A Principled Approach to Feature Selection in Models of Sentence Processing

Among theories of human language comprehension, cue-based memory retrieval has proven to be a useful framework for understanding when and how processing difficulty arises in the resolution of long-distance dependencies. Most previous work in this area has assumed that very general retrieval cues like [+subject] or [+singular] do the work of identifying (and sometimes misidentifying) a retrieval target in order to establish a dependency between words. However, recent work suggests that general, handpicked retrieval cues like these may not be enough to explain illusions of plausibility (Cunnings & Sturt, 2018), which can arise in sentences like The letter next to the porcelain plate shattered. Capturing such retrieval interference effects requires lexically specific features and retrieval cues, but handpicking the features is hard to do in a principled way and greatly increases modeler degrees of freedom. To remedy this, we use well-established word embedding methods for creating distributed lexical feature representations that encode information relevant for retrieval using distributed retrieval cue vectors. We show that the similarity between the feature and cue vectors (a measure of plausibility) predicts total reading times in Cunnings and Sturt’s eye-tracking data. The features can easily be plugged into existing parsing models (including cue-based retrieval and self-organized parsing), putting very different models on more equal footing and facilitating future quantitative comparisons.     

Effects of short-term fatigue on biomechanical and physiological aspects of double poling in high-level cross-country skiers

The study aim was to evaluate biomechanical and physiological alterations in double poling technique (DP) after a short-term fatiguing exercise. Eight high-level skiers performed a sub-maximal DP trial (20 km h⁻¹, 1°) before (PRE) and after (POST) a DP test to exhaustion while roller skiing on a treadmill. An integrated analysis of DP technique during PRE and POST included measurement of pole, joint, and centre of mass (COM) kinematics, poling forces, cycle timing, and metabolic parameters. Muscle fatigue in three upper-body muscles was assessed by calculating the Dimitrov’ fatigue index (FI_nms5) of specific electromyographic segments. FI_nms5 tended to increase in the latissimus dorsi and teres major muscles (P = 0.023 and P = 0.030, respectively) across consecutive DP cycles, as did blood lactate concentration (P = 0.001) and rating of perceived exertion (P = 0.005). The changes indicated a state of fatigue during POST and coincided with the reduction in poling force exertion capacity (P = 0.020). Pole, joint and COM kinematics did not differ between PRE and POST (P > 0.050), whereas recovery phase and cycle times were shorter at POST (P < 0.001 and P = 0.001, respectively). Short-term fatigue led to a reduction in poling force exertion capacity and cycle time in high-level skiers, without altering body and pole kinematics.     

Social Capital and Community Group Participation: Examining ‘Bridging’ and ‘Bonding’ in the Context of a Healthy Living Centre in the UK

Social capital has been widely advocated as a way of understanding and building community participation in the interest of health improvement. However, the concept as proposed by Putnam, has been criticised for presenting an overly romanticised account of complex community relations. This paper presents analysis from a qualitative evaluation of a Healthy Living Centre (HLC) in the North of England, to examine the utility of the concept of social capital in this context. We found the concepts of ‘bridging’ and ‘bonding’ social capital were useful – though not without limitations – in helping to make sense of the complexities and contradictions in participants’ experiences of community group participation. 'Bridging' helped provide an understanding of how the decline in shared social spaces such as local shops impacts on social relationships. 'Bonding' highlighted how community group membership can have positive and negative implications for individuals and the wider community. It was found that skilled group leadership was key to strengthening bridging capital. Politically, in the UK, community participation is seen as having an essential role in social change, for example, its centrality to the coalition government's idea of the ‘Big Society’. A micro‐examination of this HLC using the lens of social capital provides a valuable critical insight into community participation. It shows that this kind of initiative can be successful in building social capital, given conditions such as an appropriate setting and effective leadership. However, they cannot substitute for other kinds of investment in the physical infrastructure of a community. Copyright © 2012 John Wiley & Sons, Ltd.     

A visual embedding for the unsupervised extraction of abstract semantics

Vector-space word representations obtained from neural network models have been shown to enable semantic operations based on vector arithmetic. In this paper, we explore the existence of similar information on vector representations of images. For that purpose we define a methodology to obtain large, sparse vector representations of image classes, and generate vectors through the state-of-the-art deep learning architecture GoogLeNet for 20 K images obtained from ImageNet. We first evaluate the resultant vector-space semantics through its correlation with WordNet distances, and find vector distances to be strongly correlated with linguistic semantics. We then explore the location of images within the vector space, finding elements close in WordNet to be clustered together, regardless of significant visual variances (e.g., 118 dog types). More surprisingly, we find that the space unsupervisedly separates complex classes without prior knowledge (e.g., living things). Afterwards, we consider vector arithmetics. Although we are unable to obtain meaningful results on this regard, we discuss the various problem we encountered, and how we consider to solve them. Finally, we discuss the impact of our research for cognitive systems, focusing on the role of the architecture being used.     

Identical Twins,Deduction Theorems,and Pattern Functions: Exploring the Implicative BCSK Fragment of S5

We recapitulate (Section 1) some basic details of the system of implicative BCSK logic, which has two primitive binary implicational connectives, and which can be viewed as a certain fragment of the modal logic S5. From this modal perspective we review (Section 2) some results according to which the pure sublogic in either of these connectives (i.e., each considered without the other) is an exact replica of the material implication fragment of classical propositional logic. In Sections 3 and 5 we show that for the pure logic of one of these implicational connectives two – in general distinct – consequence relations (global and local) definable in the Kripke semantics for modal logic turn out to coincide, though this is not so for the pure logic of the other connective, and that there is an intimate relation between formulas constructed by means of the former connective and the local consequence relation. (Corollary 5.8. This, as we show in an Appendix, is connected to the fact that the ‘propositional operations’ associated with both of our implicational connectives are close to being what R. Quackenbush has called pattern functions.) Between these discussions Section 4 examines some of the replacement-of-equivalents properties of the two connectives, relative to these consequence relations, and Section 6 closes with some observations about the metaphor of identical twins as applied to such pairs of connectives.     

A comparison of deaf, normal, and brain damaged adults on a tachistoscopic task

Printed words and letters, pictures of signed words, and pictures of finger-spelled letters were presented tachistoscopically to deaf, normal, and brain-damaged adults. The findings suggest that (1) the deaf show no cerebral asymmetry for any of the material but found identification of the stimuli very difficult unless a forced-choice recognition procedure was used or the exposure time was very long; (2) both the normal and the brain-damaged groups show left hemisphere superiority for printed material but no cerebral specialization for signed material; (3) the left-right hemisphere difference is more marked in the brain-damaged group regardless of which hemisphere showed most damage. This may be explained, at least in some cases, by impaired transfer across the corpus callosum.     

Boredom: Practical consequences and a theory

Boredom is defined as a unique psychophysiological state possessing interrelated and inseparable emotional, motivational, perceptual and cognitive concomitants. Practical consequences of boredom are reviewed, including diminished performance efficiency, general life satisfaction and health. Finally, the outline of a theoretical model is presented.