University of Stirling

About: University of Stirling is a education organization based out in Stirling, Stirling, United Kingdom. It is known for research contribution in the topics: Population & Polyunsaturated fatty acid. The organization has 7722 authors who have published 20549 publications receiving 732940 citations. The organization is also known as: Stirling University.

Why 'What Works' Still Won't Work: From Evidence-Based Education to Value-Based Education

Abstract: The idea that professional practices such as education should be based upon or at least be informed by evidence continues to capture the imagination of many politicians, policy makers, practitioners and researchers. There is growing evidence of the influence of this line of thought. At the same time there is a growing body of work that has raised fundamental questions about the feasibility of the idea of evidence-based or evidence-informed practice. In this paper I make a further contribution to this discussion through an analysis of a number of assumptions that inform the discussion. I focus on the epistemological, ontological and praxeological dimensions of the discussion and in each domain identify a deficit. In the epistemological domain there is a knowledge deficit, in the ontological domain an effectiveness or efficacy deficit and in the practice domain an application deficit. Taken together these deficits not only raise some important questions about the very idea of evidence-based practice but also highlight the role of normativity, power and values. Against this background I outline the case for the idea of value-based education as an alternative for evidence-based education. As I am generally concerned about the expectations policy makers hold about what evidence can and should achieve in professional practices such as education, my contribution is primarily meant to provide educators and other professionals with arguments that can help them to resist unwarranted expectations about the role of evidence in their practices and even more so of unwarranted interventions in their practices.

Tensor Fusion Network for Multimodal Sentiment Analysis

Abstract: Multimodal sentiment analysis is an increasingly popular research area, which extends the conventional language-based definition of sentiment analysis to a multimodal setup where other relevant modalities accompany language. In this paper, we pose the problem of multimodal sentiment analysis as modeling intra-modality and inter-modality dynamics. We introduce a novel model, termed Tensor Fusion Networks, which learns both such dynamics end-to-end. The proposed approach is tailored for the volatile nature of spoken language in online videos as well as accompanying gestures and voice. In the experiments, our model outperforms state-of-the-art approaches for both multimodal and unimodal sentiment analysis.

Chemical toxins: a hypothesis to explain the global obesity epidemic.

Abstract: The number of obese people worldwide has escalated recently, revealing a complex picture of significant variations among nations and different profiles among adults and children, regions, and occupations. The commonly held causes of obesity - overeating and inactivity - do not explain the current obesity epidemic. There is evidence of a general decrease in food consumption by humans and a significant decline in their overall levels of physical activity. There is also more evidence to indicate that the body's natural weight-control mechanisms are not functioning properly in obesity. Because the obesity epidemic occurred relatively quickly, it has been suggested that environmental causes instead of genetic factors maybe largely responsible. What has, up to now, been overlooked is that the earth's environment has changed significantly during the last few decades because of the exponential production and usage of synthetic organic and inorganic chemicals. Many of these chemicals are better known for causing w...

When inverted faces are recognized: the role of configural information in face recognition.

Abstract: The identification of upright faces seems to involve a special sensitivity to “configural” information, the processing of which is less effective when the face is inverted. However the precise meaning of “configural” remains unclear. Five experiments are presented, which showed that the disruption of the processing of relational, rather than holistic, information largely determines the occurrence as well as the size of the face-inversion effect. In Experiment 1, faces could be identified either by unique combinations of local information (e.g. a specific eye colour plus hair colour) or by unique relational information (e.g. nose-mouth distance). The former showed no inversion effect, whereas the latter did. A combination of local and relational information (Experiment 2) again produced an inversion effect, although this effect was smaller than that found when only relational information was used. The results were replicated in Experiment 3 when differences in the brightness of local features were used ins...