Search

Forty million people are estimated to be in some form of modern slavery across the globe. Understanding the factors that make any particular individual or geographical region vulnerable to such abuse is essential for the development of effective interventions and policy. Efforts to isolate and assess the importance of individual drivers statistically are impeded by two key challenges: data scarcity and high dimensionality, typical of many “wicked problems”. The hidden nature of modern slavery restricts available data points; and the large number of candidate variables that are potentially predictive of slavery inflate the feature space exponentially. The result is a “small n, large p” setting, where overfitting and significant inter-correlation of explanatory variables can render more traditional statistical approaches problematic. Recent advances in non-parametric computational methods, however, offer scope to overcome such challenges and better capture the complex nature of modern slavery. We present an approach that combines non-linear machine-learning models and strict cross-validation methods with novel variable importance techniques, emphasising the importance of stability of model explanations via a Rashomon-set analysis. This approach is used to model the prevalence of slavery in 48 countries, with results bringing to light the importance of new predictive factors—such as a country’s capacity to protect the physical security of women, which has been previously under-emphasised in quantitative models. Further analyses uncover that women are particularly vulnerable to exploitation in areas where there is poor access to resources. Our model was then leveraged to produce new out-of-sample estimates of slavery prevalence for countries where no survey data currently exists.

Understanding what factors predict whether an urban migrant will end up in a deprived neighbourhood or not could help prevent the exploitation of vulnerable individuals. This study leveraged pseudonymized mobile money interactions combined with cell phone data to shed light on urban migration patterns and deprivation in Tanzania. Call detail records were used to identify individuals who migrated to Dar es Salaam, Tanzania’s largest city. A street survey of the city’s subwards was used to determine which individuals moved to more deprived areas. t-tests showed that people who settled in poorer neighbourhoods had less money coming into their mobile money account after they moved, but not before. Amachine learning approach was then utilized to predict which migrants will move to poorer areas of the city, making them arguably more vulnerable to poverty, unemployment and exploitation. Features indicating the strength and location of people’s social connections in Dar es Salaam before they moved (‘pull factors’) were found to be most predictive, more so than traditional ‘push factors’ such as proxies for poverty in the migrant’s source region.

Additional file 1: Appendix S1. Interview Guide.

Additional file 2: Appendix S2. Table S1. Participant characteristics.

On any given day, tens of millions of people find themselves trapped in instances of modern slavery. The terms "human trafficking," "trafficking in persons," and "modern slavery" are sometimes used interchangeably to refer to both sex trafficking and forced labor. Human trafficking occurs when a trafficker compels someone to provide labor or services through the use of force, fraud, and/or coercion. The wide range of stakeholders in human trafficking presents major challenges. Direct stakeholders are law enforcement, NGOs and INGOs, businesses, local/planning government authorities, and survivors. Viewed from a very high level, all stakeholders share in a rich network of interactions that produce and consume enormous amounts of information. The problems of making efficient use of such information for the purposes of fighting trafficking while at the same time adhering to community standards of privacy and ethics are formidable. At the same time they help us, technologies that increase surveillance of populations can also undermine basic human rights. In early March 2020, the Computing Community Consortium (CCC), in collaboration with the Code 8.7 Initiative, brought together over fifty members of the computing research community along with anti-slavery practitioners and survivors to lay out a research roadmap. The primary goal was to explore ways in which long-range research in artificial intelligence (AI) could be applied to the fight against human trafficking. Building on the kickoff Code 8.7 conference held at the headquarters of the United Nations in February 2019, the focus for this workshop was to link the ambitious goals outlined in the A 20-Year Community Roadmap for Artificial Intelligence Research in the US (AI Roadmap) to challenges vital in achieving the UN's Sustainable Development Goal Target 8.7, the elimination of modern slavery., Comment: A Computing Community Consortium (CCC) workshop report, 24 pages

Not only is the crime of modern slavery a hidden one, but global definitions also vary, as do cultural norms on issues such as child marriage, child labour and forced marriage. In this nuanced context, this research goes beyond the traditional regression analyses used in this space, to try to understand not just the linear predictive relationships, but also the non-linear effects, thresholds, interaction effects, and potential subgroups of countries. Consequently, this research, in the context of small data (N=70) and many potential (and correlated) predictors (107), utilises machine learning methods with the goal to better predict and explain countries' prevalence of MDS, with all its complexities. Data is collated and scraped from multiple open sources including the World Bank, Varieties of Democracy Project (Coppedge et al., 2015), The Woman Stats Project, and the Global Slavery Index, which includes a measure of slavery prevalence measured by the Gallup World Poll. Our methodology investigates the utility of theory driven feature selection and data driven feature compression, followed by the evaluation of different prediction models. The approach that gave the most accurate out of sample predictions used NMF to reduce the feature space, and a decision tree to predict MDS prevalence. Stability of the model over different random seeds was then assessed, and a Rashomon set (other reasonably accurate predictive models, which may contain a simpler more interpretable model) identified (Rudin, 2019). Within the Rashomon set, the models’ predictions and their strategies were analysed to identify: 1) whether or not the models gave diverging predictions, or whether predictions were similar enough that a simpler, more interpretable, model could be substituted in; and 2) similarities between the models that could be used to provide more stable insights.

This thesis explored the impact of digital learning settings, and their associated social practices, on an individual’s historical thinking (Wineburg, 1991). The rationale was underpinned by the assumption that information transfer between print and digital mediums is not a neutral process, and that digital learning tools serve as carriers of patterns of knowledge known as Digital Epistemologies (Lankshear & Knoble, 2006), which stand in marked contrast to the critical understandings associated with historical thinking, providing theoretical warrant for an investigation into its potential impact. Data was generated via qualitative interviews where novice (student) and expert (historian) participants were asked to evaluate historical content embedded within different types of websites. A discursive analysis was undertaken using a neo-Vygotskian conceptual framework (Wertsch, 1991) focussing on social languages, the mediational tool kit, and privileging practices. Using a Wertschian framework, it was theorised that participants would have at least two sets of evaluative criteria available to them (formal and informal), and that these would exist in a state of epistemic tension during participant evaluation. It was found that when evaluating historical content online, participants appeared to be using a hybridised adaptation of the historical thinking heuristics identified by Wineburg (1991) and the informal digital heuristics identified by Metzger, Flanagin and Medders (2010), and that these two ‘evaluative discourses’ existed in tension, producing multiple, conflicting notions of website credibility. This tension appeared to impede participants’ critical historical thinking, and an association was identified between discursive tension and school-based understandings of website credibility and evaluation. A second aim was to identify key features of expert practices, and it was found that the historians engaged in a process termed ‘formalisation’, where informal web-based heuristics were adapted for critical disciplinary purposes. The paper will conclude by proposing models of expert practice and historical thinking online, and outlining a framework by which student evaluation might be formalised to better reflect the critical practices of experts.

Advances in sampling schemes for Markov jump processes have recently enabled multiple inferential tasks. However, in statistical and machine learning applications, we often require that these continuous-time models find support on structured and infinite state spaces. In these cases, exact sampling may only be achieved by often inefficient particle filtering procedures, and rapidly augmenting observed datasets remains a significant challenge. Here, we build on the principles of uniformization and present a tractable framework to address this problem, which greatly improves the efficiency of existing state-of-the-art methods commonly used in small finite-state systems, and further scales their use to infinite-state scenarios. We capitalize on the marginal role of variable subsets in a model hierarchy during the process jumps, and describe an algorithm that relies on measurable mappings between pairs of states and carefully designed sets of synthetic jump observations. The proposed method enables the efficient integration of slice sampling techniques and it can overcome the existing computational bottleneck. We offer evidence by means of experiments addressing inference and clustering tasks on both simulated and real data sets.

The characterisation of time-series data via their most salient features is extremely important in a range of machine learning task, not least of all with regards to classification and clustering. While there exist many feature extraction techniques suitable for non-intermittent time-series data, these approaches are not always appropriate for intermittent time-series data, where intermittency is characterized by constant values for large periods of time punctuated by sharp and transient increases or decreases in value. Motivated by this, we present aggregation, mode decomposition and projection (AMP) a feature extraction technique particularly suited to intermittent time-series data which contain time-frequency patterns. For our method all individual time-series within a set are combined to form a non-intermittent aggregate. This is decomposed into a set of components which represent the intrinsic time-frequency signals within the data set. Individual time-series can then be fit to these components to obtain a set of numerical features that represent their intrinsic time-frequency patterns. To demonstrate the effectiveness of AMP, we evaluate against the real word task of clustering intermittent time-series data. Using synthetically generated data we show that a clustering approach which uses the features derived from AMP significantly outperforms traditional clustering methods. Our technique is further exemplified on a real world data set where AMP can be used to discover groupings of individuals which correspond to real world sub-populations., Comment: Paper accepted in workshop on mining and learning from time series (MiLeTS) to be held in conjunction with KDD 2015 August 10- 13 2015, Sydney, Australia

In this paper we present AnswerTree, a collaborative mobile location-based educational game designed to teach 8-12 year olds about trees and wildlife within the University of Nottingham campus. The activity is designed around collecting virtual cards (similar in nature to the popular Top TrumpsTM games) containing graphics and information about notable trees. Each player begins by collecting one card from a game location, but then he or she can only collect further cards by answering questions – whose solutions are obtainable through sharing knowledge with other cardholders. This ostensibly allows each player to become a subject expert at the start of the game, encouraging collaborative interaction for the game to be successfully completed. In this initial paper we will outline the structure and background of this location based game. AnswerTree has been authored within the Hyperplace framework, and is a first implementation of a wider process to develop a flexible, multi-purpose platform for both individual and group location-based mobile learning.

A collection of folklore and local history stories from Cnoc an Teampoill (B.), Ráth Luirc (school) (Freemount, Co. Cork), collected as part of the Schools' Folklore Scheme, 1937-1938 under the supervision of teacher Domhnall Breathnach., Hidden Treasure / Nunan, Thomas -- Funny Story / Mullane, Denis -- Collection of Riddles / Nunan, Joseph -- Weather-Lore -- Local Happenings / Goulding, James -- Local Heroes / Nunan, Thomas -- Severe Weather / Mullane, Denis -- Old Schools / Nunan, Thomas -- Old Crafts / Nunan, Thomas -- Local Marriage Customs / Moreton, Edmond -- In the Penal Times / Goulding, James -- Local Place Names / Moreton, Edmond -- Local Cures / Nunan, Thomas -- Homemade Toys / Moreton, Edmond -- Lore of Certain Days / Rahilly, Denis -- Travelling Folk / Shaughnessy, William O' -- Famine Times / Nunan, James -- Games I Play / Conner, Denis O' -- Local Roads / Nunan, Joseph -- My Home District / Conner, Denis O' -- Our Holy Wells / Goulding, James -- Herbs / Connor, Denis O' -- Potato Crop / Nunan, Joseph -- Proverbs / Conner, Denis O' -- Festival Customs / Nunan, Joseph -- Care of Our Farm Animals / Goulding, James -- Churning / Nunan, Thomas -- Care of the Feet / Conner, Denis O' -- Local Forge / Lane, Jeremiah -- Clothes Made Locally / Conner, Denis O' -- Local Patron Saint / Nunan, Joseph -- Local Fairs / Nunan, Thomas -- Landlord / Goulding, James -- Food in Olden Times / Kavanagh, Eugene -- Hurling Match / Connor, Denis O' -- Old Story / Nunan, Thomas -- Old Irish Tales / Goulding, James -- Song / Kavanagh, Eugene -- Local Monuments / Nunan, Thomas -- Bread / Nunan, Joseph -- Buying and Selling / Nunan, James -- Old Houses / Nunan, Joseph -- Local Ruins -- Old Churchyards / Nunan, Thomas, Supported by funding from the Department of Arts, Heritage and the Gaeltacht (Ireland), University College Dublin, and the National Folklore Foundation (Fondúireacht Bhéaloideas Éireann), 2014-2016.