Data Study Groups are week-long events at The Alan Turing Institute bringing together some of the country's top talent from data science, artificial intelligence, and wider fields, to analyse real-world data science challenges. NCSC: Capturing the purpose of websites This challenge was provided by the National Cyber Security Centre (NCSC), which is tasked with protecting the UK public sector, and aims at improving existing internet search technologies. As the internet grows, there is a need in industry, government and academia to better facilitate website recommendation, semantic search, and domain discovery, as well as to improve the security of the web. For instance, it is not possible to easily find all UK public sector domains with a simple search; nor it is possible for commercial organisations to easily generate a list of potential competitors or suppliers/customers from current search engines. An exciting potential approach to enable the NCSC and other organisations to leverage the latest machine learning algorithms on these challenges is to automatically learn a purposeful compact vector representation of every website or domain on the Web.
Data Study Groups are week-long events at The Alan Turing Institute bringing together some of the country's top talent from data science, artificial intelligence, and wider fields, to analyse real-world data science challenges. Automating the evaluation of local government planning applications It is estimated that England requires up to 345,000 new homes to be built each year in order to keep up with population growth. Although housebuilding has increased since reaching a low point after the financial crisis, the annual net supply of new homes would need to increase by more than 40% to reach this number. This lack of new homes contributes to higher mortgages, higher rents, less social housing and wider deprivation. One of the contributing factors is that the current planning application system remains a complex and inefficient task. Every construction project in the UK, including building or extending a house, fitting new windows on a listed building, or chopping down a tree, requires the submission of complicated planning forms and technical drawings. Each one of these documents needs to be manually validated and approved by a planning officer. Over 3.5 million applications are submitted to councils each year. On average, owing to local government budget cuts of 40% over the last 10 years, it takes three weeks for a council to start looking at a planning application. This creates large backlogs and a lack of information for application submitters, leading to additional calls and emails to chase progress which further increase workload on planning officers. At the same time, over a third (1.2 million) of the applications submitted annually are rejected, often owing to the complexity of submitting a correct application and the manual burden in processing them at councils. The overall objective of this work is to move towards the automated detection of common errors in planning applications using ML/AI approaches.
Data Study Groups are week-long events at The Alan Turing Institute bringing together some of the country's top talent from data science, artificial intelligence, and wider fields, to analyse real-world data science challenges. Smart monitoring for conservation areas WWF (World Wide Fund for Nature) monitors over 250,000 protected areas (e.g. national parks and nature reserves) and thousands of other sites and critical habitats. These sites are the foundation of global natural assets and are central to the preservation of biodiversity and human well-being. Unfortunately, they face increasing pressures from human development. In this challenge, we explore various data science techniques to automatically detect news articles that report emerging threats to key protected areas. We describe a system that identifies such news stories near real-time. This is vital to enable the wider machinery of WWF and the conservation community to engage with governments, companies, shareholders, insurers, and others to help halt the degradation or destruction of key habitats.
Data Study Groups are week-long events at The Alan Turing Institute bringing together some of the country's top talent from data science, artificial intelligence, and wider fields, to analyse real-world data science challenges. Spend Network: Automated matching of businesses to government contract opportunities This report presents the output of a week-long collaboration between Spend Network, and lead academics from the University of Manchester and the University of Oxford that attended the Data Study Group at The Alan Turing Institute. Spend Network is a platform that aims to enable efficient public procure-ment. The goal of this work was to match suppliers to tenders. Our approach consists of building vector representations of suppli-ers and tenders and identifying their compat-ibility with the distance between the respec-tive vectors. In building their representations, we make use of both their textual descriptions and the knowledge of previously awarded con-tracts. We find previous contracts informative of future procurement decisions. Our best re-sults use Correlated Topic Models (Blei et al., 2007) for extracting representations of textual descriptions.
In: Bulletin of the World Health Organization: the international journal of public health = Bulletin de l'Organisation Mondiale de la Santé, Volume 88, Issue 7, p. 509-518
Este artículo reseña la investigación que realizó un grupo de investigadoras de ANSIRH (Adelantando Nuevos Estándares en Investigación en Salud) de la Universidad de California en San Francisco (UCSF). Este estudio fue aprobado por el Comité de Investigación Humanas de UCSF y fue financiado por la Fundación David and Lucille Packard, por la Fundación Wallace Alexander Gerbode y otros donantes privados. Traducción por: Gabriela Castellanos
Data Study Groups are week-long events at The Alan Turing Institute bringing together some of the country's top talent from data science, artificial intelligence, and wider fields, to analyse real-world data science challenges. Discovering topics and trends in the UK government web archive The challenge we address in this report is to make steps towards improving search and discovery of resources within this vast archive for future archive users, and how the UKGWA collection could begin to be unlocked for research and experimentation by approaching it as data (i.e. as a dataset at scale). The UKGWA has begun to examine independently the usefulness of modelling the hyperlinked structure of its collection for advanced corpus exploration; the aim of this collaboration is to test algorithms capable of searching for documents via the topics that they cover (e.g. 'climate change'), envisioning a future convergence of these two research frameworks. This is a diachronic corpus that is ideal for studying the emergence of topics and how they feature through government websites over time, and it will indicate engagement priorities and how these change over time.
