Software and Systems Modeling最新文献

Conversational user interfaces (CUIs), such as chatbots, are becoming a common component of many software systems. Although they are evolving in many directions (such as advanced language processing features, thanks to new AI-based developments), less attention has been paid to access control and other security concerns associated with CUIs, which may pose a clear risk to the systems they interface with. In this paper, we apply model-driven techniques to model and enforce access-control policies in CUIs. In particular, we present a fully fledged framework to integrate the role-based access-control (RBAC) protocol into CUIs by: (1) modeling a set of access-control rules to specify permissions over the bot resources using a domain-specific language that tailors core RBAC concepts to the CUI domain; and (2) describing a mechanism to show the feasibility of automatically generating the infrastructure to evaluate and enforce the modeled access control policies at runtime.

Over the past few years, several software companies have emerged that offer process mining tools to assist enterprises in gaining insights into their process executions. However, the effective application of process mining technologies depends on analysts who need to be proficient in managing process mining projects and providing process insights and improvement opportunities. To contribute to a better understanding of the difficulties encountered by analysts and to pave the way for the development of enhanced and tailored support for them, this work reveals the challenges they perceive in practice. In particular, we identify 23 challenges based on interviews with 41 analysts, which we validate using a questionnaire survey. We provide insights into the relevancy of the process mining challenges and present mitigation strategies applied in practice to overcome them. While mitigation strategies exist, our findings imply the need for further research to provide support for analysts along all phases of process mining projects on the individual level, but also the technical, group, and organizational levels.

As neural networks are increasingly being integrated into mission-critical systems, it is becoming crucial to ensure that they meet various safety and liveness requirements. Toward, that end, numerous complete and sound verification techniques have been proposed in recent years, but these often suffer from severe scalability issues. One recently proposed approach for improving the scalability of verification techniques is to enhance them with abstraction/refinement capabilities: instead of verifying a complex and large network, abstraction allows the verifier to construct and then verify a much smaller network, and the correctness of the smaller network immediately implies the correctness of the original, larger network. One shortcoming of this scheme is that whenever the smaller network cannot be verified, the verifier must perform a refinement step, in which the size of the network being verified is increased. The verifier then starts verifying the new network from scratch—effectively “forgetting” its earlier work, in which the smaller network was verified. Here, we present an enhancement to abstraction-based neural network verification, which uses residual reasoning: a process where information acquired when verifying an abstract network is utilized in order to facilitate the verification of refined networks. At its core, the method enables the verifier to retain information about parts of the search space in which it was determined that the refined network behaves correctly, allowing the verifier to focus on areas of the search space where bugs might yet be discovered. For evaluation, we implemented our approach as an extension to the Marabou verifier and obtained highly promising results.

Since the dawn of programming, several developments in programming language design and programming methodology have been hailed as the end of the profession of programmer; they have all proven to be exaggerated rumors, to echo the words attributed to Mark Twain. In this short paper, we ponder the question of whether the emergence of large language models finally realizes these prophecies? Also, we discuss why even if this prophecy is finally realized, it does not change the job of the researcher in programming.