Deep Learning Applied to Computational Mechanics:A Comprehensive Review,State of the Art,and the Classics

Three recent breakthroughs due to AI in arts and science serve as motivation:An award winning digital image,protein folding,fast matrix multiplication.Many recent developments in artificial neural networks,particularly deep learning(DL),applied and relevant to computational mechanics(solid,fluids,finite-element technology)are reviewed in detail.Both hybrid and pure machine learning(ML)methods are discussed.Hybrid methods combine traditional PDE discretizations with ML methods either(1)to help model complex nonlinear constitutive relations,(2)to nonlinearly reduce the model order for efficient simulation(turbulence),or(3)to accelerate the simulation by predicting certain components in the traditional integration methods.Here,methods(1)and(2)relied on Long-Short-Term Memory(LSTM)architecture,with method(3)relying on convolutional neural networks.Pure ML methods to solve(nonlinear)PDEs are represented by Physics-Informed Neural network(PINN)methods,which could be combined with attention mechanism to address discontinuous solutions.Both LSTM and attention architectures,together with modern and generalized classic optimizers to include stochasticity for DL networks,are extensively reviewed.Kernel machines,including Gaussian processes,are provided to sufficient depth for more advanced works such as shallow networks with infinite width.Not only addressing experts,readers are assumed familiar with computational mechanics,but not with DL,whose concepts and applications are built up from the basics,aiming at bringing first-time learners quickly to the forefront of research.History and limitations of AI are recounted and discussed,with particular attention at pointing out misstatements or misconceptions of the classics,even in well-known references.Positioning and pointing control of a large-deformable beam is given as an example.

Effects of neurofeedback on the short-term memory and continuous attention of patients with moderate traumatic brain injury： A preliminary randomized controlled clinical trial

Purpose： There are some studies which showed neurofeedback therapy （NET） can be effective in clients with traumatic brain injury （TBI） history. However, randomized controlled clinical trials are still needed for evaluation of this treatment as a standard option. This preliminary study was aimed to evaluate the effect of NET on continuous attention （CA） and short-term memory （STM） of clients with moderate TBI using a randomized controlled clinical trial （RCT）. Methods： In this preliminary RCT, seventeen eligible patients with moderate TBl were randomly allo- cated in two intervention and control groups. All the patients were evaluated for CA and STM using the visual continuous attention test and Wechsler memory scale-4th edition （WMS-IV） test, respectively, both at the time of inclusion to the project and four weeks later. The intervention group participated in 20 sessions of NFT through the first four weeks. Conversely, the control group participated in the same NF sessions from the fifth week to eighth week of the project. Results： Eight subjects in the intervention group and five subjects in the control group completed the study. The mean and standard deviation of participants＇ age were （26.75 ~ 15.16） years and （2Z60 ＋_ 8.17） years in experiment and control groups, respectively. All of the subjects were male. No significant improvement was observed in any variables of the visual continuous attention test and WMS-IV test between two groups （p _~ 0.05）. Conclusion： Based on our literature review, it seems that our study is the only study performed on the effect of NET on TBl patients with control group. NET has no effect on CA and STM in patients with moderate TBI. More RCTs with large sample sizes, more sessions of treatment, longer time of follow-up and different orotocols are recommended.