This paper presents Diabot, a specialized conversational chatbot designed to assist individuals managing diabetes. The development of Diabot addresses a critical need for accurate, accessible, and empathetic diabetes information in a digital format. Three approaches were explored to enhance the chatbot's capabilities: prompt engineering, fine-tuning of Large Language Models (LLMs), and Retrieval-Augmented Generation (RAG). The implementation focuses on contextual filtering, specialized LLM utilization, and an intuitive user interface. The RAG approach emerged as the most effective solution, overcoming hardware limitations and resource constraints while maintaining response accuracy. The system architecture, implementation challenges, and future improvements are discussed, including the integration of calculation capabilities for precise numerical responses. Diabot demonstrates the potential of specialized conversational AI in healthcare, particularly for chronic condition management like diabetes.

Incorporating security patterns into software architecture is essential for robust system design.Model Driven Engineering (MDE) offers a structured approach to software development, emphasizing modeling and automation.This paper explores the integration of security patterns into software architecture using MDE techniques.We highlight the benefits of this approach, including improved level of security and enhanced maintainability.Challenges such as modeling complexity and tool support are also discussed.Through a SCADA (Supervisory Control and Data Acquisition) system case study, we demonstrate the effectiveness of integrating security patterns into software architecture using MDE.

The strategy design pattern is an essential behavioral design pattern.When developing software it is relied upon for the development of modular components.Given the constraints inherent to the inner workings of the Blockchain and features offered by Smart Contract languages, some of the design patterns commonly used in software systems development cannot be naively implemented.This paper explores a new pattern implementation of the mentioned pattern in a contract oriented language aimed at Smart-Contracts development, specifically the solidity language.The pattern's implementations discussed in this paper are intended to be deployed and run on the Ethereum Blockchain.

The free throw in basketball is one of the important skills for a player. The accuracy of success of the throw affects the final score in a game, that’s why players make more effort to have a high accuracy in scoring free throw.The studies proved that as a human being, every player has his own way to throw the ball. The player position, the ball shot angle and velocity, are the factors that change from a player to another, and therefore affect the success of the throw even for the same player.The aim of this work is to propose a method to automatically measure and correct the extracted parameters of a basketball free throw shooting. The approach aims to create an automated system to correct the position of basketball throws and track the movement of the ball.

Digitalization is transforming manufacturing systems to become more agile and smart thanks to the integration of sensors and connection technologies that help capture data at all phases of a product's life cycle.Digitalization promises to improve manufacturing flexibility, quality, productivity, and reliability.However, there is still a significant need of effective methods to develop models that could enhance the capabilities of manufacturing systems.Formal methods and tools are becoming essential to achieve this objective.Within this context, this paper introduces a formal software solution for the automatic generation of modular manufacturing control systems software.The proposed solution leverages software Model-Based Design (MBD) techniques to reduce development effort, time, and human error by automating several manual steps.

International audience

Design of complex systems goes through a multiview paradigm in which separate teams, from different business viewpoints, build partial models describing the system. As they are expressed in different languages, these partial models are called heterogeneous models. To maintain the global system's consistency, we propose a collaborative approach that combines Group Decision Making (GDM) and Model-Based Engineering. This paper presents a metamodel for collaborative decision elaboration via a set of decision policies which are instances of GDM patterns. Our approach is illustrated with a hospital Emergency Department case study and is supported by a tool allowing models alignment through GDM based processes.

Design of complex systems implies various points of view expressed by stakeholders with different areas of expertise. Each stakeholder describes his model in a Domain Specific Language, according to his point of view. Ensuring the consistency of the global system and building a cross view is a challenging task. It requires the involvement of all stakeholders to produce intermodel correspondences that satisfy their concerns. In this paper, we first introduce a metamodel of collaboration that formalizes collaborative work, then we use this metamodel to define a collaborative process for heterogeneous design models matching. This approach establishes semantic links at metamodel level by following a group decision-making process, then it refines those links semi-automatically at model level by exploiting their semantics.

Companies use these different approaches to elicit, specify, analyse and validate their requirements in different contexts.The globalization and the rapid development of information technologies sometimes require companies to work together in order to achieve common objectives as quickly as possible.We propose a Unified Requirements Engineering meta-model (UREM) that allows cooperation in the requirements engineering process between heterogeneous RE (Requirement Engineering) models.In this paper, we explore UREM as a pivot meta-model to ensure interoperability between heterogeneous RE models.

This paper falls into the context of modeling complex systems according to various viewpoints. More precisely, it presents an iterative process of heterogeneous models consistency management - by taking into account various types of evolution - based on building a correspondence model. In the case of models evolution, this process is intended to capture changes in the models, to list modifications to be made in the impacted models and finally to update the correspondence model for a future iteration.

This work is situated in the context of the application of Model Driven Engineering to complex systems view-based modelling. In fact, view-based models - called also partial models - are manipulated by different actors (designers), and are thus generally heterogeneous, that is, described with different DSLs (Domain Specific Languages). Instead of building a single global model, which is not realistic, we propose to organize the different partial models as a network of related models, which provides a global view of the system through a correspondence model. As models are modelled separately by different designers, they also evolve separately that induces a problem of consistency. To solve it, we propose a semi-automatic process based on the correspondence model allowing detecting changes, calculating their impacts, and proposing modifications to maintain the consistency among them. The approach is supported by a tool chain and illustrated by the example of a Bug Tracking System.