The effect of increased cognitive processing on reactive balance control following perturbations to the upper limb.

Reactive balance control following hand perturbations is important for everyday living as humans constantly encounter perturbations to the upper limb while performing functional tasks while standing. When multiple tasks are performed simultaneously, cognitive processing is increased, and performance on at least one of the tasks is often disrupted, owing to attentional resources being divided. The purpose here was to assess the effects of increased cognitive processing on whole-body balance responses to perturbations of the hand during continuous voluntary reaching. Sixteen participants (8 females; 22.9 ± 4.5 years) stood and grasped the handle of a KINARM - a robotic-controlled manipulandum paired with an augmented visual display. Participants completed 10 total trials of 100 mediolateral arm movements at a consistent speed of one reach per second, and an auditory n-back task (cognitive task). Twenty anteroposterior hand perturbations were interspersed randomly throughout the reaching trials. The arm movements with random arm perturbations were either performed simultaneously with the cognitive task (combined task) or in isolation (arm perturbation task). Peak centre of pressure (COP) displacement and velocity, time to COP displacement onset and peak, as well as hand displacement and velocity following the hand perturbation were evaluated. N-back response times were 8% slower and 11% less accurate for the combined than the cognitive task. Peak COP displacement following posterior perturbations increased by 8% during the combined compared to the arm perturbation task alone, with no other differences detected. Hand peak displacement decreased by 5% during the combined compared to the arm perturbation task. The main findings indicate that with increased cognitive processing, attentional resources were allocated from the cognitive task towards upper limb movements, while attentional resources for balance seemed unaltered.

Scalable Variational Gaussian Processes for Crowdsourcing: Glitch Detection in LIGO.

In the last years, crowdsourcing is transforming the way classification training sets are obtained. Instead of relying on a single expert annotator, crowdsourcing shares the labelling effort among a large number of collaborators. For instance, this is being applied in the laureate laser interferometer gravitational waves observatory (LIGO), in order to detect glitches which might hinder the identification of true gravitational-waves. The crowdsourcing scenario poses new challenging difficulties, as it has to deal with different opinions from a heterogeneous group of annotators with unknown degrees of expertise. Probabilistic methods, such as Gaussian processes (GP), have proven successful in modeling this setting. However, GPs do not scale up well to large data sets, which hampers their broad adoption in real-world problems (in particular LIGO). This has led to the very recent introduction of deep learning based crowdsourcing methods, which have become the state-of-the-art for this type of problems. However, the accurate uncertainty quantification provided by GPs has been partially sacrificed. This is an important aspect for astrophysicists in LIGO, since a glitch detection system should provide very accurate probability distributions of its predictions. In this work, we first leverage a standard sparse GP approximation (SVGP) to develop a GP-based crowdsourcing method that factorizes into mini-batches. This makes it able to cope with previously-prohibitive data sets. This first approach, which we refer to as scalable variational Gaussian processes for crowdsourcing (SVGPCR), brings back GP-based methods to a state-of-the-art level, and excels at uncertainty quantification. SVGPCR is shown to outperform deep learning based methods and previous probabilistic ones when applied to the LIGO data. Its behavior and main properties are carefully analyzed in a controlled experiment based on the MNIST data set. Moreover, recent GP inference techniques are also adapted to crowdsourcing and evaluated experimentally.