PULLAR – Pull-based Unified Learning and Labeling for Annotated Research

PurposeThe purpose of this work is to explore the new possibilities enabled by the recent introduction of RDF-star, an extension that allows for statements about statements within the Resource Description Framework (RDF). Alongside Named Graphs, this approach offers opportunities to leverage a meta-level for data modeling and data applications.Design/methodology/approachIn this extended paper, the authors build onto three modeling use cases published in a previous paper: (1) provide provenance information, (2) maintain backwards compatibility for existing models, and (3) reduce the complexity of a data model. The authors present two scenarios where they implement the use of the meta-level to extend a data model with meta-information.FindingsThe authors present three abstract patterns for actively using the meta-level in data modeling. The authors showcase the implementation of the meta-level through two scenarios from our research project: (1) the authors introduce a workflow for triple annotation that uses the meta-level to enable users to comment on individual statements, such as for reporting errors or adding supplementary information. (2) The authors demonstrate how adding meta-information to a data model can accommodate highly specialized data while maintaining the simplicity of the underlying model.Practical implicationsThrough the formulation of data modeling patterns with RDF-star and the demonstration of their application in two scenarios, the authors advocate for data modelers to embrace the meta-level.Originality/valueWith RDF-star being a very new extension to RDF, to the best of the authors’ knowledge, they are among the first to relate it to other meta-level approaches and demonstrate its application in real-world scenarios.

Data modeling is a complex task for novice designers. This paper conducts a systematic study of cognitive complexity to reveal important factors pertaining to data modeling. Four major sources of complexity principles are identified: problem solving principles, design principles, information overload, and systems theory. The factors that lead to complexity are listed in each category. Each factor is then applied to the context of data modeling to evaluate if it affects data modeling complexity. Redundant factors from different sources are ignored, and closely linked factors are merged. The factors are then integrated to come up with a comprehensive list of factors. The factors that cannot largely be controlled are dropped from further analysis. The remaining factors are employed to develop a semantic differential scale for assessing cognitive complexity. The paper concludes with implications and recommendations on how to address cognitive complexity caused by data modeling.

Technological advances led to the generation of large scale complex data. Thus, extraction and retrieval of information to automatically discover latent pattern have been largely studied in the various domains of science and technology. Consequently, machine learning experienced tremendous development and various statistical approaches have been suggested. In particular, data clustering has received a lot of attention. Finite mixture models have been revealed to be one of the flexible and popular approaches in data clustering. Considering mixture models, three crucial aspects should be addressed. The first issue is choosing a distribution which is flexible enough to fit the data. In this paper, a model based on multivariate Beta distributions is proposed. The two other challenges in mixture models are estimation of model's parameters and model complexity. To tackle these challenges, variational inference techniques demonstrated considerable robustness. In this paper, two methods are studied, namely, batch and online variational inferences and the models are evaluated on four medical applications including image segmentation of colorectal cancer, multi‐class colon tissue analysis, digital imaging in skin lesion diagnosis and computer aid detection of Malaria.

Considerable attention in the information systems and management science literature has focused on computer-based modeling environments, sometimes called integrated modeling environments or model management systems. This research has been primarily concerned with suggesting features/components of modeling environments such as improved executable modeling languages for model creation, integration, and data representation; specialized database systems for managing model data; and customized model-solver software. However, there has been little (if any) empirical guidance offered in the literature about the specific needs of business and industry for computer-based integrated modeling environments. Using a data set compiled from a national survey of modelers (analysts) and model users (decision makers), we empirically investigate the validity of several of the key assumptions of modeling environment research reported in the literature, and examine the relationships between the modeling factors: data complexity, model complexity, modeling intensity, modeler/user requirements, and need for computer-based integrated modeling environments in organizations. Our empirical analysis of the data set shows that practitioners rank automated access to model data and automated error checking (e.g., model syntax and semantics checking) high as desirable components in modeling environments. We find that users prefer to have modeling environments linked to their current modeling and modeling-support software systems. Our findings further suggest that a high percentage of modelers and users are dissatisfied with the software systems they are currently using to support their modeling activities. Finally, a covariance structure analysis of the modeling environment factors clearly shows that: (a) model complexity has a direct positive effect on modeling intensity; (b) data complexity has an insignificant direct effect on modeling intensity, but has a negative effect on modeler/user requirements; and (c) modeler/user requirements have a direct positive effect on need for computer-based integrated modeling environments in organizations.

Developmental cascade models provide a valuable framework for understanding how risk and protective factors interact over time to shape health and behavioral outcomes. Traditional statistical methods, such as logistic regression and structural equation modeling, have been instrumental in uncovering developmental pathways within prevention science. However, these methods often impose constraints on model complexity and face limitations in capturing the non-linear and interdependent nature of developmental processes. Machine learning (ML) offers complementary advantages, such as the ability to incorporate high-dimensional data, detect complex interactions, and enhance predictive accuracy. These capabilities can improve identification of at-risk individuals, support the timing of interventions across developmental stages, and refine theory-driven models. By integrating ML with developmental cascade models, researchers can more effectively identify when and how which risk accumulates and protective factors exert influence, thereby improving the tailoring and efficiency of prevention strategies. This conceptual paper outlines how ML can extend traditional analytic approaches in developmental cascade research, discusses key practical considerations for researchers including data requirements, software selection, and model validation, and highlights its potential to advance prevention science across the life course.

Maximizing supply chain performance leveraging machine learning to anticipate customer backorders

A comparison of machine learning surrogate models for net present value prediction from well placement binary data

Oilseed rape (Brassica napus L.) is one of the most important oil crops worldwide. Machine learning metchods are gaining increasing im­portance in oilseed rape production, enabling yield prediction, crop condition assessment, and support for agronomic decision-making. This paper presents a review of studies on the use of various algorithms – from classical models such as linear regression, decision trees, random forest, and support vector machine to more advanced ones, including neural networks. Applications of remote sensing data, unnamed aerial vehicles and satellite imagery, as well as ground-based sensors for monitoring diseases, environmental stresses, and nutrient requirements are discussed. The potential of machine learning to enhance efficiency and sustainability in rapeseed production is highlighted, along with challenges related to data quality, model complexity, and interpretability.

This review examines the increasing application of artificial intelligence (AI) and/or machine learning (ML) in microalgae processes, focusing on their ability to improve production efficiency, yield, and process control. AI/ML technologies are used in various aspects of microalgae processes, such as real-time monitoring, species identification, the optimization of growth conditions, harvesting, and the purification of bioproducts. Commonly employed ML algorithms, including the support vector machine (SVM), genetic algorithm (GA), decision tree (DT), random forest (RF), artificial neural network (ANN), and deep learning (DL), each have unique strengths but also present challenges, such as computational demands, overfitting, and transparency. Despite these hurdles, AI/ML technologies have shown significant improvements in system performance, scalability, and resource efficiency, as well as in cutting costs, minimizing downtime, and reducing environmental impact. However, broader implementations face obstacles, including data availability, model complexity, scalability issues, cybersecurity threats, and regulatory challenges. To address these issues, solutions, such as the use of simulation-based data, modular system designs, and adaptive learning models, have been proposed. This review contributes to the literature by offering a thorough analysis of the practical applications, obstacles, and benefits of AI/ML in microalgae processes, offering critical insights into this fast-evolving field.

Multivariate density estimation is a fundamental problem in Applied Statistics and Machine Learning. Given a collection of data sampled from an unknown distribution, the task is to approximately reconstruct the generative distribution. There are two different approaches to the problem, the parametric approach and the non-parametric approach. In the parametric approach, the approximate distribution is represented by a model from a predetermined family. In this thesis, we adopt the parametric approach and investigate the use of a model family called latent tree models for the task of density estimation. Latent tree models are tree-structured Bayesian networks in which leaf nodes represent observed variables, while internal nodes represent hidden variables. Such models can represent complex relationships among observed variables, and in the meantime, admit efficient inference among them. Consequently, they are a desirable tool for density estimation. While latent tree models are studied for the first time in this thesis for the purpose of density estimation, they have been investigated earlier for clustering and latent structure discovery. Several algorithms for learning latent tree models have been proposed. The state-of-the-art is an algorithm called EAST. EAST determines model structures through principled and systematic search, and determines model parameters using the EM algorithm. It has been shown to be capable of achieving good trade-off between fit to data and model complexity. It is also capable of discovering latent structures behind data. Unfortunately, it has a high computational complexity, which limits its applicability to density estimation problems. In this thesis, we propose two latent tree model learning algorithms specifically for density estimation. The two algorithms have distinct characteristics and are suitable for different applications. The first algorithm is called HCL. HCL assumes a predetermined bound on model complexity and restricts to binary model structures. It first builds a binary tree structure based on mutual information and then runs the EM algorithm once on the resulting structure to determine the parameters. As such, it is efficient and can deal with large applications. The second algorithm is called Pyramid. Pyramid does not assume predetermined bounds on model complexity and does not restrict to binary tree structures. It builds model structures using heuristics based on mutual information and local search. It is slower than HCL. However, it is faster than EAST and is only slightly inferior to EAST in terms of the quality of the resulting models. In this thesis, we also study two applications of the density estimation techniques that we develop. The first application is to approximate probabilistic inference in Bayesian networks. A Bayesian network represents a joint distribution over a set of random variables. It often happens that the network structure is very complex and making inference directly on the network is computational intractable. We propose to approximate the joint distribution using a latent tree model and exploit the latent tree model for faster inference. The idea is to sample data from the Bayesian network, learn a latent tree model from the data offline, and when online, make inference with the latent tree model instead of the original Bayesian network. HCL is used here because the sample size needs to be large to produce accurate approximation and it is possible to predetermine a bound on the online running. Empirical evidence shows that this method can achieve good approximation accuracy at low online computational cost. The second application is classification. A common approach to this task is to formulate it as a density estimation problem: One constructs the class-conditional density for each class and then uses the Bayes rule to make classification. We propose to estimate those class-conditional densities using either EAST or Pyramid. Empiricalevidence shows that this method yields good classification performances. Moreover, the latent tree models built for the class-conditional densities are often meaningful, which is conducive to user confidence. A comparison between EAST and Pyramid reveals that Pyramid is significantly more efficient than EAST, while it results in more or less the same classification performance as the latter.

Groundwater, a vital resource for various human activities and ecosystems, necessitates efficient management and sustainable utilisation. Groundwater potential zone mapping plays a pivotal role in identifying areas where groundwater resources are abundant, thereby aiding decision makers in optimal resource allocation. This review paper presents an in-depth analysis of diverse methods employed for groundwater potential zone mapping, offering a comprehensive overview of their strengths, weaknesses and recent advancements. The review covers traditional methods rooted in hydrogeological principles, as well as modern techniques that harness the power of geospatial technologies and machine learning. Furthermore, the paper explores the integration of remote sensing and geographic information systems for spatial data analysis, emphasising their role in enhancing the accuracy of potential zone mapping. In the context of recent advancements, the review sheds light on the emergence of hybrid methods that combine the strengths of multiple approaches, resulting in improved prediction accuracy and robustness. Challenges associated with each method, such as data quality, model complexity and interpretability, are critically examined, providing insights into the potential limitations and avenues for improvement. The review also emphasises the importance of validation and uncertainty assessment, ensuring the reliability of potential zone mapping results. Finally, this review paper serves as a comprehensive guide for researchers, practitioners and policymakers engaged in groundwater resource management. By offering a holistic understanding of the diverse methods available for groundwater potential zone mapping, this paper contributes to informed decision making and the advancement of sustainable groundwater-management practices.

Network Traffic Classification (NTC) plays a critical role in modern network security and management, used in various applications such as performance monitoring, anomaly detection, and bandwidth management. Traditional methods like port-based identification and deep packet inspection (DPI) face limitations due to evolving encryption techniques and dynamic port usage. Recent advances in machine learning have improved NTC but often struggle with data labeling, encrypted traffic, and model complexity. This paper introduces a novel Information Bottleneck Neural Network (IBNN) to address these challenges by efficiently extracting essential features from raw traffic data while minimizing irrelevant information. The IBNN framework balances model complexity and discriminative feature preservation, leading to enhanced classification accuracy and robustness, particularly for encrypted traffic. Through information-centric feature extraction and relevant information compression, the proposed method outperforms existing techniques about 5–7% in terms of accuracy, as well as F1 score on both ISCX 2012 and UNB CIS datasets, delivering superior resilience and adaptability across diverse network scenarios. Extensive experiments demonstrate the efficacy of IBNN, positioning it as a significant advancement in the field of NTC.

Traffic flow prediction is crucial in transportation management and optimizing control strategies and infrastructure planning. However, traditional prediction methods often need help handling traffic data's inherent uncertainty and complexity. This paper explores the application of Grey prediction models as a promising approach for efficient traffic flow prediction. Grey prediction models, rooted in Grey systems theory, offer a robust framework for modelling systems with incomplete and uncertain information. The paper provides a comprehensive overview of Grey prediction models, including the Grey Verhulst, Grey-Markov, and Grey differential equation models. It discusses the process of data collection, pre-processing, and model application in traffic flow prediction. Case studies and experimental results demonstrate the effectiveness of Grey prediction models in accurately forecasting traffic flow patterns. Despite challenges such as data sparsity and model complexity, Grey prediction models offer significant advantages over traditional methods, including improved prediction accuracy and adaptability to dynamic traffic conditions. The paper concludes with recommendations for practitioners and researchers interested in leveraging Grey prediction models for efficient traffic management.

The so-called Energiewende is a complex task with a variety of stakeholders, regulations, technical infrastructure, and proposed solutions. Buildings are an important sector for increasing resource and energy efficiency, as in Germany around 35% of end energy usage can be attributed to them. Digital applications can help reduce these emissions through more efficient planning, operating, renovation, or demolition. Depending on the task and the parties involved, the complexity of descriptions and data models can vary from simple energy efficiency labels used in the labeling process of buildings to complex individual data models used for simulation. However, these specific information systems are often non-transferable, are hard to compare in their restrictions and requirements, and hence increase the overall costs of digital solutions. A good understanding is needed to preserve economic and ecological benefits while maintaining privacy and security aspects. To compare digital applications, a variety of solutions are identified. This paper first provides an overview of the identified solutions, namely frameworks, taxonomies, and ontologies. While the identified frameworks focus more on technological aspects and are complex to use, they provide an in-depth understanding. Taxonomies can be used for a scientific (e.g., classification of methods) comparison and provide simple relationships. Ontologies provide relationships and definitions while being task-dependent. We conclude that comparing the underlying data models of digital applications is a complex task and dependent on the application and its infrastructure. However, a variety of tasks refers to the same tools and data. After discussing these approaches, we then give an overview of digital applications developed by German researchers. Last, we give an insight on how to combine these aspects in our ongoing research. To summarize, in this paper we give an overview of the complexity involved in transferring digital solutions in the building sector, provide a method used for comparing applications and describe a solution to compare infrastructure and digital tools built for it.

Arithmetic coding (AC) is widely used for lossless data compression, and parallelization of arithmetic coding is relatively simple because all symbols can be encoded independently. On the other hand, Parallel adaptive arithmetic coding stores the data model for each data block (the model contains the data structure used to calculate the probability distribution of the data) and these data models need to be frequently accessed and modified during encoding and decoding, so the access latency and computational complexity of the data model are important factors affecting performance. In this paper, we present a new data structure called partial prefix sum array which can quickly calculate the probability distribution of the data, so that parallel adaptive arithmetic encoding and decoding can be accelerated. Furthermore, we use a mode of coalescing access to access global memory, thereby improving the throughput of global memory. Experimental results for 6 files on NVIDIA Tesla M60 GPU show that our GPU adaptive arithmetic encoding and decoding run 1.61x- 2.75x times and 1.03x-2.22x times faster than previously presented GPU implementation, respectively.