Prior Knowledge Research Articles

The relation of complex problem solving with reflective abstraction: a systematic literature review

<span lang="EN-US">Complex problem solving (CPS) is a new paradigm in solving problems and is one of the soft skills needed to face the industrial revolution 4.0. Reflective abstraction is associating and modifying pre-existing conceptions into new situations. This article reviews research on CPS and reflective abstraction. This research is needed to know the relationship between reflective abstraction and CPS. The systematic writing of this review was assisted by the Publish or Perish 7 application, Mendeley, and VOSviewer. A literature search was performed through the ScienceDirect and ERIC databases. Based on the search results with the term “complex problem solving” and several exclusion criteria, 58 articles were found, whereas with the word “reflective abstraction” there were 23 articles, totaling 81 papers. Based on the literature review, it was found that there is a relationship between CPS and reflective abstraction by obtaining common ground in the form of prior knowledge. CPS requires prior knowledge from reflective abstraction to integrate the most relevant information. To improve CPS, efforts and special attention can be made to build initial knowledge through reflective abstraction. This article contributes to further research and becomes a study for the themes of CPS and reflective abstraction in learning and education.</span>

Towards a disease prediction system: BioBERT-based medical profile representation

<p>Healthcare professionals are increasingly interested in predicting diseases before they manifest, as this can prevent more serious health conditions and even save lives. Machine learning techniques are now playing an important role in healthcare, including in the early prediction of diseases based on prior medical knowledge. However, one of the biggest challenges is how to represent medical information in a way that can be processed by machine learning algorithms. Medical histories are often in a format that computers cannot read, so filtering and converting this information into numerical representations is a crucial step. This process has become easier with the advancement of natural language processing techniques. In this paper, we propose three representations of medical information, two of which are based on BioBERT, the latest text representation techniques for the biomedical sector. The efficiency of these representations is tested on the MIMIC-III database, which contains information on 46,520 patients. The focus of the study is on predicting Coronary Artery Disease, and the results demonstrate the effectiveness of the proposed approach. The study highlights the importance of medical history in disease prediction and demonstrates the potential of machine learning techniques to advance healthcare.</p>

Consumer preference and willingness to pay for 3D-printed chocolates: A discrete choice experiment

This study explores the emerging field of 3D food printing, specifically examining consumer preferences for 3D-printed chocolate bars. While 3D printing has revolutionized various industries, the consumer acceptance of 3D-printed food remains a relatively underexplored topic. A discrete choice experiment was conducted to investigate consumer preferences and willingness to pay for 3D-printed chocolate bars in the UK. Key attributes included in the study were price, chocolate type, and the use of 3D printing technology. Using a mixed logit model, preference heterogeneity was assessed, revealing a general inclination towards 3D-printed chocolates, with an estimated increase in willingness to pay of £0.21 per 100g. Furthermore, the influence of individual (participants’) characteristics, such as age, gender, prior knowledge of 3D food printing, and food technology neophobia, on perceived benefits associated with this technology and willingness to pay was also explored, using a structural equation modeling analysis. The results indicated that only perceived hedonic benefits, such as improved texture, positively impacted the willingness to pay, while food technology neophobia acted as a significant barrier. The results of this study therefore contribute valuable insights into the consumer acceptance of 3D-printed foods, specifically chocolates, highlighting potential avenues for market adoption and emphasizing the importance of effective communication in shaping consumer attitudes toward novel food processing technologies.

PROFESSIONAL MOTIVATION OF STUDENTS AS A FACTOR FOR INTEGRATING FUNDAMENTAL AND CLINICAL DISCIPLINES IN HIGHER MEDICAL EDUCATION

This article addresses the pressing issue of medical education: the integration of fundamental and clinical disciplines. Given the rapid pace of scientific and technological advancements, the professionalism of modern doctors largely relies on their natural and scientific training, which begins in the early stages of medical university education. Medical and biological physics play an important role in this training, but integrating them into medical education is challenging due to the complexity of the discipline's content, which students must grasp at the outset of their education, and the lack of understanding among first-year students about the connections between physics, mathematics, and medicine. The article discusses various approaches to teaching medical and biological physics, focusing on creating professional motivation among students to master fundamental sciences. It highlights the example of designing an X-ray computed tomography (CT) scanner as a means of motivating students to excel in their studies. The importance of leveraging students' prior knowledge from school textbooks on physics, biology, and chemistry to solve tasks fundamental to clinical practice is emphasized. Using the practical lesson topic “Physical foundations of X-ray diagnostics and X-ray therapy” as an example, the article demonstrates the need to focus on the properties of X-ray radiation relevant to medical applications, such as rheography, fluorography, and CT scans. Professional motivation is further enhanced through activities like participating in a biophysics scientific circle and visiting university clinics and training centers, where students can witness the practical application of their theoretical knowledge. The article underscores the importance of integrating fundamental and clinical disciplines while adhering to pedagogical principles such as scientific rigor, continuity, and consistency.

Learning cooperative strategies in multi-agent encirclement games with faster prey using prior knowledge

Learning cooperative strategies in multi-agent encirclement games with faster prey using prior knowledge

How pre-trained large model can help, when SAM meets image restoration

In the process of image capture, degradation is inevitable due to noise, motion, down-sampling and so on. Therefore, image restoration is essential to improve the quality of images to enhance their visual effects and benefit downstream tasks. Adding prior knowledge can help the model better understand the image content and restoration requirements, thus improving the quality and efficiency. Semantic-level prior information can be generated by pre-trained large-scale models, such as segment anything models (SAM), and applied to a large number of downstream tasks. SAM has demonstrated powerful robustness and stability in restoration tasks, such as denoising, super-resolution, low-light enhancement, etc. Meanwhile, as an interactive component, SAM brings more control for users during the repair process. In this paper, we focus on the importance of SAM as prior information and systematically summarize a series of recent works combining SAM prior and low-level image restoration from three perspectives. In addition, we have summarized some potential problems and future directions of SAM.

Adaptive impact analysis in flexible multibody systems based on hierarchically refined IGA models

AbstractUsually, detailed impact simulation models within flexible multibody systems have to be set up manually rather than being generated automatically. This is because the process requires prior knowledge of the time and location of the impact, as well as the element resolution within the contact area. If the penalty method is used to determine the occurring contact forces, the corresponding penalty factor also needs to be determined manually. This work, however, presents an adaptive algorithm to simulate impacts within flexible multibody systems fully automatically using reduced isogeometric analysis models, the floating frame of reference formulation, and quasistatic contact models for an efficient but still accurate simulation. The adaptive algorithm detects impacts in the system, determines the contact locations on the bodies, refines the contact area, and determines the penalty factor, and therefore automatically simulates impacts. The work shows how to automatically simulate impacts in flexible multibody systems without user action or prior knowledge of impact location and size. The first application example simulates significant elastodynamic effects within a long flexible rod. The goal is to validate the algorithm by preserving the wave propagation and energy of the system. The second application example simulates the impacts of two flexible double pendulums. This setup is a suitable benchmark for the complete adaptive impact analysis procedure as the flexible double pendulums undergo large rigid body motions.

Capturing the situated, dynamic nature of EMI-lecture listening comprehension in real time

Abstract English-medium instruction (EMI) potentially offers pedagogical efficiency by simultaneously providing access to academic content and English affordances. EMI’s efficacy and effectiveness, however, remain unproven with questions concerning students’ language proficiency unresolved. As listening to lectures holds a central role in tertiary education and the development of language skills is founded on second-language listening ability, key concerns for stakeholders are knowing how much of an EMI lecture students can understand and what factors impact their lecture listening comprehension. Current methodologies used to capture listening comprehension data, such as tests, surveys, summaries, and transcript markings do not capture the idiosyncratic, volatile, and multifaceted reactions to aural text phenomena that learners encounter during a lecture. This study, therefore, uses an innovative foot switch mechanism to capture comprehension data in real time as students participated in an EMI lecture. These data were then used to guide a stimulated recall. The data showed learners failing to comprehend extensive sections of the lecture while deeper analysis identified sections of the lecture where issues concerning prior taught knowledge, top-down schema building, lexical, and bottom-up identification of sounds coincided for multiple learners. Extrapolating from these points of convergence, pedagogical recommendations are provided.

Decision-making strategies implemented in SolFinder 1.0 to identify eco-efficient aircraft trajectories: application study in AirTraf 3.0

Abstract. The optimization of aircraft trajectories involves balancing operating costs and climate impact, which are often conflicting objectives. To achieve compromised optimal solutions, higher-level information such as preferences of decision-makers must be taken into account. This paper introduces the SolFinder 1.0 module, a decision-making tool designed to identify eco-efficient aircraft trajectories, which allow for the reduction of the flight's climate impact with limited cost penalties compared to cost-optimal solutions. SolFinder 1.0 offers flexible decision-making options that allow users to select trade-offs between different objective functions, including fuel use, flight time, NOx emissions, contrail distance, and climate impact. The module is included in the AirTraf 3.0 submodel, which optimizes trajectories under atmospheric conditions simulated by the ECHAM/MESSy Atmospheric Chemistry model. This paper focuses on the ability of the module to identify eco-efficient trajectories while solving a bi-objective optimization problem that minimizes climate impact and operating costs. SolFinder 1.0 enables users to explore trajectory properties at varying locations of the Pareto fronts without prior knowledge of the problem results and to identify solutions that limit the cost of reducing the climate impact of a single flight.

Matching Process Concerning Children With Disabilities in Family Foster Care

ABSTRACTChildren with disabilities are over‐represented in foster care and exposed to more disruptions. This is unfortunate because they need long‐term relationships and predictability, as well as access to special schools and health care. Disruption can be the result of a poor match between the child and the foster family. In this article, we focus on the matching process in relation to children with disabilities. Zeijlmans et al. describe matching as the process by which social workers move from principles of good matching to realistic matching, depending on existing premises. Using their model, we analyse the matching process for children with disabilities in Sweden. The study is based on 31 individual interviews and six group interviews with social workers and social work managers. They describe how they take into account the family climate, family composition and prior knowledge and/or experience of disability in the foster home. However, disruptions occur even when these conditions are met. Stable placements seem to be more related to the fostered child's progress when the family's efforts are rewarded. We also address the complex issue of what disability actually means in the context of foster care, given the interrelatedness of trauma, maltreatment and impairment.

Cross-lingual Cross-temporal Summarization: Dataset, Models, Evaluation

Abstract While summarization has been extensively researched in natural language processing (NLP), cross-lingual cross-temporal summarization (CLCTS) is a largely unexplored area that has the potential to improve cross-cultural accessibility and understanding. This paper comprehensively addresses the CLCTS task, including dataset creation, modeling, and evaluation. We (1) build the first CLCTS corpus with 328 (+127) instances for hDe-En and 289 (+212) for hEn-De, leveraging historical fiction texts and Wikipedia summaries in English and German; (2) examine the effectiveness of popular transformer end-to-end models with different intermediate finetuning tasks; (3) explore the potential of GPT-3.5 as a summarizer; (4) report evaluations from humans, GPT-4, and several recent automatic evaluation metrics. Our results indicate that intermediate task finetuned end-to-end models generate bad to moderate quality summaries while GPT-3.5, as a zero-shot summarizer, provides moderate to good quality outputs. GPT-3.5 also seems very adept at normalizing historical text. To assess data contamination in GPT-3.5, we design an adversarial attack scheme in which we find that GPT-3.5 performs slightly worse for unseen source documents compared to seen documents. Moreover, it sometimes hallucinates when the source sentences are inverted against its prior knowledge with a summarization accuracy of 0.67 for plot omission, 0.71 for entity swap, and 0.53 for plot negation. Overall, our regression results of model performances suggest that longer, older, and more complex source texts (all of which are more characteristic for historical language variants) are harder to summarize for all models, indicating the difficulty of the CLCTS task. Regarding evaluation, we observe that both GPT-4 and BERTScore correlate moderately with human evaluations but GPT-4 is prone to giving lower scores.

Comparative study for the performance of pure artificial intelligence software sensor and self-organizing map assisted software sensor in predicting 5-day biochemical oxygen demand for Kauma Sewage Treatment Plant effluent in Malawi

Introduction: Modeling plays a crucial role in understanding wastewater treatment processes, yet conventional deterministic models face challenges due to complexity and uncertainty. Artificial intelligence offers an alternative, requiring no prior system knowledge. This study tested the reliability of the Adaptive Fuzzy Inference System (ANFIS), an artificial intelligence algorithm that integrates both neural networks and fuzzy logic principles, to predict effluent Biochemical Oxygen Demand. An important indicator of organic pollution in wastewater.Materials and Methods: The ANFIS models were developed and validated with historical wastewater quality data for the Kauma Sewage Treatment Plant located in Lilongwe City, Malawi. A Self Organizing Map (SOM) was applied to extract features of the raw data to enhance the performance of ANFIS. Cost-effective, quicker, and easier-to-measure variables were selected as possible predictors while using their respective correlations with effluent. Influents’ temperature, pH, dissolved oxygen, and effluent chemical oxygen demand were among the model predictors.Results and Discussions: The comparative results demonstrated that for the same model structure, the ANFIS model achieved correlation coefficients (R) of 0.92, 0.90, and 0.81 during training, testing, and validation respectively, whereas the SOM-assisted ANFIS Model achieved R Values of 0.99, 0.87 and 0.94. Overall, despite the slight decrease in R-value during the testing stage, the SOM- assisted ANFIS model outperformed the traditional ANFIS model in terms of predictive capability. A graphic user interface was developed to improve user interaction and friendliness of the developed model. Integration of the developed model with supervisory control and data acquisition system is recommended. The study also recommends widening the application of the developed model, by retraining it with data from other wastewater treatment facilities and rivers in Malawi.

On Bayesian premium credibility: maximum entropy prior for unknown claims process and numerical example

In the complex field of insurance, determining the credibility of the most efficient premium is a major challenge, especially when the distribution of claims remains unknown. This study uses a Bayesian methodology to address this issue. In this framework, the initial assumptions concerning the unidentified parameters of the claims process are expressed by means of a distribution function. The objective is to establish premium credibility by maximizing the insurance company's expected utility while respecting specific constraints. By applying the maximum entropy method, we can derive the posterior distribution of the loss process from the observed data available. This approach is advantageous because it offers modeling flexibility and enables estimates to be progressively updated as new data becomes available. It is also particularly useful in situations where data is limited or incomplete, as is often the case in the insurance sector. The study highlights the characteristics and advantages of this Bayesian method. It highlights how the approach handles uncertainty and integrates prior knowledge into the estimation process. In addition, several numerical illustrations are included to demonstrate the practical application of the method and its impact on determining premium credibility. In summary, this research offers an innovative perspective on risk management in the insurance sector. It proposes a method that not only adheres to the playful approach to risk management, but also to the premium credibility approach.

Comparative Study of Multi‐objective Bayesian Optimization and NSGA‐III based Approaches for Injection Molding Process

AbstractInjection molding is a prevalent method for producing plastic components, yet determining the ideal process parameters has predominantly relied on heuristic approaches. In this research, a data‐driven injection molding process optimization framework is developed to simultaneously minimize warpage, cycle time, and clamping force. Employing multi‐objective Bayesian optimization (MBO), the framework is applied to a fan blade model for verification. Incorporating those objectives enables the selection of an injection molding machine with the proper maximum clamping force within acceptable tolerance and production time. Furthermore, non‐moldable regimes are considered in design space, which allows less prior knowledge for setting design space. The optimized results are compared with those obtained using NSGA‐III, a well‐established genetic algorithm‐based optimization technique, as a benchmark. The optimization frameworks produce a Pareto front for the three‐dimensional outputs, revealing distinct trade‐off relationships between cycle time and warpage, as well as clamping force and warpage. The MBO framework demonstrates a superior Pareto front compared to NSGA‐III when utilizing a limited data set, underscoring its benefits in scenarios where costly simulations or experiments are necessary. These findings are anticipated to contribute to the optimization of manufacturing processes, ultimately enhancing productivity in real‐world industries.

Sand fill cleanout on wireline enables access to additional perforation zones in gas well producer

An Australian operator wanted to explore options for wireline conveyed sand bailing in one of their offshore deviated ‘S’ shape gas producing wells. The objective was to remove approximately 24 m of sand fill inside 9.625″ production casing, with an internal diameter (ID) of 8.681″. To limit the number of cleanout runs, the latest technology in wireline deployed suction tools with a 4.25″ overall diameter (OD) was chosen, which provided increased recovery volume, plus real-time surface control and monitoring. Because of the tool’s overall diameter, the landing nipple also required milling to allow access. Therefore, a specially designed 4.412″ mill bit was manufactured and run on a milling tool. The electric line deployed milling technology successfully milled through the landing nipple and a ball catcher in four runs, thus increasing the ID sufficiently to allow access for the cleanout tool string. A well fluid column of about seven bars is required to enable operation of the suction tool. The cleanout tool was configured with three bailer sections (seven max) for the first run, to determine sand recovery optimisation. Three different sizes of bailer filters are provided; prior knowledge of debris particle size is therefore advantageous. In this case, a slickline bailing sand sample had been recovered, which immediately aided filter choice. Seventeen bailing runs were conducted over 5 days, successfully recovering a total of 918 L of sand, equating to about 2.4 tons. Ultimately, the process enabled perforation of two additional gas zones to increase production.

Enhancing Oil–Water Flow Prediction in Heterogeneous Porous Media Using Machine Learning

The rapid and accurate forecasting of two-phase flow in porous media is a critical challenge in oil field development, exerting a substantial impact on optimization and decision-making processes. Although the Convolutional Long Short-Term Memory (ConvLSTM) network effectively captures spatiotemporal dynamics, its generalization in predicting complex engineering problems remains limited. Similarly, although the Fourier Neural Operator (FNO) demonstrates adeptness at learning operators for solving partial differential equations (PDEs), it struggles with three-dimensional, long-term prediction. In response to these limitations, we introduce an innovative hybrid model, the Convolutional Long Short-Term Memory-Fourier Neural Operator (CL-FNO), specifically designed for the long-term prediction of three-dimensional two-phase flows. This model integrates a 3D convolutional encoder–decoder structure to extract and generate hierarchical spatial features of the flow fields. It incorporates physical constraints to enhance the model’s forecasts with robustness through the infusion of prior knowledge. Additionally, a temporal function, constructed using gated memory-forgetting mechanisms, augments the model’s capacity to analyze time series data. The efficacy and practicality of the CL-FNO model are validated using a synthetic three-dimensional case study and application to an actual reservoir model.

An exploratory study of the relationship between pulse transit time and blood pressure based on causal inference

Aim: We propose to examine the causal relationship between the noninvasive features represented by pulse transit time (PTT) and blood pressure (BP), with the aim of mitigating the impact of confounding factor(s) and thus improving the performance of BP estimation. Methods: We identified the causal graph of BP and the important noninvasive features extracted from electrocardiogram (ECG) and photoplethysmogram (PPG) via fast causal inference (FCI) algorithm, with the orientations of the edges in the causal graph being determined by the causal generative neural networks (CGNN) algorithm. With the knowledge obtained from the causal graph, we further used hierarchical regression model to estimate BP, and validated the proposed method on 17 subjects. Results: We found that the obtained causal graph was almost consistent with the prior knowledge, and heart rate (HR) was one of the main confounding factors of PTT and BP. Incorporating HR into the hierarchical regression model to eliminate its confounding effect on the PTT-based BP estimation, the mean value of SBP and DBP estimation was improved by 1.27 and 1.89 mmHg, respectively, and the mean absolute difference (MAD) was improved by 2.28 and 3.60 mmHg, respectively. Conclusion: Causal inference-based method has the potential to clarify the causal relationship between BP and related wearable noninvasive features, which can further shed light on developing new methods for cuffless BP with acceptable accuracy.

When Load is Low, Working Memory is Shielded From Long-Term Memory's Influence.

Previous studies found that episodic long-term memory (eLTM) enhances working memory (WM) performance when both novel and previously learnt word pairs must be retained on a short-term basis. However, there is uncertainty regarding how and when WM draws on eLTM. Three possibilities are (a) that people draw on eLTM only if WM capacity is exceeded; (b) that there is always a contribution of eLTM to WM performance, irrespective of whether prior knowledge is helpful or not; or (c) benefits of prior knowledge are specific to comparisons between conditions which are similarly ambiguous concerning whether LTM may be useful. We built on the assumption that under conditions of a contribution from LTM, these LTM traces of memoranda could benefit or hamper performance in WM tasks depending on the match between the traces stored in LTM and the ones to-be stored in WM in the current trial, yielding proactive facilitation (PF) and proactive interference (PI), respectively. Across four experiments, we familiarized participants with some items before they completed a separate WM task. In accordance with possibility (a) we show that there are indeed conditions in which only WM contributes to performance. Performance deteriorated with the addition of stimuli from eLTM when WM load was low, but not when it was high; and an exchange of information between LTM and WM occurred only when WM capacity was exceeded, with PI and PF effects affecting immediate memory performance in verbal and visual tasks only at higher set sizes.

Retrosynthesis Zero: Self-Improving Global Synthesis Planning Using Reinforcement Learning.

The field of computer-aided synthesis planning (CASP) has witnessed significant growth in recent years. Still, many CASP programs rely on large data sets to train neural networks, resulting in limitations due to the data quality and prior knowledge from chemists. In response, we propose Retrosynthesis Zero (ReSynZ), a reaction template-based method that combines Monte Carlo Tree Search with reinforcement learning inspired by AlphaGo Zero. Unlike other single-step reaction template-based CASP methods, ReSynZ takes complete synthesis paths for complex molecules, determined by reaction rules, as input for training the neural network. ReSynZ enables neural networks trained with relatively small reaction data sets (tens of thousands of data) to generate multiple synthesis pathways for a target molecule and suggest possible reaction conditions. On multiple data sets of molecular retrosynthesis, ReSynZ demonstrates excellent predictive performance compared to existing algorithms. The advantages, such as self-improving model features, flexible reward settings, the potential to surpass human limitations in chemical synthesis route planning, and others, make ReSynZ a valuable tool in chemical synthesis design.

Analyzing the Role of Problem Finding in the Creative Design with Regard to the Effect of Prior Knowledge in the Design Process

Statement of the problem. Despite many researches in the field of design process, there is still mostly the main reliance of the design process on ‘solving the problem’, and designers are evaluated based on their ability to provide the solution. In fact, creativity is defined in ‘quality of the solution’. Results. However, what is less discussed is the ‘problem-finding’. considering that the design problem – unlike problems in other fields – is vague and ill-defined, and is not predetermined; Also, as prior knowledge of the designer is so effective in all stages of the design process and can hinder creativity, due to the desiner’s reference to his own previous experience, so problem finding should be considered as one of the important abilities of a designer. In this regard, students should also be taught to be creative in the problem-finding phase, and teachers should not provide the design problem in a completely defined and fixed way. Conclusions. This research believes that creativity does not occur only in the ideation phase, and the problem finding should be considered as an integral and essential part of the creative design process, whether in professional contexts or education. If the designer can find a novel problem, or see the existing problems from a new angle, it can be expected that there will be no fixation, and creativity will happen in the design.