Digitalized manufacturing processes necessitate a shift from traditional production control systems to more intelligent frameworks. Existing systems reliant on programmable logical controllers fall short in handling the influx of data generated by the Internet of Things layer and the array of IT systems integral to modern operations. This article presents technical developments made within MESLedger, a collaborative project aiming to enhance manufacturing execution by integrating blockchain technology and artificial intelligence. The first part of our work investigates how deep reinforcement learning, particularly the proposed Dual Attention Network for Multi-Objective Proximal Policy Optimization, can optimize job scheduling by balancing objectives such as makespan and energy consumption. This approach formulates the Flexible Job Shop Scheduling Problem as a Markov decision process and uses actor-critic networks enhanced with dual attention mechanisms to generate adaptive and context-aware scheduling strategies. The second part explores how blockchain technology, specifically Hyperledger Fabric, can be used to secure communication within collaborative industrial environments by managing critical manufacturing processes through smart contracts and decentralized architectures. Together, these components form the foundation of an intelligent manufacturing execution systems environment capable of dynamically analyzing data and orchestrating operations based on customizable performance priorities.

The Complexity of Buildings nowadays has made the digitalization a must, by connecting the whole technical data chain of the built environment using Building Information Modeling (BIM) processes among others; Scan to BIM, 3D Modeling and Data Exchange… etc., makes it easy to monitor and to manage better the information covering the entire lifecycle of a project. Moreover, an extended collaboration model operates on a strong lifecycle system, provides an efficient environment for managing complex data from the Architecture, Engineering, and Construction (AEC) industry and ensures smooth interactions between designers, suppliers, and builders. In this paper, the research team investigates with industrial partners how to develop solutions that can improve BIM workflows by using lifecycle approaches. Those workflows contain the steps adopted to reach the goal of integrating BIM and lifecycle approaches which called Building Lifecycle Management (BLM) that includes Technical Data Management and Lifecycle Assessment, to reduce the environmental impacts of buildings. BIM-based lifecycle approach is an emerging and promising research area addressing the critical concern of comprehensive lifecycle management within the AEC industry.

Extracting relevant clauses from legal contracts is a challenging task due to the complex structure and specialized language of legal documents. Accurate clause identification is critical for contract analysis but often requires legal expertise and significant manual effort. In this paper, we propose an efficient framework for extracting clauses from legal contracts by formulating the task as a Question Answering (QA) problem, enhanced through knowledge distillation in a teacher–student architecture. The teacher model, built upon a transformer from the BERT family pre-trained on contract-specific data, is fine-tuned to capture deep contextual understanding, while a lightweight student model learns to replicate the teacher’s performance with reduced resource demands. To support this, we develop a custom extractive QA dataset using contracts from the SEC EDGAR database by integrating selected span-based question-answer pairs from the CUAD dataset and manually annotating additional clause types. Through a comprehensive evaluation, our distilled model achieves an AUPR of 0.723 and a precision of 0.682 at 80

Modern manufacturing environments frequently experience disruptions such as urgent orders, resource failures, and varying resource availability. This paper addresses these challenges in the dynamic Flexible Job Shop Scheduling Problem (FJSP) by extending an existing reinforcement learning model to accommodate real-time production changes. The initial sections examine the original FJSP model, emphasizing its core constraints and optimization methodology. Subsequently, newly introduced constraints such as additional job arrivals and resource breakdowns are incorporated into the baseline formulation. The resulting framework continually revises resource allocations and operation sequences in response to these events, ensuring efficient and feasible scheduling. This work demonstrates a practical strategy for adapting static approaches to dynamic contexts, thereby supporting more resilient and real-time decision-making in modern flexible manufacturing systems.

Network intrusion detection systems are crucial for securing information technology and operational technology networks against cyberattacks. While machine learning and deep learning techniques hold significant promise for enhancing these systems, their performance is highly dependent on how network traffic data is transformed and represented. In a survey of recent popular papers, we identified four main categories of data representations: numerical, pixel-based, sequence-based, and graph-based approaches. The identified transformations capture information either from network traffic packets, flows, or both. Using insights from the literature and additional experiments conducted on the CICIDS-2017 dataset, we assessed each representation not only in terms of its ability to enhance detection performance but also in terms of computational efficiency. Our findings highlight the need for future research to improve data transformation techniques, especially in terms of dataset labeling and inference time reporting, to support the development of more robust and practical network intrusion detection systems.

Digitalized manufacturing processes necessitate a shift from traditional production control systems to more intelligent frameworks. Existing systems reliant on Programmable Logical Controllers (PLCs) fall short in handling the influx of data generated by the Internet of Things (IoT) layer and the array of IT systems integral to modern operations. This paper proposes the development of a novel manufacturing execution environment capable of dynamically analyzing data, orchestrating operations, and making informed decisions, especially in response to malfunctions. Furthermore, such environment aims to optimize production activities based on diverse priorities such as cost, energy efficiency, and production time. This necessitates a reevaluation of manufacturing operations, incorporating both standard and parametric design elements. The success of such environment is crucial for industries motivated to achieve optimal performance through intelligent data utilization. From a scientific perspective, the challenge lies in devising scalable algorithms capable of autonomously driving digital production systems while maintaining a high level of adaptability and efficiency. This paper presents MESLedger, a collaborative project aiming at improving manufacturing execution through the integration of blockchain technology to secure communication in collaborative environments and artificial intelligence to optimize major aspects of production. It also emphasizes on the significance of this effort in meeting industrial demands and outline the scientific advancements required to realize the project's objective.

Network Intrusion Detection Systems play a pivotal role in preventing cyber attacks by identifying threats within computer networks. Recent advancements in deep learning techniques positioned them as highly effective methods in detecting a diverse range of cyber attacks. However, the "Black-Box" nature of deep models makes understanding their decisions very challenging, and renders them susceptible to adversarial attacks. In this paper, we propose the use of Explainable AI (XAI) approaches in deep-learning-based network traffic classifiers to validate their decisions' rationale and soundness. In particular, we combine the popular Grad-CAM technique with a reverse lookup algorithm to explain models trained using image-transformed raw network traffic sessions, encompassing general, malware, and encrypted traffic data. Model behaviors were analyzed by mapping the highly impacting pixels to their corresponding raw features, to facilitate investigating the meaningfulness of the features learned by the model. Experimental results indicate cases of consistent highlighting of pixels associated with network layers across specific traffic types. However, models occasionally used unexpected features during the classification process, raising security vulnerability concerns that merit serious investigation. The proposed approach serves as a valid method to explain the behavior of general black-box image-based network traffic classification models and assess their robustness. The implementation code is available at https://github.com/ayahdev/XAI-Image-Based-IDS.

Building Information Modeling (BIM) is a smart process that uses 3D models and information management to assist Architecture/ Engineering/Construction and Facility Management (AEC-FM) practitioners in design, construct and operate buildings and infrastructure more efficiently and more economically. BIM integrates new technologies and tools to facilitate the exchange of information between the different stakeholders involved in a project, from the initial conception to the final demolition. However, one of the challenges BIM is facing is how to manage all those information generated throughout the entire lifecycle of any built asset, which can span several decades and involve multiple changes and updates. Unlike the manufacturing industry (such as aeronautics or automobiles), which has established a set of procedures and standards to document and control the product development and delivery across the whole lifecycle, the AEC-FM sector is still in the process of adopting and implementing lifecycle management, assessment concepts and practices. In this paper, we review the literature on how life-cycle approaches can be integrated with BIM enabled projects in the AEC-FM sector. We also discuss some of the benefits that lifecycle approaches can bring along with the digitalization of the sector, in terms of mitigating the environmental impacts and increasing the performance of buildings.

We present a consensus mechanism in this paper that is designed specifically for supply chain blockchains, with a core focus on establishing trust among participating stakeholders through a novel reputation-based approach. The prevailing consensus mechanisms, initially crafted for cryptocurrency applications, prove unsuitable for the unique dynamics of supply chain systems. Unlike the broad inclusivity of cryptocurrency networks, our proposed mechanism insists on stakeholder participation rooted in process-specific quality criteria. The delineation of roles for supply chain participants within the consensus process becomes paramount. While reputation serves as a well-established quality parameter in various domains, its nuanced impact on non-cryptocurrency consensus mechanisms remains uncharted territory. Moreover, recognizing the primary role of efficient block verification in blockchain-enabled supply chains, our work introduces a comprehensive reputation model. This model strategically selects a leader node to orchestrate the entire block mining process within the consensus. Additionally, we innovate with a Schnorr Multisignature-based block verification mechanism seamlessly integrated into our proposed consensus model. Rigorous experiments are conducted to evaluate the performance and feasibility of our pioneering consensus mechanism, contributing valuable insights to the evolving landscape of blockchain technology in supply chain applications.

The preservation of research grants’ data is essential for long-term accessibility, auditing, and informed decision-making. This paper presents the implementation of the Open Archival Information System (OAIS) model for archiving research grants’ data. The proposed system ensures structured storage, retrieval, and management of critical research outputs, including reports, publications, and financial records. While the implementation has addressed key challenges such as data ingestion, metadata management, and security, certain aspects require further refinement to optimize scalability and accessibility. Despite these challenges, the OAIS-based approach provides a robust foundation for future-proofing research data, ensuring its availability for strategic planning, performance evaluation, and institutional decision-making. The study highlights lessons learned from the initial deployment and outlines enhancements needed to improve system efficiency. Ultimately, this model serves as a sustainable framework for research data archiving, contributing to a more data-driven research ecosystem.

The rise of Industry 4.0 technologies, such as the Internet of things (IoT), Cyber-Physical Systems (CPS), cloud computing, and Artificial Intelligence (AI). Has transformed traditional manufacturing into smart, data-driven systems. This shift has increased the complexity of production scheduling, especially in the Flexible Job-Shop Scheduling Problem (FJSP), which is already a complex problem by itself and further complicated by the integration of robotic job transfers. To address these challenges, we present the Dual Attention Network for MultiObjective Proximal Policy Optimization (DANMO-PPO) model, a Deep Reinforcement Learning (DRL) framework that formulates the FJSP as a Markov Decision Process (MDP). Our model employs actor-critic networks with a dual-attention mechanism to prioritize scheduling actions based on operational and machinelevel features. It is trained using the Proximal Policy Optimization (PPO) algorithm and guided by a weighted reward function balancing makespan and energy consumption. Experimental results show that DANMO-PPO effectively learns adaptive scheduling policies suited for complex, real-time industrial environments.

Business Process Mining is considered one of the emerging fields that focuses on analyzing Business Process Models (BPM), by extracting knowledge from event logs generated by various Supply chain Management information systems, such as ERP and PLM, for the purpose of auditing, monitoring, and analysis of business activities for future improvement and optimization throughout the entire lifecycle of such processes, from creation to conclusion. Such a framework will help enhance the accuracy of anomaly detection in the global Supply Chain, improve the multi-level business processes workflow, and optimize the processes in the Supply Chain in terms of security and automation. In this work, a Bidirectional Long Short-Term Memory (Bi-LSTM)-autoencoder Neural Network was utilized for the prediction of the execution of cases, through training and testing the model on event traces extracted from event logs related to a procurement business process model, which is one of the main components of the supplychain. The approach consisted of three phases: preprocessing the logs, classification, and categorization in addition to all the activities related to implementing the Bi-LSTM model, including network design, training, and model selection. Our results showed that our model was able to predict the next activity in the sequence as well as detect anomalous ones with over 80

Efficient contract management is essential for ensuring sustainable and reliable supply chains; yet, traditional methods remain manual, error-prone, and inefficient, leading to delays, financial risks, and compliance challenges. AI and blockchain technology offer a transformative alternative, enabling the establishment of automated, transparent, and self-executing smart contracts that enhance efficiency and sustainability. As part of AI-driven smart contract automation, we previously implemented contractual clause extraction using question answering (QA) and named entity recognition (NER). This paper presents the next step in the information extraction process, relation extraction (RE), which aims to identify relationships between key legal entities and convert them into structured business rules for smart contract execution. To address RE in legal contracts, we present a novel hierarchical transformer model that captures sentence- and document-level dependencies. It incorporates global and segment-based attention mechanisms to extract complex legal relationships spanning multiple sentences. Given the scarcity of publicly available contractual datasets, we also introduce the contractual relation extraction (ContRE) dataset, specifically curated to support relation extraction tasks in legal contracts, that we use to evaluate the proposed model. Together, these contributions enable the structured automation of legal rules from unstructured contract text, advancing the development of AI-powered smart contracts.

In recent years, significant strides in various domains have been fueled by the convergence of large-scale datasets and sophisticated machine learning algorithms. Nevertheless, the utilization of these datasets poses challenges, including privacy concerns, data ownership issues, and resource limitations. Cooperative data learning approaches have emerged as a solution, allowing multiple parties to collaboratively train machine learning models using their distributed data. While Federated Learning (FL) addresses the issue of privacy concerns, reluctance among data owners to share their data remains a challenge. It is imperative to provide incentives for participation in these cooperative learning settings to boost effectiveness and promote the widespread adoption of such approaches. This paper introduces an RL-ICDL-BC framework that seamlessly integrates principles of incentive design and cooperative learning, fostering effective collaboration among data owners. The framework's primary objective is to motivate and reward participants for contributing their models while simultaneously preserving privacy and ensuring fairness in the learning process. Experimental evaluations utilizing Covid-19 datasets and diverse collaborative learning scenarios demonstrate the effectiveness of the proposed framework. The results reveal that incentivizing cooperative data learning leads to increased participation rates, improved model performance, and enhanced fairness in the learning process. Despite the challenges posed by non-iid data, the experiments yield outstanding outcomes, showcasing a Covid19 virus detection accuracy rate of approximately 99%. This exceptional accuracy underscores the efficacy of our proposed approach in effectively detecting and mitigating the transmission of infectious diseases.

Achieving an alignment of the components within an Enterprise Architecture (EA) is challenging since it reflects both the business and IT views, and it must be regularly updated in response to the changes of the firm. Blockchain technology, and more specifically the means by which the logic of smart contracts may be extended to the business, is one avenue that has been explored and that may still be fruitful in addressing the issue. Through the use of smart contracts, we offer a new form of activity for operational processes that may identify when the process in which they are engaged exhibits unexpected behavior and so provide early warning of the need to update the EA. This research goes in depth as well as proposes a model to allow for a continuous alignment between the IT and business operations. Not only during normal circumstances would this model be able to be upheld but the research focuses on instances where both operations might experience unfavorable situations due to unforeseen circumstances. In these instances by implementing a blockchain solution both IT and business operations can stay intact without the inclusion of a third party allowing for the blockchain to make autonomous decisions as well as providing a middle ware between both entities to continue their operations.

Purpose. To assess the impact of mining waste on the heavy metal content of water surfaces, plants, and topsoil near the tailings dam of a Zn-Pb mine using both biotests and analytical methods. Methodology. A battery of microbiotests on different animal and plant species was carried out, making it possible to evaluate the toxic effect of residues and surrounding soils on living organisms. Furthermore, the possible relationship between the observed toxicity and the results of the physicochemical analysis of the samples was studied. Findings. The tests showed that the topsoil in contact with the tailings dam is slightly toxic to the living organisms used while the mining tailings are toxic or even very toxic. The heavy metal content of the samples is particularly high for Fe, Zn, Pb and Cu. The correlation of physic-chemical parameters and the results of microbiotests using the principal components analysis (PCA) and the multiple correspondence factor analysis (MCFA) indicate that the toxicity of tailings and the surrounding topsoil can be associated with anthropogenic mining activity. Originality. The study aimed to assess the impact of mining waste on the heavy metal content using biotests and analytical methods. The evaluation considers the concentrations of the samples (highly concentrated samples and samples after dilution) and the different phases of exposure (solid, liquid) for a more detailed assessment of the potential toxicity of the samples. Practical value. It is important to conduct a comprehensive assessment of mining waste and the risks it may pose to humans and the environment in order to develop an adequate rehabilitation plan.

Responding to inquiries within the legal field is notably challenging due to the complexity and variability of legal documents. Delivering precise responses to legal questions often requires domain-specific expertise, posing difficulties even for experienced professionals. The Question-Answering (QA) task, a subtask of Natural Language Processing (NLP), is designed to generate answers to natural language questions. In this study, we explore how QA systems can improve contract analysis by accurately extracting relevant clauses using advanced deeplearning models. By carefully creating a subset of the CUAD dataset, focusing only on relevant categories, we aimed to improve the accuracy of our models. We thoroughly tested several top transformer models to evaluate their performance in extracting important clauses from complex legal documents. Our findings show the great promise of these models in automating and enhancing the accuracy of legal document analysis.

Appears in: INTED2024 Proceedings Publication year: 2024Pages: 1135-1143ISBN: 978-84-09-59215-9ISSN: 2340-1079doi: 10.21125/inted.2024.0360Conference name: 18th International Technology, Education and Development ConferenceDates: 4-6 March, 2024Location: Valencia, Spain

Product design is often a process that involves multiple parties collaborating with each other to design a final product. The involvement of multiple parties induces several risks associated with cyber security and intellectual property theft. These risks are hard to address especially in the case of traditional centralized platforms which may be prone to misconfiguration and software vulnerabilities. As a potential solution, we aim at addressing the issue of design data exchange through decentralized platforms such as the blockchain. Our solution leverages data formats that can segment product data models and gives the ability to control access to data through a decentralized platform which can be fully integrated with PLM processes through APIs. A proof of concept of this solution using the open-source Odoo PLM platform as well as the Hyperledger Fabric blockchain development platform is demonstrated.

Entity recognition and extraction from contracts play a crucial role in automating contract analysis and extracting valuable information. Named Entity Recognition (NER) techniques are used for identifying and classifying specific entities such as parties, dates, amounts, and clauses within contracts. In this study, we create a high-quality NER dataset from various types of English language contracts by considering their structure, and the legal terminology used within these documents. We present a systematic approach to manually annotate contracts with appropriate entity labels, ensuring accuracy and consistency. The resulting NER dataset serves as a valuable resource for training and evaluating NER models for contract analysis tasks. We evaluate the performance of NER on this dataset using a range of methods. These methods include Conditional Random Fields, various Bidirectional LSTM configurations, and BERT models. Each of these models brings different strengths and capabilities to the task of entity recognition, allowing for a comprehensive evaluation and the selection of the best models over the dataset. Among these, the NER model based on Contracts–BERT–base from the Legal–BERT family, which is pre-trained specifically on English contracts, outperformed all others, achieving an impressive overall F1 score of 0.94.