In this perspective article,we first recall the historic background of human-cyber-physical systems(HCPSs),and then introduce and clarify important concepts.We discuss the key challenges in establishing the scientific...In this perspective article,we first recall the historic background of human-cyber-physical systems(HCPSs),and then introduce and clarify important concepts.We discuss the key challenges in establishing the scientific foundation from a system engineering point of view,including(1)complex heterogeneity,(2)lack of appropriate abstractions,(3)dynamic black-box integration of heterogeneous systems,(4)complex requirements for functionalities,performance,and quality of services,and(5)design,implementation,and maintenance of HCPS to meet requirements.Then we propose four research directions to tackle the challenges,including(1)abstractions and computational theory of HCPS,(2)theories and methods of HCPS architecture modelling,(3)specification and verification of model properties,and(4)software-defined HCPS.The article also serves as the editorial of this special section on cyber-physical systems and summarises the four articles included in this special section.展开更多
Neural networks, as an important computing model, have a wide application in artificial intelligence (AI) domain. From the perspective of computer science, such a computing model requires a formal description of its b...Neural networks, as an important computing model, have a wide application in artificial intelligence (AI) domain. From the perspective of computer science, such a computing model requires a formal description of its behaviors, particularly the relation between input and output. In addition, such specifications ought to be verified automatically. ReLU (rectified linear unit) neural networks are intensively used in practice. In this paper, we present ReLU Temporal Logic (ReTL), whose semantics is defined with respect to ReLU neural networks, which could specify value-related properties about the network. We show that the model checking algorithm for theΣ2∪Π2 fragment of ReTL, which can express properties such as output reachability, is decidable in EXPSPACE. We have also implemented our algorithm with a prototype tool, and experimental results demonstrate the feasibility of the presented model checking approach.展开更多
基金the Capacity Development Fund of Southwest University,China(No.SWU116007)the National Natural Science Foundation of China(Nos.61732019,61672435,61811530327,and 62032019)。
文摘In this perspective article,we first recall the historic background of human-cyber-physical systems(HCPSs),and then introduce and clarify important concepts.We discuss the key challenges in establishing the scientific foundation from a system engineering point of view,including(1)complex heterogeneity,(2)lack of appropriate abstractions,(3)dynamic black-box integration of heterogeneous systems,(4)complex requirements for functionalities,performance,and quality of services,and(5)design,implementation,and maintenance of HCPS to meet requirements.Then we propose four research directions to tackle the challenges,including(1)abstractions and computational theory of HCPS,(2)theories and methods of HCPS architecture modelling,(3)specification and verification of model properties,and(4)software-defined HCPS.The article also serves as the editorial of this special section on cyber-physical systems and summarises the four articles included in this special section.
基金This work is supported by the National Natural Science Foundation of China under Grant No.61872371the Open Fund from the State Key Laboratory of High Performance Computing of China(HPCL)under Grant No.202001-07the Natural Key Research and Development Program of China under Grant No.2018YFB0204301.
文摘Neural networks, as an important computing model, have a wide application in artificial intelligence (AI) domain. From the perspective of computer science, such a computing model requires a formal description of its behaviors, particularly the relation between input and output. In addition, such specifications ought to be verified automatically. ReLU (rectified linear unit) neural networks are intensively used in practice. In this paper, we present ReLU Temporal Logic (ReTL), whose semantics is defined with respect to ReLU neural networks, which could specify value-related properties about the network. We show that the model checking algorithm for theΣ2∪Π2 fragment of ReTL, which can express properties such as output reachability, is decidable in EXPSPACE. We have also implemented our algorithm with a prototype tool, and experimental results demonstrate the feasibility of the presented model checking approach.
