Modelling Wall-Pressure Spectra in Turbulent Boundary Layers Using Neural Networks

In this work, we model the spectra of wall-pressure fluctuations beneath subsonic, supersonic and hypersonic turbulent boundary layers (TBLs) at zero pressure gradient using neural networks (NNs). We collect and compile data pertaining to wall-pressure fluctuation spectra from several experimental and computational studies on TBLs. In contrast to conventional methods of hand-tuning the parameters of a model to fit the available data, the use of modern powerful statistical learning techniques such as neural networks provide an automatic and quick way to fit a model. We explore four different scenarios of making use of the compiled data. In comparison with COMPRA-G, an empirical model recently proposed to account for compressibility effects in TBLs, we achieve a better fit to observed data using the NN model, particularly at low frequencies of the spectra.

An Introduction to Topological Data Analysis: Fundamental and Practical Aspects for Data Scientists

With the recent explosion in the amount, the variety, and the dimensionality of available data, identifying, extracting, and exploiting their underlying structure has become a problem of fundamental importance for data analysis and statistical learning. Topological data analysis (tda) is a recent and fast-growing field providing a set of new topological and geometric tools to infer relevant features for possibly complex data. It proposes new well-founded mathematical theories and computational tools that can be used independently or in combination with other data analysis and statistical learning techniques. This article is a brief introduction, through a few selected topics, to basic fundamental and practical aspects of tda for nonexperts.

Nonparametric inference for interventional effects with multiple mediators

Understanding the pathways whereby an intervention has an effect on an outcome is a common scientific goal. A rich body of literature provides various decompositions of the total intervention effect into pathway-specific effects. Interventional direct and indirect effects provide one such decomposition. Existing estimators of these effects are based on parametric models with confidence interval estimation facilitated via the nonparametric bootstrap. We provide theory that allows for more flexible, possibly machine learning-based, estimation techniques to be considered. In particular, we establish weak convergence results that facilitate the construction of closed-form confidence intervals and hypothesis tests and prove multiple robustness properties of the proposed estimators. Simulations show that inference based on large-sample theory has adequate small-sample performance. Our work thus provides a means of leveraging modern statistical learning techniques in estimation of interventional mediation effects.

Beyond Traditional Measures of Personality with BISCUIT and BARE: A New Statistical Learning Technique and Behavioral Item Pool to Push Personality Psychology Forward

This dissertation investigates two ways in which personality psychology should move beyond the traditional approach of measuring personality with broad domains composed of trait descriptors, as exemplified by the Big Five taxonomy. The first study (Chapter 2) suggests an alternative to the traditional approach of aggregating personality items into domains. Mounting evidence indicates that, compared to domains, narrower measures of personality account for more variance in criteria and describe personality-criterion relationships more accurately. Analysis of individual personality items is the most granular approach to studying personality and is typically performed with statistical learning techniques (SLTs). The first study: (a) champions a new statistical learning technique, BISCUIT; (b) finds that BISCUIT provides a balance between prediction and parsimony; and (c) replicates previous findings that the broadness of the Big Five traits hinder their predictive power.The second study (Chapter 3) suggests an alternative to the traditional approach of measuring personality with trait descriptors, or "traditional personality items." Of the three patterns commonly associated with personality (cognitions, emotions, and behaviors), behaviors are the least studied; traditional personality items tend to measure cognitions and emotions. Historically, yearlong patterns of specific behaviors have been thought of as criteria of personality measures, but the second study posits they should be classified as personality items because they measure patterns of behavior, a component of personality. The second study reviews and extends two pilot studies that indicated behavioral frequencies predict life outcomes, sometimes better than traditional personality items. The second study: (a) estimates the extent to which behavioral frequencies strengthen personality-criterion relationships above traditional personality items; (b) determines that some criteria are differentially predicted by personality item type; and (c) publishes an updated, public-domain item pool of behavioral frequencies: the BARE (Behavioral Acts, Revised and Expanded) Inventory.

Verbal Working Memory, Long-Term Knowledge, and Statistical Learning

Evidence supporting the idea that serial-order verbal working memory is underpinned by long-term knowledge has accumulated over more than half a century. Recent studies using natural-language statistics, artificial statistical-learning techniques, and the Hebb repetition paradigm have revealed multiple types of long-term knowledge underlying serial-order verbal working memory performance. These include (a) element-to-element association knowledge, which slowly accumulates through extensive exposure to an exemplar; (b) position–element knowledge, which is acquired through several encounters with an exemplar; and (c) whole-sequence knowledge, which is captured by the Hebb repetition paradigm and acquired rapidly with a few repetitions. Arguably, the first two are a basis for fluent and efficient language usage, and the third is a basis for vocabulary learning. Thus, statistical-learning mechanisms (and possibly episodic-learning mechanisms) may form the foundation of language acquisition and language processing, which characterize linguistic long-term knowledge for verbal working memory.