GUI-Based DL-Network Designer for KISTI’s Supercomputer Users

With the increase in research on AI(Artificial Intelligence),the importance of DL(Deep Learning)in various fields,such as materials,biotechnology,genomes,and new drugs,is increasing significantly,thereby increasing the number of deep-learning framework users.However,to design a deep neural network,a considerable understanding of the framework is required.To solve this problem,a GUI(Graphical User Interface)-based DNN(Deep Neural Network)design tool is being actively researched and developed.The GUI-based DNN design tool can design DNNs quickly and easily.However,the existing GUI-based DNN design tool has certain limitations such as poor usability,framework dependency,and difficulty encountered in changing GUI components.In this study,a deep learning algorithm that solves the problem of poor usability was developed using a template to increase the accessibility for users.Moreover,the proposed tool was developed to save and share only the necessary parts for quick operation.To solve the framework dependency,we applied ONNX(Open Neural Network Exchange),which is an exchange standard for neural networks,and configured it such that DNNs designed with the existing deep-learning framework can be imported.Finally,to address the difficulty encountered in changing GUI components,we defined and developed the JSON format to quickly respond to version updates.The developed DL neural network designer was validated by running it with KISTI’s supercomputer-based AI Studio.

A mixed behaviour equilibrium model with mode choice and its application to the endogenous demand of automated vehicles

This study develops a mixed behavioural equilibrium model with explicit consideration of mode choice(MBE-MC)in a transportation system where fully automated vehicles(AV)coexist with conventional human-driven vehicles(HV).For the mode choice,travellers select among three options,following a logit modal split:driving their private HV,or taking an AV mobility service provided by either a firm or the government.For the route choice,the HV drivers follow the random utility maximisation principle while central agents route the AV passengers following the Cournot Nash(firm agent)or Social Optimal(government agent)principles.We consider two types of travel costs(i.e.travel time and monetary travel cost)to characterise the new features(e.g.expanded link capacity and reduced value of time)of the mixed AVeHV transportation system.We model the MBE-MC problem in a combined modeeroute choice framework and formulate it as a route-based variational inequality(VI)problem.We show the equivalence between the VI formulation and the MBE-MC problem,and the existence of a solution to the MBE-MC problem.Then,we modify a partial linearisation algorithm for solving the proposed model.Numerical results validate the equilibrium conditions and show the efficacy of the new model in capturing the features of the mixed AVeHV transportation system.The impact patterns of different parameters on(1)the network performance in terms of AV share and system cost and(2)on the solution efficiency are analysed.