Search

The AI and Society discourse has previously drawn attention to the ways that digital systems embody the values of the technology development community from which they emerge through the development and deployment process. Research shows how this effect leads to a particular treatment of gender in computer systems development, a treatment which lags far behind the rich understanding of gender that social studies scholarship reveals and people across society experience. Many people do not relate to the narrow binary gender options of male or female, and many people express their gender identity in much richer ways than the sex/gender binary female/woman and male/man Boolean terms will allow. We ask: are "born-digital" gendered datasets in digital systems experienced as marginalising by those who express their identity beyond the male/female binary? Case Study: Ireland. To answer this universal question, this paper presents the findings of an empirical case study of people in Ireland with diverse gender identities and expressions, and their experiences with public data systems and new technologies. In spite of great social changes in Ireland which have led to constitutional change in favour of LGBTQI + people, born-digital systems were experienced by respondents as embodying socio-cultural values which were no longer accepted in society at large. For many of the respondents, digital technologies routinely marginalise them in all kinds of ways. These systems keep alive violence and oppression long after civil rights have been enshrined in constitutional law. This study is just one example of the way assumptions about digital are disengaged from society-at-large. It is a call to arms to all who are passionate about socially-responsible technology. [ABSTRACT FROM AUTHOR]

Aim: The aim of this study was to: (1) use cognitive task analysis to describe final year nursing students situation awareness in recognising, responding and escalating care of deteriorating patients in ward settings; and (2) make recommendations for training and practice. Design: A mixed methods cognitive task analysis with a convergent triangulation design. Method: Data collection involved observations of 33 final year nursing students in simulated deteriorating patient scenarios and retrospective cognitive interviews. A process tracing technique was applied to identify the cues to deterioration participants perceived; how cue perception altered as situational demands increased; the extent that participants made connections between perceived cues and reached a situational understanding; and the factors that influenced and constrained participants situation awareness. Qualitative and quantitative findings are woven together and presented using descriptive statistics, illustrative quotations and timeline extractions. Results: The median cue perception was 65.4% and 57.6% in the medical and surgical scenarios, respectively. Perception was negatively influenced by incomplete vital sign monitoring as situations escalated; limited physical assessments; passive scanning behaviours; poor task automaticity; and excessive cognitive demands. Incomplete perception, poor cue integration and underdeveloped mental models influenced situational understanding. Escalation calls did not always accurately reflect situations and a reporting mindset was evident. Clinical exposure to deteriorating patients was described as variable and opportunistic. Reporting Method: The study is reported in accordance with the Good Reporting of a Mixed Methods Study (GRAMMS) checklist. Patient or Public Contribution: Patients and public were not involved in this research. [ABSTRACT FROM AUTHOR]

Inverse reinforcement learning (IRL) aims to estimate the reward function of optimizing agents by observing their response (estimates or actions). This paper considers IRL when noisy estimates of the gradient of a reward function generated by multiple stochastic gradient agents are observed. We present a generalized Langevin dynamics algorithm to estimate the reward function R(θ); specifically, the resulting Langevin algorithm asymptotically generates samples from the distribution proportional to exp(R(θ)). The proposed adaptive IRL algorithms use kernel-based passive learning schemes. We also construct multi-kernel passive Langevin algorithms for IRL which are suitable for high dimensional data and achieve variance reduction. The performance of the proposed IRL algorithms are illustrated on examples in adaptive Bayesian learning, logistic regression (high dimensional problem) and constrained Markov decision processes. We prove weak convergence of the proposed IRL algorithms using martingale averaging methods. We also analyze the tracking performance of the IRL algorithms in non-stationary environments where the utility function R(θ) has a hyper-parameter that jump changes over time as a slow Markov chain which is not known to the inverse learner. In this case, martingale averaging yields a Markov switched diffusion limit as the asymptotic behavior of the IRL algorithm. [ABSTRACT FROM AUTHOR]

Highlights • Co-simulation and building energy models provide a framework to test control algorithms. • Machine learning algorithms are effective to develop predictive optimisation models. • Predictive controls for residential can reduce heating systems costs up to 40%. • Smart predictive control in buildings results in lower carbon footprint up to 39%. • The smart grid dynamically controls dwellings to balance electricity supply/demand. Abstract This paper assesses the performance of control algorithms for the implementation of demand response strategies in the residential sector. A typical house, representing the most common building category in Ireland, was fully instrumented and utilised as a test-bed. A calibrated building simulation model was developed and used to assess the effectiveness of demand response strategies under different time-of-use electricity tariffs in conjunction with zone thermal control. Two demand response algorithms, one based on a rule-based approach, the other based on a predictive-based (machine learning) approach, were deployed for control of an integrated heat pump and thermal storage system. The two algorithms were evaluated using a common demand response price scheme. Compared to a baseline reference scenario, the following reductions were observed: electricity end-use expenditure (20.5% rule-based and 41.8% predictive algorithm), utility generation cost (18.8% rule-based and 39% predictive algorithm), carbon emissions (20.8% rule-based and 37.9% predictive algorithm). [ABSTRACT FROM AUTHOR]

Background the aim of this study was to retrospectively operationalise the World Guidelines for Falls Prevention and Management (WGFPM) falls risk stratification algorithm using data from The Irish Longitudinal Study on Ageing (TILDA). We described how easy the algorithm was to operationalise in TILDA and determined its utility in predicting falls in this population. Methods participants aged ≥50 years were stratified as 'low risk', 'intermediate' or 'high risk' as per WGFPM stratification based on their Wave 1 TILDA assessments. Groups were compared for number of falls, number of people who experienced one or more falls and number of people who experienced an injury when falling between Wave 1 and Wave 2 (approximately 2 years). Results 5,882 participants were included in the study; 4,521, 42 and 1,309 were classified as low, intermediate and high risk, respectively, and 10 participants could not be categorised due to missing data. At Wave 2, 17.4%, 43.8% and 40.5% of low-, intermediate- and high-risk groups reported having fallen, and 7.1%, 18.8% and 18.7%, respectively, reported having sustained an injury from falling. Conclusion the implementation of the WGFPM risk assessment algorithm was feasible in TILDA and successfully differentiated those at greater risk of falling. The high number of participants classified in the low-risk group and lack of differences between the intermediate and high-risk groups may be related to the non-clinical nature of the TILDA sample, and further study in other samples is warranted. [ABSTRACT FROM AUTHOR]

Background Physical activity (PA) interventions capable of producing health benefits cost effectively are a public health priority across the Western world. 'Men on the Move' (MOM), a community-based PA intervention for men, demonstrated significant health benefits up to 52-weeks (W) post-baseline. This article details the economic evaluation of MOM with a view to determining its cost-effectiveness as a public health intervention to be rolled out nationally in Ireland. Methods Cost-effectiveness was determined by comparing the costs (direct and indirect) of the programme to its benefits, which were captured as the impact on quality-adjusted life-years (QALYs). For the benefits, cost–utility analysis was conducted by retrospectively adapting various health-related measures of participants to generate health states using Brazier et al.'s (2002) short form-6D algorithm. This in turn allowed for 'utility measures' to be generated, from which QALYs were derived. Results Findings show MOM to be cost-effective in supporting an 'at risk' cohort of men achieves significant improvements in aerobic fitness, weight loss and waist reduction. The total cost per participant (€125.82 for each of the 501 intervention participants), the QALYs gained (11.98 post-12-W intervention, or 5.3% health improvement per participant) and estimated QALYs ratio costs of €3723 represents a cost-effective improvement when compared to known QALY guidelines. Conclusions The analysis shows that the cost per QALY achieved by MOM is significantly less than the existing benchmarks of £20 000 and €45 000 in the UK and Ireland respectively, demonstrating MOM to be cost-effective. [ABSTRACT FROM AUTHOR]

We examined recent atmospheric mercury concentrations measured with a high temporal resolution of 15 min at Mace Head, a GAW station on the western coast of Ireland. We attributed a direct contribution of 34 % (0.44 ngm-3) to primary sources. Additionally, a steep decline (0.05 ngyr-1) in mercury concentrations was observed between 2013 and 2018. Using a stereo algorithm we reconstructed 99.9 % of the atmospheric mercury. A conservative analysis demonstrated no decrease in total gaseous mercury (TGM) associated with atmospheric species typically used as tracers for oceanic emissions. The results show that the atmospheric mercury mass is mainly loaded in a baseline factor with an ongoing decline. Moreover, we exploit temporal variation and wind pattern effects in the measured atmospheric species; the results show that the diurnal variation and seasonality in TGM observed in Mace Head are closely related to other species linked to primary sources and can be explained by transport from continental areas. [ABSTRACT FROM AUTHOR]