arXiv subjects

Paul Wilson

Publications and source records attributed to Paul Wilson.

30 records · Page 2Linked to original sources

arXiv2022

Is big team research fair in national research assessments? The case of the UK Research Excellence Framework 2021

Collaborative research causes problems for research assessments because of the difficulty in fairly crediting its authors. Whilst splitting the rewards for an article amongst its authors has the greatest surface-level fairness, many important evaluations assign full credit to each author, irrespective of team size. The underlying rationales for this are labour reduction and the need to incentivise collaborative work because it is necessary to solve many important societal problems. This article assesses whether full counting changes results compared to fractional counting in the case of the UK's Research Excellence Framework (REF) 2021. For this assessment, fractional counting reduces the number of journal articles to as little as 10% of the full counting value, depending on the Unit of Assessment (UoA). Despite this large difference, allocating an overall grade point average (GPA) based on full counting or fractional counting give results with a median Pearson correlation within UoAs of 0.98. The largest changes are for Archaeology (r=0.84) and Physics (r=0.88). There is a weak tendency for higher scoring institutions to lose from fractional counting, with the loss being statistically significant in 5 of the 34 UoAs. Thus, whilst the apparent over-weighting of contributions to collaboratively authored outputs does not seem too problematic from a fairness perspective overall, it may be worth examining in the few UoAs in which it makes the most difference.

cs.DL↗

arXiv2022

The Cost of Compositionality: A High-Performance Implementation of String Diagram Composition

String diagrams are an increasingly popular algebraic language for the analysis of graphical models of computations across different research fields. Whereas string diagrams have been thoroughly studied as semantic structures, much less attention has been given to their algorithmic properties, and efficient implementations of diagrammatic reasoning are almost an unexplored subject. This work intends to be a contribution in such a direction. We introduce a data structure representing string diagrams in terms of adjacency matrices. This encoding has the key advantage of providing simple and efficient algorithms for composition and tensor product of diagrams. We demonstrate its effectiveness by showing that the complexity of the two operations is linear in the size of string diagrams. Also, as our approach is based on basic linear algebraic operations, we can take advantage of heavily optimised implementations, which we use to measure performances of string diagrammatic operations via several benchmarks.

math.CT↗

arXiv2022

Towards Confident Detection of Prostate Cancer using High Resolution Micro-ultrasound

MOTIVATION: Detection of prostate cancer during transrectal ultrasound-guided biopsy is challenging. The highly heterogeneous appearance of cancer, presence of ultrasound artefacts, and noise all contribute to these difficulties. Recent advancements in high-frequency ultrasound imaging - micro-ultrasound - have drastically increased the capability of tissue imaging at high resolution. Our aim is to investigate the development of a robust deep learning model specifically for micro-ultrasound-guided prostate cancer biopsy. For the model to be clinically adopted, a key challenge is to design a solution that can confidently identify the cancer, while learning from coarse histopathology measurements of biopsy samples that introduce weak labels. METHODS: We use a dataset of micro-ultrasound images acquired from 194 patients, who underwent prostate biopsy. We train a deep model using a co-teaching paradigm to handle noise in labels, together with an evidential deep learning method for uncertainty estimation. We evaluate the performance of our model using the clinically relevant metric of accuracy vs. confidence. RESULTS: Our model achieves a well-calibrated estimation of predictive uncertainty with area under the curve of 88$\%$. The use of co-teaching and evidential deep learning in combination yields significantly better uncertainty estimation than either alone. We also provide a detailed comparison against state-of-the-art in uncertainty estimation.

eess.IV↗

arXiv2022

Categories of Differentiable Polynomial Circuits for Machine Learning

Reverse derivative categories (RDCs) have recently been shown to be a suitable semantic framework for studying machine learning algorithms. Whereas emphasis has been put on training methodologies, less attention has been devoted to particular \emph{model classes}: the concrete categories whose morphisms represent machine learning models. In this paper we study presentations by generators and equations of classes of RDCs. In particular, we propose \emph{polynomial circuits} as a suitable machine learning model. We give an axiomatisation for these circuits and prove a functional completeness result. Finally, we discuss the use of polynomial circuits over specific semirings to perform machine learning with discrete values.

cs.LG↗

arXiv2021

Categorical Foundations of Gradient-Based Learning

We propose a categorical semantics of gradient-based machine learning algorithms in terms of lenses, parametrised maps, and reverse derivative categories. This foundation provides a powerful explanatory and unifying framework: it encompasses a variety of gradient descent algorithms such as ADAM, AdaGrad, and Nesterov momentum, as well as a variety of loss functions such as as MSE and Softmax cross-entropy, shedding new light on their similarities and differences. Our approach to gradient-based learning has examples generalising beyond the familiar continuous domains (modelled in categories of smooth maps) and can be realized in the discrete setting of boolean circuits. Finally, we demonstrate the practical significance of our framework with an implementation in Python.

cs.LG↗

arXiv2021

A Bayesian Hurdle Quantile Regression Model for Citation Analysis with Mass Points at Lower Values

Quantile regression presents a complete picture of the effects on the location, scale, and shape of the dependent variable at all points, not just the mean. We focus on two challenges for citation count analysis by quantile regression: discontinuity and substantial mass points at lower counts. A Bayesian hurdle quantile regression model for count data with a substantial mass point at zero was proposed by King and Song (2019). It uses quantile regression for modeling the nonzero data and logistic regression for modeling the probability of zeros versus nonzeros. We show that substantial mass points for low citation counts will nearly certainly also affect parameter estimation in the quantile regression part of the model, similar to a mass point at zero. We update the King and Song model by shifting the hurdle point past the main mass points. This model delivers more accurate quantile regression for moderately to highly cited articles, especially at quantiles corresponding to values just beyond the mass points, and enables estimates of the extent to which factors influence the chances that an article will be low cited. To illustrate the potential of this method, it is applied to simulated citation counts and data from Scopus.

stat.ME↗

arXiv2021

Category Theory in Machine Learning

Over the past two decades machine learning has permeated almost every realm of technology. At the same time, many researchers have begun using category theory as a unifying language, facilitating communication between different scientific disciplines. It is therefore unsurprising that there is a burgeoning interest in applying category theory to machine learning. We aim to document the motivations, goals and common themes across these applications. We touch on gradient-based learning, probability, and equivariant learning.

cs.LG↗

arXiv2021

Reverse Derivative Ascent: A Categorical Approach to Learning Boolean Circuits

We introduce Reverse Derivative Ascent: a categorical analogue of gradient based methods for machine learning. Our algorithm is defined at the level of so-called reverse differential categories. It can be used to learn the parameters of models which are expressed as morphisms of such categories. Our motivating example is boolean circuits: we show how our algorithm can be applied to such circuits by using the theory of reverse differential categories. Note our methodology allows us to learn the parameters of boolean circuits directly, in contrast to existing binarised neural network approaches. Moreover, we demonstrate its empirical value by giving experimental results on benchmark machine learning datasets.

cs.LO↗

arXiv2019

Exocometary Science

Evidence for exocomets, icy bodies in extrasolar planetary systems, has rapidly increased over the past decade. Volatiles are detected through the gas that exocomets release as they collide and grind down within their natal belts, or as they sublimate once scattered inwards to the regions closest to their host star. Most detections are in young, 10 to a few 100 Myr-old systems that are undergoing the final stages of terrestrial planet formation. This opens the exciting possibility to study exocomets at the epoch of volatile delivery to the inner regions of planetary systems. Detection of molecular and atomic gas in exocometary belts allows us to estimate molecular ice abundances and overall elemental abundances, enabling comparison with the Solar Nebula and Solar System comets. At the same time, observing star-grazing exocomets transiting in front of their star (for planetary systems viewed edge-on) and exozodiacal dust in the systems' innermost regions gives unique dynamical insights into the inward scattering process producing delivery to inner rocky planets. The rapid advances of this budding subfield of exoplanetary science will continue in the short term with the upcoming JWST, WFIRST and PLATO missions. In the longer term, the priority should be to explore the full composition of exocomets, including species crucial for delivery and later prebiotic synthesis. Doing so around an increasingly large population of exoplanetary systems is equally important, to enable comparative studies of young exocomets at the epoch of volatile delivery. We identify the proposed LUVOIR and Origins flagship missions as the most promising for a large-scale exploration of exocometary gas, a crucial component of the chemical heritage of young exo-Earths.

astro-ph.EP↗

arXiv2015

Distributions for cited articles from individual subjects and years

The citations to a set of academic articles are typically unevenly shared, with many articles attracting few citations and few attracting many. It is important to know more precisely how citations are distributed in order to help statistical analyses of citations, especially for sets of articles from a single discipline and a small range of years, as normally used for research evaluation. This article fits discrete versions of the power law, the lognormal distribution and the hooked power law to 20 different Scopus categories, using citations to articles published in 2004 and ignoring uncited articles. The results show that, despite its popularity, the power law is not a suitable model for collections of articles from a single subject and year, even for the purpose of estimating the slope of the tail of the citation data. Both the hooked power law and the lognormal distributions fit best for some subjects but neither is a universal optimal choice and parameter estimates for both seem to be unreliable. Hence only the hooked power law and discrete lognormal distributions should be considered for subject-and-year-based citation analysis in future and parameter estimates should always be interpreted cautiously.

cs.DL↗

arXiv2015

Regression for citation data: An evaluation of different methods

Citations are increasingly used for research evaluations. It is therefore important to identify factors affecting citation scores that are unrelated to scholarly quality or usefulness so that these can be taken into account. Regression is the most powerful statistical technique to identify these factors and hence it is important to identify the best regression strategy for citation data. Citation counts tend to follow a discrete lognormal distribution and, in the absence of alternatives, have been investigated with negative binomial regression. Using simulated discrete lognormal data (continuous lognormal data rounded to the nearest integer) this article shows that a better strategy is to add one to the citations, take their log and then use the general linear (ordinary least squares) model for regression (e.g., multiple linear regression, ANOVA), or to use the generalised linear model without the log. Reasonable results can also be obtained if all the zero citations are discarded, the log is taken of the remaining citation counts and then the general linear model is used, or if the generalised linear model is used with the continuous lognormal distribution. Similar approaches are recommended for altmetric data, if it proves to be lognormally distributed.

cs.DL↗

arXiv2013

Best Practices for Scientific Computing

Scientists spend an increasing amount of time building and using software. However, most scientists are never taught how to do this efficiently. As a result, many are unaware of tools and practices that would allow them to write more reliable and maintainable code with less effort. We describe a set of best practices for scientific software development that have solid foundations in research and experience, and that improve scientists' productivity and the reliability of their software.

cs.MS↗