Incorporating Demographic Embeddings Into Language Understanding

Meaning depends on context. This applies in obvious cases like deictics or sarcasm as well as more subtle situations like framing or persuasion. One key aspect of this is the identity of the participants in an interaction. Our interpretation of an utterance shifts based on a variety of factors, including personal history, background knowledge, and our relationship to the source. While obviously an incomplete model of individual differences, demographic factors provide a useful starting point and allow us to capture some of this variance. However, the relevance of specific demographic factors varies between situations—where age might be the key factor in one context, ideology might dominate in another. To address this challenge, we introduce a method for combining demographics and context into situated demographic embeddings—mapping representations into a continuous geometric space appropriate for the given domain, showing the resulting representations to be functional and interpretable. We further demonstrate how to make use of related external data so as to apply this approach in low‐resource situations. Finally, we show how these representations can be incorporated into improve modeling of real‐world natural language understanding tasks, improving model performance and helping with issues of data sparsity.     

Utilizing physiological measures to facilitate phobia treatment with individuals with autism and intellectual disability: A case study

Despite the conceptualization of phobic fear as a multi‐factorial response, the measurement and treatment of phobias is often not comprehensive. The extant literature on the treatment of phobias for individuals with intellectual disabilities has exclusively focused on indices of behavioral avoidance. The present study demonstrates the treatment of behavioral avoidance and physiological reactivity in a child diagnosed with autism, intellectual disability, and a dog phobia. The results indicate that the individual often experienced physiological reactivity in the presence of the feared stimulus despite engaging in behavioral approach. Treatment of both behavioral avoidance and physiological reactivity resulted in successful generalization of treatment effects across three additional dogs and maintenance of treatment effects during a six‐month follow‐up assessment. Implications for examining multiple aspects of the anxious response when treating children with intellectual disabilities and severe language delays are discussed. Copyright © 2010 John Wiley & Sons, Ltd.     

Perception and action influences on discrete and reciprocal bimanual coordination

For nearly four decades bimanual coordination, “a prototype of complex motor skills” and apparent “window into the design of the brain,” has been intensively studied. Past research has focused on describing and modeling the constraints that allow the production of some coordination patterns while limiting effective performance of other bimanual coordination patterns. More recently researchers have identified a coalition of perception-action constraints that hinder the effective production of bimanual skills. The result has been that given specially designed contexts where one or more of these constraints are minimized, bimanual skills once thought difficult, if not impossible, to effectively produce without very extensive practice can be executed effectively with little or no practice. The challenge is to understand how these contextual constraints interact to allow or inhibit the production of complex bimanual coordination skills. In addition, the factors affecting the stability of bimanual coordination tasks needs to be re-conceptualized in terms of perception-related constraints arising from the environmental context in which performance is conducted and action constraints resident in the neuromotor system.     

The Effects of Intermittent Illumination on a Visual Inspection Task

Two experiments are described in which eye movements were monitored as subjects performed a simple target-spotting task under conditions of intermittent illumination produced by varying the display-screen frame rate on a computer VDU. In Experiment 1, subjects executed a saccade from a fixation point to a target which appeared randomly at a fixed eccentricity of 14 character positions to the left or right. Saccade latency did not differ reliably as a function of screen refresh rate, but average saccade extent at 70 Hz and 110 Hz was reliably shorter than at 90 Hz and 100 Hz. Experiment 2 examined the same task using a range of target eccentricities (7, 14, and 28 character positions to the left and right) and across a wider range of screen refresh rates. The results confirmed the curvilinear relationship obtained in Experiment 1, with average saccade extent reliably shorter at refresh rates of 50 Hz and 125 Hz than at 75 Hz and 100 Hz. While the effect was greater for remote targets, analyses of the proportional target error failed to show a reliable interaction between target eccentricity and display refresh rate. In contrast to Experiment 1, there was a pronounced effect of refresh rate on saccade latency (corrected for time to write the screen frame), with shorter latencies at higher refresh rates. It may be concluded that pulsation at frequencies above fusion disrupts saccade control. However, the curvilinear functional relationship between screen refresh rate and saccade extent obtained in these studies differs from previously reported effects of intermittent illumination on the average size of 'entry saccades' (the first saccade to enter a given word) in a task involving word identification (Kennedy & Murray, 1993a, 1996). This conflict of data may arise in part because withinword adjustments in viewing position, which are typical of normal reading, influence measures of average saccade extent.     

The relationship between infanticide and pregnancy block in mice

Newly pregnant mice are more likely to exhibit pregnancy block when females are exposed to infanticidal males rather than to noninfanticidal males. These data suggest that females are able to assess the risk to their future infants and adjust their reproductive tactic. Those females that did not exhibit pregnancy block in the presence of infanticidal males suffered a greater loss of pups than those in the presence of noninfanticidal males, demonstrating the advantage of pregnancy block. Females served by infanticidal or noninfanticidal males showed no difference in pregnancy block, but if the litter was produced, there was a greater loss of pups sired by infanticidal males.