Search

Chemical engineering research and design 187, 164-173 (2022). doi:10.1016/j.cherd.2022.08.061 special issue: "100 years of IChemE", Published by Elsevier, Amsterdam

This study evaluated the effects of ischemic conditioning on power output and bar velocity in the bench press exercise. Ten healthy males (age: 25 ± 2 years; body mass: 92 ± 8 kg; bench press one repetition maximum −1RM: 145 ± 13 kg), took part in two experimental sessions (with and without ischemia), 1 week apart in random and counterbalanced order. In the ischemic condition, cuffs placed around the upper part of the arms were inflated to 80% of arterial occlusion pressure before each set, while in the control condition there was no blood flow restriction. The exercise protocol included 5 sets of three repetitions each, against a resistance equal to 60% 1RM, with 5 min recovery intervals between sets. There was a main effect of condition for mean power output (MP) and mean bar velocity (MV) (p = 0.01), with overall MP being higher in ischemia than in control by 5.6 ± 4.1% (mean ± 90% compatibility limits), a standardized effect size (ES) of 0.51. Overall MV was also higher by 5.5 ± 4.0%, ES = 0.63. Peak power output (PP) and peak bar velocity (PV) were similar in set 1 of the control and ischemia condition (1039 ± 105 vs. 1054 ± 82 W; 684 ± 74 vs. 696 ± 53 W; 1.09 ± 0.07 vs. 1.12 ± 0.09 m/s; 0.81 ± 0.05 vs. 0.82 ± 0.05 m/s, p = 0.67 to 0.99, mean ± standard deviation). However, from set 3 onward (p = 0.03 to 0.001), PP and PV were higher in ischemia compared with control, with the highest difference observed in set 5 (10.9 ± 5.9%, ES = 0.73 for PP and 8.6 ± 4.6%; ES = 0.89 for PV). These results indicate that ischemia used before each set of the bench press exercise increases power output and bar velocity and this may be used as performance-enhancing stimulus during explosive resistance training. © Copyright © 2021 Wilk, Krzysztofik, Jarosz, Krol, Leznicka, Zajac, Stastny and Bogdanis.

This study examined changes in movement velocity and surface electromyographic (sEMG) activity of the pectoralis major (PM) and triceps brachii (TB) muscles during the bench press exercise to failure against different loads. Fourteen men performed a set to failure with maximum intended velocity, against low (40%-1 repetition maximum-RM), moderate (60%-1RM), and heavy loads (80%-1RM). Number of repetitions, volume load, mean and peak velocity, and total time increased with decreasing load (40% > 60% > 80%, p

Gaussian processes are distributions over functions that are versatile and mathematically convenient priors in Bayesian modelling. However, their use is often impeded for data with large numbers of observations, $N$, due to the cubic (in $N$) cost of matrix operations used in exact inference. Many solutions have been proposed that rely on $M \ll N$ inducing variables to form an approximation at a cost of $\mathcal{O}(NM^2)$. While the computational cost appears linear in $N$, the true complexity depends on how $M$ must scale with $N$ to ensure a certain quality of the approximation. In this work, we investigate upper and lower bounds on how $M$ needs to grow with $N$ to ensure high quality approximations. We show that we can make the KL-divergence between the approximate model and the exact posterior arbitrarily small for a Gaussian-noise regression model with $M\ll N$. Specifically, for the popular squared exponential kernel and $D$-dimensional Gaussian distributed covariates, $M=\mathcal{O}((\log N)^D)$ suffice and a method with an overall computational cost of $\mathcal{O}(N(\log N)^{2D}(\log\log N)^2)$ can be used to perform inference., Comment: Extended version of http://proceedings.mlr.press/v97/burt19a.html (arxiv version: arXiv:1903.03571 ). Published in Journal of Machine Learning Research: http://jmlr.org/papers/v21/19-1015.html. Code available at: https://github.com/markvdw/RobustGP

This study evaluated the effects of continuous and intermittent blood flow restriction (BFR) with 70% of full arterial occlusion pressure on bar velocity during the bench press exercise against a wide range of resistive loads. Eleven strength-trained males (age: 23.5 ± 1.4 years; resistance training experience: 2.8 ± 0.8 years, maximal bench press strength – 1RM = 101.8 ± 13.9 kg; body mass = 79.8 ± 10.4 kg), performed three different testing protocols in random and counterbalanced order: without BFR (NO-BFR); intermittent BFR (I-BFR) and continuous BFR (C-BFR). During each experimental session, subjects performed eight sets of two repetitions each, with increasing loads from 20 to 90% 1RM (10% steps), and 3 min rest between each set. In the C-BFR condition occlusion was kept throughout the trial, while in the I-BFR, occlusion was released during each 3 min rest interval. Peak bar velocity (PV) during the bench press exercise was higher by 12–17% in both I-BFR and C-BFR compared with NO-BFR only at the loads of 20, 30, 40, and 50% 1RM (p < 0.001), while performance at higher loads remained unchanged. Mean bar velocity (MV) was unaffected by occlusion (p = 0.342). These results indicate that BFR during bench press exercise increases PV and this may be used as an enhanced stimulus during explosive resistance training. At higher workloads, bench press performance was not negatively affected by BFR, indicating that the benefits of exercise under occlusion can be obtained while explosive performance is not impaired. © Copyright © 2020 Wilk, Gepfert, Krzysztofik, Stastny, Zajac and Bogdanis.

We identify a new variational inference scheme for dynamical systems whose transition function is modelled by a Gaussian process. Inference in this setting has either employed computationally intensive MCMC methods, or relied on factorisations of the variational posterior. As we demonstrate in our experiments, the factorisation between latent system states and transition function can lead to a miscalibrated posterior and to learning unnecessarily large noise terms. We eliminate this factorisation by explicitly modelling the dependence between state trajectories and the Gaussian process posterior. Samples of the latent states can then be tractably generated by conditioning on this representation. The method we obtain (VCDT: variationally coupled dynamics and trajectories) gives better predictive performance and more calibrated estimates of the transition function, yet maintains the same time and space complexities as mean-field methods. Code is available at: github.com/ialong/GPt., Comment: 10 pages, 4 figures, 3 tables. Published in the proceedings of the Thirty-sixth International Conference on Machine Learning (ICML), 2019

Generalising well in supervised learning tasks relies on correctly extrapolating the training data to a large region of the input space. One way to achieve this is to constrain the predictions to be invariant to transformations on the input that are known to be irrelevant (e.g. translation). Commonly, this is done through data augmentation, where the training set is enlarged by applying hand-crafted transformations to the inputs. We argue that invariances should instead be incorporated in the model structure, and learned using the marginal likelihood, which correctly rewards the reduced complexity of invariant models. We demonstrate this for Gaussian process models, due to the ease with which their marginal likelihood can be estimated. Our main contribution is a variational inference scheme for Gaussian processes containing invariances described by a sampling procedure. We learn the sampling procedure by back-propagating through it to maximise the marginal likelihood.

GPflow is a Gaussian process library that uses TensorFlow for its core computations and Python for its front end. The distinguishing features of GPflow are that it uses variational inference as the primary approximation method, provides concise code through the use of automatic differentiation, has been engineered with a particular emphasis on software testing and is able to exploit GPU hardware.