Efficiency investigation from shallow to deep neural network techniques in human activity recognition

In the last years, several researchers measured different recognition rates with different artificial neural network (ANN) techniques on public data sets in the human activity recognition (HAR) problem. However an overall investigation does not exist in the literature and the efficiency of complex and deeper ANNs over shallow networks is not clear. The purpose of this paper is to investigate the recognition rate and time requirement of different kinds of ANN approaches in HAR. This work examines the performance of shallow ANN architectures with different hyper-parameters, ANN ensembles, binary ANN classifier groups, and convolutional neural networks on two public databases. Although the popularity of binary classifiers, classifier ensembles and deep learning have been significantly increasing, this study shows that shallow ANNs with appropriate hyper-parameters in combination with extracted features can reach similar or higher recognition rate in less time than other artificial neural network methods in HAR. With a well-tuned ANN we outperformed all previous results on two public databases. Consequently, instead of the more complex ANN techniques, the usage of simple ANN with two or three layers can be an appropriate choice for activity recognition.     

Human rights,harm, and climate change mitigation

Abstract

A number of philosophers have resisted impersonal explanations of our obligation to mitigate climate change, and have developed accounts according to which these obligations are explained by human rights or harm-based considerations. In this paper I argue that several of these attempts to explain our mitigation obligations without appealing to impersonal factors fail, since they either cannot account for a plausibly robust obligation to mitigate, or have implausible implications in other cases. I conclude that despite the appeal of the motivations for rejecting the appeal to impersonal factors, such factors must play a prominent role in explaining our mitigation obligations.     

What’s wrong with human extinction?

Abstract

This paper explores what could be wrong with the fact of human extinction. I first present four reasons why we might consider human extinction to be wrong: (1) it would prevent millions of people from being born; (2) it would mean the loss of rational life and civilization; (3) it would cause existing people to suffer pain or death; (4) it would involve various psychological traumas. I argue that looking at the question from a contractualist perspective, only reasons (3) and (4) are admissible. I then consider what implications this limitation on reasons has for the wrongfulness of various forms of human extinction.     

Experience level influences the effect of attentional focus on sprint performance

Two experiments evaluated the influence of attentional focus on 10-meter sprint time and start kinetics in a group of collegiate soccer players and highly experienced sprinters. In Experiment 1, the collegiate soccer players were asked to perform 10-meter sprints under an external focus condition, an internal focus condition and a control condition. For the 10-meter sprint time, the results showed that both the external focus and control conditions resulted in significantly faster sprint times than the internal focus condition. There were no significant differences observed between the external focus and control conditions. There were also no significant differences observed across any of the conditions for a select set of kinetic variables. In Experiment 2, the highly experienced sprinters performed the same 10-meter sprint task using the same instructional conditions as in Experiment 1. For the 10-meter sprint time and kinetic variables, there were no significant differences observed across any of the conditions. These results provide new evidence that experience level mediates the influence of attentional focus on sprint performance.     

Audiovisual integration increases the intentional step synchronization of side-by-side walkers

When people walk side-by-side, they often synchronize their steps. To achieve this, individuals might cross-modally match audiovisual signals from the movements of the partner and kinesthetic, cutaneous, visual and auditory signals from their own movements. Because signals from different sensory systems are processed with noise and asynchronously, the challenge of the CNS is to derive the best estimate based on this conflicting information. This is currently thought to be done by a mechanism operating as a Maximum Likelihood Estimator (MLE). The present work investigated whether audiovisual signals from the partner are integrated according to MLE in order to synchronize steps during walking. Three experiments were conducted in which the sensory cues from a walking partner were virtually simulated. In Experiment 1 seven participants were instructed to synchronize with human-sized Point Light Walkers and/or footstep sounds. Results revealed highest synchronization performance with auditory and audiovisual cues. This was quantified by the time to achieve synchronization and by synchronization variability. However, this auditory dominance effect might have been due to artifacts of the setup. Therefore, in Experiment 2 human-sized virtual mannequins were implemented. Also, audiovisual stimuli were rendered in real-time and thus were synchronous and co-localized. All four participants synchronized best with audiovisual cues. For three of the four participants results point toward their optimal integration consistent with the MLE model. Experiment 3 yielded performance decrements for all three participants when the cues were incongruent. Overall, these findings suggest that individuals might optimally integrate audiovisual cues to synchronize steps during side-by-side walking.     

Chemosignalling effects of human tears revisited: Does exposure to female tears decrease males’ perception of female sexual attractiveness?

Gelstein et al. reported the results of three experiments suggesting a dampening influence of inhalation of female emotional tears on males’ arousal and perception of female sexual attractiveness, specifically in non-sexual situations. This prompted the hypothesis that crying exerts its influence on others not only via the auditory and visual mode but also via chemosignals. In three studies, we attempted to replicate and extend Gelstein et al.’s findings by including an additional condition with irritant tears, by using pictures of sexually attractive women, and by testing related hypotheses on the pro-social effects of exposure to tears. All three studies, separately or combined in a meta-analysis, failed to replicate the original inhibitory effects of tears. In addition, sniffing tears did not affect measures of connectedness, aggression and pro-social behaviour. It is concluded that the effects of female tears on male arousal and perception of female sexual attractiveness, if any, are very weak at best. Rather, it seems that crying exerts its strong inter-personal effects through the visual and auditory sensory channels.     

Improving Human‐Machine Cooperative Classification Via Cognitive Theories of Similarity

Acquiring perceptual expertise is slow and effortful. However, untrained novices can accurately make difficult classification decisions (e.g., skin‐lesion diagnosis) by reformulating the task as similarity judgment. Given a query image and a set of reference images, individuals are asked to select the best matching reference. When references are suitably chosen, the procedure yields an implicit classification of the query image. To optimize reference selection, we develop and evaluate a predictive model of similarity‐based choice. The model builds on existing psychological literature and accommodates stochastic, dynamic shifts of attention among visual feature dimensions. We perform a series of human experiments with two stimulus types (rectangles, faces) and nine classification tasks to validate the model and to demonstrate the model's potential to boost performance. Our system achieves high accuracy for participants who are naive as to the classification task, even when the classification task switches from trial to trial.     

Return of experimentally induced chocolate craving after extinction in a different context: divergence between craving for and expecting to eat chocolate

Unlike in fear conditioning, little attention has been devoted to extinction and renewal in appetitive conditioning, despite its relevance for extinction-based addiction treatments. We developed a paradigm, using a specific tray as a conditioned stimulus (CS) for eating chocolate (unconditioned stimulus, US), to investigate the effects of context change on acquisition and extinction of conditioned chocolate craving using an ABA renewal design. In Study 1 (n=32), participants successfully acquired chocolate craving, but unlike what is commonly observed in fear conditioning, craving did not extinguish. In Study 2, we separately assessed craving and US expectancy in a between-subjects design (n=64). US-expectancy data showed acquisition, extinction and renewal in the ABA group. The craving data did not follow this pattern, suggesting different mechanisms for craving and US expectancy. Similarities and differences between craving and US expectancy, as well as practical implications, are discussed.     

The Human Condition and the Gift: Towards a Theoretical Perspective on Close Relationships

Hannah Arendt’s exposition of the human condition provides the basic framework for a theoretical perspective on close relationships. According to Arendt, the human condition is comprised of three modes of activity: labor, work, and action. Labor is need-driven behavior, work concerns goal-directed activity and the fabrication of things, and action involves the mutual validation of unique individuals. Within this framework, the gift is the means by which relational ties are made concrete. I propose a model of gift-giving organized by two axes: whether or not the partner is singularized by the gift and whether or not the gift is given with an expectation of a return gift. I then apply this model to the three modes of the human condition.