Distinguishing crop from weed species at the seedling stage is a fine-grained morphological discrimination problem. We study it on the public Plant Seedlings benchmark, which contains twelve species (three crops and nine weeds) imaged at the seedling stage. Automated classifiers report near-ceiling accuracy on this benchmark. Yet its images are acquired in controlled trays containing soil, gravel, rulers, barcodes, and printed labels, so a model may identify a species from its acquisition context rather than its morphology. We audit two architectures, EfficientNet-B7 and ViT-B/16, trained on the V2 dataset (5539 images) and probed with plant-only, background-only, and background-swapped inputs. Near-ceiling models (96.1% and 96.4% over three seeds) recover the correct species for up to 47.7% of samples from the background alone (chance 8.3%) and lose about 60 points under background swapping. Context reliance is therefore a property of the benchmark, not any single architecture. Tracing this to its source, an independent learned representation of the plant-free background alone identifies the species at 72.4%. The reliance is correctable end to end: a consistency-regularisation scheme retains 92.1% full-image accuracy for the transformer with no segmentation at inference, at an architecture-dependent cost. Reported accuracy thus partly measures acquisition context, not morphology; morphological grounding should be measured and reported alongside accuracy.

Agriculture is critical in global food security but faces challenges like declining arable land and climate change. Machine learning, particularly deep learning with Convolutional Neural Networks (CNN), has shown promise in addressing agricultural problems. Efficient and accurate identification of plant diseases is essential for guaranteeing the long-term viability of food production. This paper investigates the use of CNNs to classify tomato diseases, employing a deep learning approach automatically. We utilize a publicly available dataset consisting of more than 16,000 photos of tomato leaves. These images cover both healthy and affected leaves by nine different diseases. Our proposal involves the development of three different CNN models, each with a different architectural design. In addition, we applied some Natural Language Generation (NLG) to generate reports and possible recommendations for detected diseases. The results of our research emphasize the capacity of deep learning to effectively and precisely detect diseases in tomatoes. This creates opportunities for the development of pragmatic smartphone applications that empower farmers by providing prompt disease diagnosis, empowering them to make well-informed decisions on crop management and pesticide utilization.

Agriculture is a sector that represents a real indicator of the country’s development, and it is a sector that affects national security. The palm tree is a significant, affordable, and nutrient-rich food source that gives millions of people access to wholesome food globally. It can get a lot of illnesses, much like other plants. DL has been effectively implemented in agriculture and has been employed in many other industries. This paper examines the latest advancements in the use of deep learning (DL) for the diagnosis of palm tree diseases. Particularly Convolutional Neural Networks (CNN) are often utilised in agricultural research due to its strong image processing ability.

Population growth necessitates the urgent enhancement of sustainable food production systems. However, the overuse of fertilizers significantly undermines soil fertility, posing a dual threat to agricultural productivity and environmental integrity. This paper introduces an innovative machine learning (ML) methodology integrated with interpretable artificial intelligence (IAI) aimed at promoting sustainable soil management practices. We conduct a thorough investigation into interpretable ML models specifically designed for the classification of soil fertility. Our approach meticulously analyzes model outcomes while pinpointing critical features that influence predictions regarding soil fertility. The results of this method exhibit remarkable promise, achieving high accuracy in predicting soil fertility.

Agriculture is considered one of the most vital industries worldwide responsible for people's food and the country's economic development.Nevertheless, this sector is the largest consumer of freshwater.Access to water, soil kind, weather conditions, fertilizers, and illnesses are crucial aspects of agricultural activities.Moreover, water scarcity has become a significant issue for farmers to tackle irrigation activity.While irrigation depends on water availability, building smart irrigation systems that control and manage water use efficiency is highly needed.This paper examines the most modern technologies used in irrigation systems over the last few years.This survey can easily be thorough and beneficial to academics while providing essential recommendations for the new era of irrigation systems technologies.The findings indicate widespread use of the Internet of Things (IoT) and artificial intelligence algorithms, while digital twins and blockchain technologies are less common and still in the early stages.To reduce costs and improve security in irrigation systems, it is strongly recommended to adopt digital twin technology.By utilizing digital twins, farmers can enhance operations, optimize resource utilization, and increase efficiency.Furthermore, digital twins contribute to identifying vulnerabilities and implementing robust security measures, ensuring protection against potential threats and disruptions.

With the extensive availability of technological systems all over the world and the increasing diffusion of information and documents by users, especially for the Arabic population, the development of semantic plagiarism detection systems has become essential due to its importance for protecting the rights of authors. Developing such a system that can covers the content of all Arabic documents using ontology as a semantic resource would be a complex and time-consuming task and would require intelligent natural language-processing capabilities. In the context of Arabic plagiarism detection systems using an Arabic ontology and permitting these systems to support semantic representation to more efficiently verify the originality of the research and meet the needs of researchers, this paper presents a novel approach for addressing plagiarism named Multi-agents Indexing System. The proposed system is composed of three phases: (1) natural language processing phase, (2) indexing phase and (3) evaluation phase. Our experimentations are based on the training dataset released for the AraPlagDet curpus. The obtained results indicated that the proposed system has improved the performance of plagiarism detection in Arabic documents with semantic indexing and mutli-agents system.

Recently, Internet has developed into a new technology known as Internet of Things (IoT), enabling the interconnection of billions sensors, actuators, devices, as well as users, which are applied for huge data generating. Moreover, due to the numerous and heterogeneous included devices, and services delivered, the IoT service discovery, selection, and composition is a challenging task. Given that several services have the same functionality but different non-functional criteria (QoS). Thus it is interesting to create an automatic, dynamic, and optimal IoT service composition system to respond on real time to large scale of services, including QoS. Accordingly, we propose, in this paper, an approach for IoT service composition based automatic planning (AP) enhanced by genetic algorithm (GA). The objective is to produce a plan composition satisfying the optimal QoS. In addition, we employ the cloud technology to establish an optimal IoT service composition with less memory consumption, and expanded scalability of our suggested framework. The simulation results have illustrated the efficiency of our approach to resolve the composition problem in a distributed and dynamic IoT system.

Smart agriculture is the use of automation, sensors, and data to improve crop yields, reduce costs and increase crop quality. Predictive analytics, machine learning, and artificial intelligence are used. Another benefit of using artificial intelligence in agriculture is the ability to optimize the consumption of resources such as water and fertilizer to achieve the greatest yield. However, to date, the technology has not been used to optimize the management of agricultural operations. The objective of this paper is to introduce the concept of smart farming and discuss applying deep reinforcement learning for optimal management in the field. We propose a smart farming system that uses Deep Reinforcement Learning to optimize the operation of a farm. The system uses Deep Learning to recognize patterns in a large dataset of real-world agricultural data, such as crop yields and weather, and uses this knowledge to recommend actions. It also dynamically learns optimal control policies for a variety of crops, depending on the farmer's goals and current conditions.

Quran is the most significant religious document in the Arabic language in Islamic law. Several Quranic search engines have been designed and widely used for the past two decades. However, these search engines have certain limitations. For example, in many cases, the search is unable to retrieve relevant verses because it is based on keywords or root search and does not rely on the semantic relation between words in the query. The main objective of the present paper is to design a semantic search engine based on ontology as an index. In our work, we focus on creating a new ontology for Quranic document based on a set of useful words extracted from the Quranic Earab book with grammatical functions that serve as concepts. This ontology will be used as an index in information retrieval. The main idea is to create links between the existing Quranic words in the same verse, which will be used with user's query to find the desired verses. We developed a graphical user interface with free and multiple inputs that convert users' Arabic queries into SPARQL queries and then retrieve relevant verses from the ontology. The obtained results show that our proposal provides heightened precision and recall compared to other search engines.

Recently, a large number of automated applications are developed to improve the retrieval of different types of knowledge. However, there are few automated applications of semantic web technologies (ontology) for the retrieval of Islamic knowledge and in particular for Arabic language, despite the strong demand and need for this knowledge by Muslims and also by non-Muslims. In this paper, we present AraFamOnto, an Arabic ontology-based inheritance calculation system. The use of ontology is becoming increasingly important to store knowledge about the person's family relationships in order to facilitate research, the processing of information about the person and family members, and the calculation of the inheritance of the deceased person's heirs. We present a practical method to limit the time needed to process family data and reduce human effort in the search for family relationships to calculate the Islamic Inheritance correctly.

The way we purchase products online has been revolutionized due to disruptive digital technologies making all the information available at our fingertips. This paper inspects the effect of reviews on mobile phones sales. Noting the inconsistent results on the effect of reviews in previous literature, this study examines how the sentiments on different mobile phone brands in online reviews affect their sales. In our work, we analyze reviews and comments about smartphones from Amazon as data set and classify the reviews text by negative/positive, very negative/very positive, and neutral sentiment. The information contained in reviews is very helpful for both the shoppers and product makers. In this paper, we conclude the following unique characteristics through more than 400,000 real smartphone reviews: (1) Short average length; (2) Power-law distribution; (3) Large span of length; (4) Notable difference in sentiment polarity. Based on the characteristics mentioned above, a series of comparative experiments have been done for sentiment classification and our model achieved an accuracy score of 76.80%

With the availability of text data in various forms on social media platforms, text mining and sentiment analysis have received huge attention. The task of deriving information from this volume of data in order to extract knowledge is very complex and expensive because it is usually unstructured and contains noise. Recently, there is a growing need for implementing various approaches and models for efficiently processing this type of data and extracting useful information. This process is known as sentiment analysis, which includes: data gathering, data pre-processing, feature engineering and labelling, finally the application of various natural language processing and machine learning algorithms. This paper provides an overview of the most recent methods used in text mining and sentiment analysis along with their detailed description and a discussion of obtained results.CCS Concepts

Agriculture is now a trillion-dollar industry, making significant contributions to the growth of several developing as well as developed countries. The huge rise in the growing demand of food and making it sustainable for people is encouraging the need for smart farming. There is a great potential to transform traditional farming profoundly by integrating Internet of things (IoT), Blockchain, and Geospatial technologies to emerge as Smart Farming. Blockchain based farming provides farmers various instant agricultural data at one secured platform, represents a unique opportunity to bring greater efficiency, sustainable crop production, tackle food scarcity, and adds transparency and traceability to the exchange of data related to farming management. Uses of Blockchain in farming management is not only improving the food traceability but also making farming safer for farmers as well as consumers involved, less uncertain and more profitable to the farmers. This paper describes the use of Blockchain technology in farming to manage the practices in a smarter as well as sustainable way, by presenting the decentralized infrastructure with added immutable geospatial technology and IoT sensors capabilities. Our proposed system architecture will explain how Blockchain technology with GIS & IoT will revolutionize the traditional farming practices. Moreover, Blockchain preserves the stakeholders privacy by enhancing IoT framework with more reliable and secure data. Likewise, geospatial technologies create the greater impact by providing visualization and decision making through analytics by transforming traditional farming into sustainable farming.

In recent years, Automatic Natural Language Processing (ANLP) for Arabic language has received a great amount of attention for the development of several applications such as question answering, information retrieval and translation, etc.However, there are a few automated applications using Semantic Web technologies for retrieving Arabic-language documents despite the high demand and need for this content.In addition, the Arabic language presents serious challenges to researchers and developers of NLP applications.These challenges are due to the complexity of the morphological, syntactic and semantic characteristics specific to the Arabic text, which requires the use of semantic resources such as ontology.In our work, we propose a new approach based on ontology and multi-agent systems to index and filter Arabic documents.Our proposal is composed of five layers, each layer contains several agents: (1) Lexical Layer; (2) Syntactic Layer; (3) Semantic Layer; (4) Indexing Layer; and GUI/Interface Layer.Our Arabic ontology is manually constructed on the basis of schemes and their semantics meanings.We use also combination of Arabic WordNet contents and Arabic VerbNet in the process of constructing the ontology.We use the semantic similarity to find the relevant documents according to the user's queries.The aim of this paper is to study the effect of patterns in solving the problem of the semantic indexing system (SIS).The main objective is to improve the quality of the indexing process to ensure the accuracy of the information search of relevant documents based on us ers' multiword queries, and also to reduce indexing and search time.Indeed, our experiments are conducted on the basis of the combination of two Arab corpus: OSAC and SemEval.We compared our results Ontological Approach Based on Multi-Agent System for Indexing and Filtering Arabic Documentswith Lucene in-dex for the same data and, we found that our approach achieves much better results than the other.

With the innovation of new Information and Communication Technologies and the needs of information and knowledge sharing among the city, a smart city system aims to improve it's citizens life quality by offering a set of public services. Within this context, this research work proposes a smart city approach for using these technologies in public services delivery to allow real-time interactions with citizens; we propose a service providing framework based on context-aware recommendation approach to improve the city's digital services according to citizen’s context and the backend as a service (BAAS) approach on cloud to enhance the scalability of the system in large workloads. In this study, we evaluated the scalability and interactivity of the proposed approach by measuring response time and exchanged data metrics. To estimate the scalability, we performed series of load tests where response time values increased with the addition of requests in different cases but remained acceptable. In terms of interactivity, the exchanged data between the citizen application and the city’s backend was measured at maximum by 117 Kb/s. The experiments revealed that context-aware recommendation approach optimised interactions by reducing the amount of exchanged data, and the BAAS approach improved the scalability of the system and allowed handling concurrent requests. GRAPHICAL ABSTRACT

Web services are meaningful only if potential users may find and execute them. Universal description discovery and integration (UDDI) help businesses, organizations, and other web services providers to discover and reach to the service(s) by providing the URI of the WSDL file. However, it does not offer a mechanism to choose a web service based on its quality. The standard also lacks sufficient semantic description in the content of web services. This lack makes it difficult to find and compose suitable web services during analysis, search, and matching processes. In addition, a central UDDI suffers from one centralized point problem and the high cost of maintenance. To get around these problems, the authors propose in this chapter a novel framework based on mobile agent and metadata catalogue for web services discovery. Their approach is based on user profile in order to discover appropriate web services, meeting customer requirements in less time and taking into account the QoS properties.

Under the impact of the innovating Information and Communication Technologies (ICT) in the smart city context and the needs of information sharing and digital services, a smart city system aims to improve the life quality of its citizens by offering a set of public services. The aim of this research work, is to propose a smart city approach that exploits These technologies in the delivery of public services to allow greater real-time interaction with citizens, this research contributes to the domain of smart city public services delivery by treating the scalability problem faced by these systems; we propose a service providing framework based on the backend as a service approach on cloud to improve the quality of the smart city digital services according to the citizen's needs while enhancing the scalability of the system in large workloads. For illustrating the functioning of the approach we have used information about the city of Biskra administrative institutions as an application example of the system.

The increasing deployment of information and communication technologies in our daily life allows administrators to manage their work with more flexibility and comfortability. In this paper, we present our work, which takes advantage of different technologies such as ontology and agent system for smart school management, especially in the classroom. This work is designed to help administrators to identify any activities that occur at the classroom level, such as tracking the attendance of the student or teacher. We use the ontology to model the context in an intelligent environment for facilitating information integration and knowledge sharing between heterogeneous knowledge and information sources in smart school. In this paper, we describe the interaction of intelligent multi-agent systems running on network, sensor networks and ontologies into the architecture aiming at increasing the management capabilities in smart school and generate automatically detailed reports of the environment with minimal time and effort.