Optimized operation scheme of flash-memory-based neural network online training with ultra-high endurance

With the rapid development of machine learning,the demand for high-efficient computing becomes more and more urgent.To break the bottleneck of the traditional Von Neumann architecture,computing-in-memory(CIM)has attracted increasing attention in recent years.In this work,to provide a feasible CIM solution for the large-scale neural networks(NN)requiring continuous weight updating in online training,a flash-based computing-in-memory with high endurance(10^(9) cycles)and ultrafast programming speed is investigated.On the one hand,the proposed programming scheme of channel hot electron injection(CHEI)and hot hole injection(HHI)demonstrate high linearity,symmetric potentiation,and a depression process,which help to improve the training speed and accuracy.On the other hand,the low-damage programming scheme and memory window(MW)optimizations can suppress cell degradation effectively with improved computing accuracy.Even after 109 cycles,the leakage current(I_(off))of cells remains sub-10pA,ensuring the large-scale computing ability of memory.Further characterizations are done on read disturb to demonstrate its robust reliabilities.By processing CIFAR-10 tasks,it is evident that~90%accuracy can be achieved after 109 cycles in both ResNet50 and VGG16 NN.Our results suggest that flash-based CIM has great potential to overcome the limitations of traditional Von Neumann architectures and enable high-performance NN online training,which pave the way for further development of artificial intelligence(AI)accelerators.

An Effective Online Collaborative Training in Developing Listening Comprehension Skills

The COVID-19 outbreak severely affected formal face-to-face classroom teaching and learning.ICT-based online education and training can be a useful measure during the pandemic.In the Pakistani educational context,the use of ICT-based online training is generally sporadic and often unavailable,especially for developing English-language instructors’listening comprehension skills.The major factors affecting availability include insufficient IT resources and infrastructure,a lack of proper online training for speech and listening,instructors with inadequate academic backgrounds,and an unfavorable environment for ICT-based training for listening comprehension.This study evaluated the effectiveness of ICT-based training for developing secondary-level English-language instructors’listening comprehension skills.To this end,collaborative online training was undertaken using random sampling.Specifically,60 private-school instructors in Chakwal District,Pakistan,were randomly selected to receive online-listening training sessions using English dialogs.The experimental group achieved significant scores in the posttest analysis.Specifically,there were substantial improvements in the participants’listening skills via online training.Given the unavailability of face-to-face learning during COVID-19,this study recommends using ICT-based online training to enhance listening comprehension skills.Education policymakers should revise curricula based on online teaching methods and modules.

Examining the Effectiveness of Virtual Training under the Kirkpatrick Model:A Post-COVID Study

Teaching and training through online tools were used by most Higher Education Institute(HEIs)worldwide during COVID-19 to cater to the needs of students who stay far away from educational institutions.After the pandemic,the virtual model of education and training became a trend.The students and teachers are also more used to this trend which is giving more opportunities to both learners and instructors.One of the notable benefits of virtual teaching and learning platform is that it provides a flexible environment to gain knowledge,skills,and attitude simultaneously along with formal off-line education.From the earlier studies it is found that the majority of studies have focused on traditional,offline training methods and only a few studies have focused on virtual training.Therefore,the present study examined the effectiveness of virtual training using the Kirkpatrick model.For this purpose,a research instrument based on the Kirkpatrick model was constructed and distributed among the UG and PG students in Mangalore City.A total of 143 responses were collected,of which 132 were completed responses and all of them were considered for analysis.Descriptive statistics and one sample t-test are employed to analyse and interpret the data.The findings revealed that virtual training is more effective compared to traditional and offline training methods.Henceforth,the training and education provided through virtual platforms made significant contributions to the employability of youngsters.

Online LS-SVM for function estimation and classification

An online algorithm for training LS-SVM (Least Square Support VectorMachines) was proposed for the application of function estimation and classification. Online LS-SVMmeans that LS-SVM can be trained in an incremental way, and can be pruned to get sparseapproximation in a decremental way. When a SV (Support Vector) is added or removed, the onlinealgorithm avoids computing large-scale matrix inverse. Thus the computation cost is reduced. Onlinealgorithm is especially useful to realistic function estimation problem such as systemidentification. The experiments with benchmark function estimation problem and classificationproblem show the validity of this online algorithm.

Aerodynamicmodeling using an end-to-end learning attitude dynamics network for flight control

A novel identification method of aerodynamicmodels using a physics neural network,named the attitude dynamics network,which incorporates the attitude dynamics of an aircraft without any prior aerodynamic knowledge,is proposed.Then a learning controller,which combines feedback linearization with sliding mode control,is developed by introducing the learned aerodynamicmodels.The merit of the identification method is that the aerodynamicmodels can be learned end-to-end by the physics network directly from the flight data.Consequently,the paper uses an offline scheme and an online scheme to combine the identification process and the control process.In the offline scheme,learning the aerodynamic models and controlling the aircraft compose a cascade system,whereas the online scheme,similar to Learn-to-Fly,is a parallel system.Specifically,in the offline scheme,the physics neural network is trained by sufficient offline flight data,and then the trained network is substituted into the controller.The online scheme refers to the controller making the aircraft fly to generate flight data and sending these data to the deep network at the time of training,while the deep network provides the trained aerodynamic models to the controller at other times.Simulation results show that both under nominal and disturbance aerodynamic conditions,the network trained offline with a large amount of nominal data approximate the aerodynamicmodels well.Thus,the performance of the controller reaches a good level;for the online scheme,the predictive capability of the network increases and the performance of the controller improves with more training data.