AdaptiveSystems: An Integrated Framework for Adaptive Systems Design and Development Using MPS JetBrains Domain-Specific Modeling Environment

There is growing interest in applying artificial intelligence (AI) to automate and support complex decision-making tasks. However, it remains unclear how algorithms compare to human judgment in contexts requiring semantic understanding and domain expertise. We examine this in the context of the judge assignment problem, matching submissions to suitably qualified judges. Specifically, we tackled this problem at the Harvard President’s Innovation Challenge, the university’s premier venture competition awarding over $500,000 to student and alumni startups. This setting represents a real-world environment where high-quality judge assignment is essential. We developed an AI-based judge assignment algorithm, the Hybrid Lexical-Semantic Similarity Ensemble (HLSE), and deployed it at the competition. We then evaluated its performance against human expert assignments using blinded match-quality scores from judges on 309 judge-venture pairs. Using a Mann-Whitney U statistic-based test, we found no statistically significant difference in assignment quality between the two approaches (AUC=0.48, p=0.40); on average, algorithmic matches were rated 3.90 and manual matches 3.94 on a 5-point scale, where 5 indicates an excellent match. Furthermore, manual assignments that previously required a full week could be automated in several hours by the algorithm during deployment. These results demonstrate that HLSE achieves human-expert-level matching quality while offering greater scalability and efficiency, underscoring the potential of AI-driven solutions to support and enhance human decision-making for judge assignment in high-stakes settings.

The agricultural sector is rapidly evolving through digital technologies, creating significant opportunities to apply Artificial Intelligence (AI) for improving crop productivity, reducing losses, and optimizing resource utilization. This review specifically examines two key challenges in modern agriculture: the timely and accurate detection of crop diseases and the generation of precise fertilizer recommendations. We present a structured analysis of recent deep learning advancements, focusing on computer vision techniques such as Convolutional Neural Networks (CNNs), Vision Transformers (ViTs), and Generative Adversarial Networks (GANs) for image-based disease diagnosis, as well as NLP and knowledge-graph approaches for integrating agronomic information. Additionally, we evaluate data-driven fertilizer recommendation frameworks that incorporate soil characteristics, climatic factors, and crop growth patterns using hybrid deep learning and ensemble models. The review also explores the role of multimodal learning, IoT-based sensing, and cloud–edge computing in enabling real-time agricultural decision-making. Finally, we highlight current limitations—including dataset scarcity, generalization issues, explainability gaps, and scalability concerns—and outline future research directions for building intelligent, interpretable, and adaptive AI systems for sustainable agriculture.

Recently, many models of applying artificial intelligence (AI) techniques into the analysis of clinical data have been proposed. Unfortunately, most models provide little help when specific "cause–effect" relation of data is not available, or even known. In this paper, an innovative method, called closest reasonable centroids (CRC), is directed to address this issue. Our present application domain was a clinical data set of the weight changes of 274 prematurely born babies who had nutritional deficiency problem and were given TPN treatments to improve their nutritional needs. Experimental result shows that the CRC's differentiability is comparable to that of the back-propagation neural networks (BPN) and better than that of statistical method. Also, from the health conditions of babies and their nutritional treatments, the proposed method can roughly predict their weight changes and provide some suggested feasible formula. All of the above results have been double confirmed by the clinicians, implicating that CRC could be used as assistant tool.

Risk analysis is process to identify and develop profile of risk. Most risk analysis procedures focuses in technical perspectives, whereas business perspective is less involved. Business impacts are very important to be considered because it is related to business continuity of organization. Therefore, identification of business aspects is needed to be done so that aspects can be used as metrics in risk analysis of IT security threats. This paper proposes metrics of risk analysis in IT security threats based on business aspects. Methods for validating business aspects as metrics are literature analysis, linear regression analysis, correlation analysis and significancy analysis. The defined metrics will be used to develop risk analysis procedure in future work so it is expected to result better output in risk profile.

Open adaptive systems are the basis for a promising new generation of embedded systems with huge economic potential. In many application domains, however, the systems are safety-critical and an appropriate safety assurance approach is still missing.In recent years, models at runtime have emerged as a promising way to systematically engineer adaptive systems. This approach seems to provide the indispensable leverage for applying safety assurance techniques in adaptive systems. Therefore, this survey analyzes the state-of-the-art of models at runtime from a safety engineering point of view in order to assess the potential of this approach and to identify open gaps that have to be closed in future research to yield a safety assurance approach for open adaptive systems.KeywordsAdaptive SystemSafety EngineeringFault Tree AnalysisSafety CaseSafety GoalThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

The growing number of available data graphs in the form of RDF Linked Data enables the development of semantic exploration applications in many domains. Often, the users are not domain experts and are therefore unaware of the complex knowledge structures represented in the data graphs they interact with. This hinders users’ experience and effectiveness. Our research concerns intelligent support to facilitate the exploration of data graphs by users who are not domain experts. We propose a new navigation support approach underpinned by the subsumption theory of meaningful learning, which postulates that new concepts are grasped by starting from familiar concepts which serve as knowledge anchors from where links to new knowledge are made. Our earlier work has developed several metrics and the corresponding algorithms for identifying knowledge anchors in data graphs. In this paper, we assess the performance of these algorithms by considering the user perspective and application context. The paper address the challenge of aligning basic level objects that represent familiar concepts in human cognitive structures with automatically derived knowledge anchors in data graphs. We present a systematic approach that adapts experimental methods from Cognitive Science to derive basic level objects underpinned by a data graph. This is used to evaluate knowledge anchors in data graphs in two application domains - semantic browsing (Music) and semantic search (Careers). The evaluation validates the algorithms, which enables their adoption over different domains and application contexts.

With the ever-increasing amount of data, the world has stepped into the era of “Big Data”. Presently, the analysis of massive and complex data and the extraction of relevant information, have been become essential tasks in many fields of studies, such as health, biology, chemistry, social science, astronomy, and physics. However, compared with the development of data storage and management technologies, our ability to gain useful information from the collected data does not match our ability to collect the data. This gap has led to a surge of research activity in the field of visual analytics. Visual analytics employs interactive visualization to integrate human judgment into algorithmic data-analysis processes. In this paper, the aim is to draw a complete picture of visual analytics to direct future research by examining the related research in various application domains. As such, a novel categorization of visual-analytics applications from a technical perspective is proposed, which is based on the dimensionality of visualization and the type of interaction. Based on this categorization, a comprehensive survey of visual analytics is performed, which examines its evolution from visualization and algorithmic data analysis, and investigates how it is applied in various application domains. In addition, based on the observations and findings gained in this survey, the trends, major challenges, and future directions of visual analytics are discussed.

In many situations, it is of interest for authentication systems to adapt to context (e.g., when the user’s behavior differs from the previous behavior). Hence, representing the context with appropriate and well-designed models is crucial. We provide a comprehensive overview and analysis of research work on C ontext M odelling f or A daptive A uthentication systems (CM4AA). To this end, we pursue three goals based on the Systematic Mapping Study (SMS) and Systematic Literature Review (SLR) research methodologies. We first present a SMS to structure the research area of CM4AA ( goal 1 ). We complement the SMS with an SLR to gather and synthesise evidence about context information and its modelling for adaptive authentication systems ( goal 2 ). From the knowledge gained from goal 2, we determine the desired properties of the context information model and its use for adaptive authentication systems ( goal 3 ). Motivated to find out how to model context information for adaptive authentication, we provide a structured survey of the literature to date on CM4AA and a classification of existing proposals according to several analysis metrics. We demonstrate the ability of capturing a common set of contextual features that are relevant for adaptive authentication systems independent from the application domain. We emphasise that despite the possibility of a unified framework, no standard for CM4AA exists.

The paper presents three large projects that were developed using OPEN Global Research and Development Process over a 5 year span. The three projects are significantly different in the domain of applicability and scope but were all developed using the same process for OPEN Global collaboration. The OPEN Global Collaboration process is based on the identification of global competence centers staffed with domain experts that are managed by a central team responsible for high-level requirements, overall architecture and quality assurance. The key enabler for OPEN Global Collaboration process is the selection of the domain experts of the global competence centers. These domain experts are internationally recognized experts and they are responsible for the complete specification and development of the components under their responsibility. This is a key insight as it signficantly reduces the required communication between the central and remote teams. As a consequence, the communication is centered on the interface between components and not on the detailed component specifications.

Motivated by the complex production management with difficulties in error-prone assembly system and inaccurate supply chain inventory, this paper designs a novel manufacturing execution system (MES) architecture for intelligent monitoring based on wireless sensor network (WSN). The technical perspective includes analysis on the proposed manufacturing resource mutual inductance method under active sensing network, appreciation technology of multisource information, and dynamic optimization technology for manufacturing execution processes. From business perspective, this paper elaborates the impact of RFID investment on complex product by establishing a three-stage supply chain model that involves two suppliers carrying out Stackelberg games (manufacturer and retailer). The optimal cost threshold values of technology investment are examined for both the centralized and the decentralized scenarios utilizing quantitative modeling methods. By analyzing and comparing the optimal profit with or without investment on WSN, this paper establishes a supply chain coordination and boosting model. The results of this paper have contributed significantly for one to make decision on whether RFID should be adopted among its members in supply chain. The system performance and model extension are verified via numerical analyses.

easyABMS: A domain-expert oriented methodology for agent-based modeling and simulation

Compensating for the loss of future tree values in the model of Fuzzy knowledge units

Studies in the field of temperature fluctuations resistance are promising for agrobiology since these changes largely determine distribution and timing of plant vegetation. Contemporary methods have significantly expanded the possibilities in research of adaptive systems: molecular reactions and immunochemical methods lead to specify the order and degree of genes expression and presence of certain proteins in tissues at a particular stage, hence steadily increasing the number of participants in the response to temperature shocks.

Due to a sudden increase of data when internet of things, which uses sensor network such as RFID and NFC, was applied, interest in collection, processing, analysis and utilisation of big data is increasing. These big data, which are mostly unstructured data, are showing a great degree of difference in the aspects of data movement speed, variety, speed, and quantity when compared to the existing typical relational data. Especially, as an open source platform, the Hadoop system, which is used for processing of big data that occurs in logistics which is a major utilisation field of internet of things in the technical perspective, is providing opportunities for anyone to build a platform with comparatively lower cost. This thesis attempts to provide data that can be referred to in introducing platform that is adequate to our company by designing various Hadoop platforms that can be utilised in processing big data which occurs in logistics based on IoT and conducting comparative evaluation of its performance. Also, we present the integrated verification results for the data analysis performance of the analysis software system running on top of realised prototypes by employing American Standard Association (ASA) dataset.

During last years Adaptive Automation (AA) has received considerable attention in the academic community, in labs, in technology companies, because of the large use of automation in several domains (e.g. aviation; manufacturing; medicine; road, rail, and maritime transportation). AA is a potential solution to the problems associated with humanautomation interaction, regardless of the complexity of the application domain. The adaptive automation concept was firstly proposed about 30 years ago (Rouse, 1976), but technology has provided the empirical evidence of its effectiveness in more recent times. Several studies have shown that adaptive systems can regulate operator workload and enhance performance while preserving the benefits of automation (Hilburn et al., 1997; Kaber & Endsley, 2004; Moray et al., 2000; Prinzel et al., 2003). Still, inappropriate design of adaptive systems may even bring to a worse performance than full manual systems (Parasuraman et al., 1999). Therefore, methods and skills for designing adaptive automation systems should be fully mastered, before taking the implementation step. Before analyzing the concept and impact of adaptation features, the meaning and definition of automation should be defined, in order to stress how human characteristics and limitations influence the use (or misuse) of automation Andy Clark (2003) successfully tried to summarize the technological relationship between human and automated system: “humans have always been adept at dovetailing our minds and skills to the shape of our current tools and aids. But when those tools and aids start dovetailing back – when our technologies actively, automatically, and continually tailor themselves to us just as we do to them – then the line between tool and human becomes flimsy indeed”. Several researchers tried to define what adaptive automation is, using different and complementary concepts. In order to achieve a complete, if it is possible, definition of adaptive automation, it is needed to analyse the definition and the implication of automation itself. AA aims at optimizing the cooperation and at efficiently allocating labor between an automated system and its human users (Kay, 2006) and it can be considered as an