmark data sets and underline its superior capacity to deal with highly incomplete matrices.

Use of a Digital Modified Checklist for Autism in Toddlers - Revised with Follow-up to Improve Quality of Screening for Autism.

OBJECTIVES: To assess changes in quality of care for children at risk for autism spectrum disorders (ASD) due to process improvement and implementation of a digital screening form. STUDY DESIGN: The process of screening for ASD was studied in an academic primary care pediatrics clinic before and after implementation of a digital version of the Modified Checklist for Autism in Toddlers - Revised with Follow-up with automated risk assessment. Quality metrics included accuracy of documentation of screening results and appropriate action for positive screens (secondary screening or referral). Participating physicians completed pre- and postintervention surveys to measure changes in attitudes toward feasibility and value of screening for ASD. Evidence of change was evaluated with statistical process control charts and χ2 tests. RESULTS: Accurate documentation in the electronic health record of screening results increased from 54% to 92% (38% increase, 95% CI 14%-64%) and appropriate action for children screening positive increased from 25% to 85% (60% increase, 95% CI 35%-85%). A total of 90% of participating physicians agreed that the transition to a digital screening form improved their clinical assessment of autism risk. CONCLUSIONS: Implementation of a tablet-based digital version of the Modified Checklist for Autism in Toddlers - Revised with Follow-up led to improved quality of care for children at risk for ASD and increased acceptability of screening for ASD. Continued efforts towards improving the process of screening for ASD could facilitate rapid, early diagnosis of ASD and advance the accuracy of studies of the impact of screening.

Information-Theoretic Compressive Measurement Design.

An information-theoretic projection design framework is proposed, of interest for feature design and compressive measurements. Both Gaussian and Poisson measurement models are considered. The gradient of a proposed information-theoretic metric (ITM) is derived, and a gradient-descent algorithm is applied in design; connections are made to the information bottleneck. The fundamental solution structure of such design is revealed in the case of a Gaussian measurement model and arbitrary input statistics. This new theoretical result reveals how ITM parameter settings impact the number of needed projection measurements, with this verified experimentally. The ITM achieves promising results on real data, for both signal recovery and classification.

Dual-wavelength pump-probe microscopy analysis of melanin composition.

Pump-probe microscopy is an emerging technique that provides detailed chemical information of absorbers with sub-micrometer spatial resolution. Recent work has shown that the pump-probe signals from melanin in human skin cancers correlate well with clinical concern, but it has been difficult to infer the molecular origins of these differences. Here we develop a mathematical framework to describe the pump-probe dynamics of melanin in human pigmented tissue samples, which treats the ensemble of individual chromophores that make up melanin as Gaussian absorbers with bandwidth related via Frenkel excitons. Thus, observed signals result from an interplay between the spectral bandwidths of the individual underlying chromophores and spectral proximity of the pump and probe wavelengths. The model is tested using a dual-wavelength pump-probe approach and a novel signal processing method based on gnomonic projections. Results show signals can be described by a single linear transition path with different rates of progress for different individual pump-probe wavelength pairs. Moreover, the combined dual-wavelength data shows a nonlinear transition that supports our mathematical framework and the excitonic model to describe the optical properties of melanin. The novel gnomonic projection analysis can also be an attractive generic tool for analyzing mixing paths in biomolecular and analytical chemistry.

Quantifying Risk for Anxiety Disorders in Preschool Children: A Machine Learning Approach.

Early childhood anxiety disorders are common, impairing, and predictive of anxiety and mood disorders later in childhood. Epidemiological studies over the last decade find that the prevalence of impairing anxiety disorders in preschool children ranges from 0.3% to 6.5%. Yet, less than 15% of young children with an impairing anxiety disorder receive a mental health evaluation or treatment. One possible reason for the low rate of care for anxious preschoolers is the lack of affordable, timely, reliable and valid tools for identifying young children with clinically significant anxiety. Diagnostic interviews assessing psychopathology in young children require intensive training, take hours to administer and code, and are not available for use outside of research settings. The Preschool Age Psychiatric Assessment (PAPA) is a reliable and valid structured diagnostic parent-report interview for assessing psychopathology, including anxiety disorders, in 2 to 5 year old children. In this paper, we apply machine-learning tools to already collected PAPA data from two large community studies to identify sub-sets of PAPA items that could be developed into an efficient, reliable, and valid screening tool to assess a young child's risk for an anxiety disorder. Using machine learning, we were able to decrease by an order of magnitude the number of items needed to identify a child who is at risk for an anxiety disorder with an accuracy of over 96% for both generalized anxiety disorder (GAD) and separation anxiety disorder (SAD). Additionally, rather than considering GAD or SAD as discrete/binary entities, we present a continuous risk score representing the child's risk of meeting criteria for GAD or SAD. Identification of a short question-set that assesses risk for an anxiety disorder could be a first step toward development and validation of a relatively short screening tool feasible for use in pediatric clinics and daycare/preschool settings.