Knowledge-based Reasoning Research Articles

놀이중심 교육과정의 놀이기록에서 전문적 시각으로 살펴본 교사의 PCK

Objectives This study aims to explore the Pedagogical Content Knowledge (PCK) of teachers in the context of play-based early childhood education. The focus is on how teachers observe, document, and support play activities in educational settings that demand a high level of professional expertise. By analyzing teachers' play documentation from a professional vision perspective, the study seeks to identify what aspects teachers focus on during play situations and how their PCK is manifested. Methods This study analyzed 19 play documentation cases from the 2019 revised Nuri Curriculum to examine teachers' selective attention and knowledge-based reasoning. The analysis of selective attention focused on the specific elements of play that teachers observed, the aspects of children they concentrated on, the areas they emphasized to support and suggest during play, and the experiences and learnings they highlighted. For knowledge-based reasoning, the study assessed teachers' inferential actions based on observational evidence, their professional knowledge, and educational theories. Results First, the analysis revealed that teachers focused on the transformation of play through chance encounters and on all the relationships children form. Second, in the interplay of play and activities, the teachers' PCK was evident in their attuned listening to children's language and behavior while implementing their educational plans, as well as their pedagogical PCK in responding spontaneously to children's needs and supporting their learning. Conclusions The examination of teachers' PCK in play documentation indicated that it is an intertwining of educational knowledge, contextual (practical) knowledge, and knowledge about children. This results in the manifestation of PCK through planning, observation, support, and interpretation by teachers. The study highlighted the importance of teachers’ roles in implementing a play-based curriculum that aims to deepen, expand, and reorganize educational content and experiences to promote higher levels of learning and support.

Improving Teachers’ Professional Vision Through a Video-based Reflection Program: A Case Study in Mexican Primary Schools

Enhancing teachers’ professional vision (TPV) may contribute to increasing the quality of teaching-learning processes through higher perceived efficacy, leading to autonomy-supportive environment. This is of particular importance in Physical Education (PE), due to the fact that this subject presents different challenges and more dynamic situations compared to others. One of the latest approaches for enhancing TPV is video-based reflection, which allows teachers to observe recorded lessons and analyze them. The aim of this research was to test the effect of a video-based reflection program on TPV in PE. Two teachers participated in a 3-step video-analysis program consisting of self-reflection, peer reflection, and expert feedback. Notable changes were found in TPV at the end of the 6-month intervention, both in terms of teachers’ selective attention to classroom events and their knowledge-based reasoning. Our findings suggest that video-based reflection programs could represent an important part of any teacher training program aiming to increasing PE teachers’ responsiveness to diverse in-class situations.

Memory-Enhanced Knowledge Reasoning with Reinforcement Learning

In recent years, the emergence of large-scale language models, such as ChatGPT, has presented significant challenges to research on knowledge graphs and knowledge-based reasoning. As a result, the direction of research on knowledge reasoning has shifted. Two critical issues in knowledge reasoning research are the algorithm of the model itself and the selection of paths. Most studies utilize LSTM as the path encoder and memory module. However, when processing long sequence data, LSTM models may encounter the problem of long-term dependencies, where memory units of the model may decay gradually with an increase in time steps, leading to forgetting earlier input information. This can result in a decline in the performance of the LSTM model in long sequence data. Additionally, as the data volume and network depth increase, there is a risk of gradient disappearance. This study improved and optimized the LSTM model to effectively address the problems of gradient explosion and gradient disappearance. An attention layer was employed to alleviate the issue of long-term dependencies, and ConvR embedding was used to guide path selection and action pruning in the reinforcement learning inference model. The overall model achieved excellent reasoning results.

Visual Chain-of-Thought Prompting for Knowledge-Based Visual Reasoning

Knowledge-based visual reasoning remains a daunting task since it not only requires machines to interpret the concepts and relationships from visual scenes but also associate them with external world knowledge to conduct a chain of reasoning on open-world questions. Previous works, however, treat visual perception and language-based reasoning as two independent modules, failing to attend to both modules throughout all stages of reasoning. To this end, we propose Visual Chain-of-thought Prompting (VCTP) for knowledge-based reasoning, which involves the interaction between visual content and natural language in an iterative step-by-step reasoning manner. VCTP contains three stages, see, think, and confirm. The see stage scans the image and grounds the visual concept candidates with a visual perception model. The think stage adopts a pre-trained large language model (LLM) to attend to key visual concepts from natural language questions adaptively. It then transforms key visual context into text context for prompting with a visual captioning model, and adopts the LLM to generate the answer. The confirm stage further uses the LLM to generate the supporting rationale to the answer, which is then passed through a cross-modality classifier to verify that it’s consistent with the visual context. We iterate through the think-confirm stages to ensure the verified rationale is consistent with the answer. We conduct experiments on a range of knowledge-based visual reasoning datasets. We found our VCTP enjoys several benefits, 1). it achieves better performance than the previous few-shot learning baselines; 2). it enjoys the total transparency and trustworthiness of the whole reasoning process by providing rationales for each reasoning step; 3). it is computation-efficient compared with other fine-tuning baselines. Our code is available at https://github.com/UMass-Foundation-Model/VisualCoT.git

The subject matters for the professional vision of classroom management: an exploratory study with biology and mathematics expert teachers

Teacher professional vision of classroom management is one crucial aspect of teacher expertise that has so far been studied without considering the role of teachers’ subjects. However, subject teaching is characterized by typical settings and activities that might require different classroom management strategies. This small-scale explorative study investigates whether twenty expert teachers from two secondary school subjects (biology and mathematics) differ in their professional vision of classroom management. Using video clips of two settings as stimuli, teachers’ eye-tracking data and retrospective think-aloud data were recorded. Think-aloud data were investigated with quantitative content analysis and epistemic network analysis. Expert teachers’ visual attention, their noticing of classroom management events, and their knowledge-based reasoning were compared for both groups. Results reveal subject-specific aspects of expert teachers’ professional vision of classroom management in terms of events noticed and their reasoning about these events. Expert biology teachers were more concerned with suggesting alternative classroom management strategies, particularly strategies addressing aspects to consider when planning activities such as providing structure or preparing the classroom. In contrast, mathematics teachers were more evaluative in their analysis of events and focused more on behavioral management or ensuring students’ engagement in the moment.

A conceptual framework for assessing transdisciplinary STEM practices

ABSTRACT Background Integrated STEM education has attracted increasing attention in global educational reforms. While transdisciplinary approaches are deemed useful pedagogy for integrated STEM curriculum design, few assessment instruments are available for appraising the efficacy of teaching and learning in transdisciplinary settings. Purpose This study proposes a framework to conceptualize the meaning of transdisciplinary STEM practices and demonstrates how the framework can be used to develop an assessment for transdisciplinary practices. Sources of evidence The conceptual framework is contextualized in design context. The concepts of need-to-do and need-to-know are used to illustrate how design engages students in transdisciplinary practices. The need-to-do cycle includes the four design phases, whereas the need-to-know part concerns the relevant STEM knowledge applied in these phases. Knowledge-based reasoning is identified as the focused practice in transdisciplinary STEM assessments. Based on the conceptual framework, we formulate the assessment template and scoring rubrics following the four steps suggested in the principled assessment design approach. Two dimensions – the types of disciplinary knowledge and the levels of reasoning – are examined to describe students’ transdisciplinary STEM practices in a design context. An exemplary STEM assessment task on a mini-wire-controlled car is presented. Main argument The conceptual framework proposed in this study presents one possible solution to align the goals of integrated STEM programs and their evaluation instruments. Conclusions The assessment of transdisciplinary STEM practices can contribute to our understanding of how students apply and reason with disciplinary knowledge in transdisciplinary settings.

Knowledge-Based Reasoning Network for Relation Detection.

With the rapid growth of large-scale knowledge bases (KBs), knowledge base question answering (KBQA) has attracted increasing attention recently. Relation detection plays an important role in the KBQA system, which finds a compatible answer by analyzing the semantics of questions and querying and reasoning with multiple KB triples. Significant progress has been made by deep neural networks. However, existing methods often concern on detecting single-hop relation without path reasoning, and a few of these methods exploit the multihop relation reasoning, which involves the answer reasoning from the noisy and abundant relational paths in the KB. Meanwhile, the relatedness between question and answer candidates has received little attention and remains unsolved. This article proposes a novel knowledge-based reasoning network (KRN) for relation detection, including both single-hop relation and multihop relation. To address the semantic gap problem in question-answer interaction, we first learn attentive question representations according to the influence of answer aspects. Then, we learn the single-hop relation sequence through different levels of abstraction and adopt the KB entity and structure information to denoise the multihop relation detection task. Finally, we adopt a Siamese network to measure the similarity between question representation and relation representation for both single-hop and multihop relation KBQA tasks. We conduct experiments on two well-known benchmarks, SimpleQuestions and WebQSP, and the results show the superiority of our approach over the state-of-the-art models for both single-hop and multihop relation detection. Our model also achieves a significant improvement over existing methods on KBQA end task. Further analysis demonstrates the robustness and the applicability of the proposed approach.

Knowledge-based Reasoning and Learning under Partial Observability in Ad Hoc Teamwork

AbstractAd hoc teamwork (AHT) refers to the problem of enabling an agent to collaborate with teammates without prior coordination. State of the art methods in AHT are data-driven, using a large labeled dataset of prior observations to model the behavior of other agent types and to determine the ad hoc agent’s behavior. These methods are computationally expensive, lack transparency, and make it difficult to adapt to previously unseen changes. Our recent work introduced an architecture that determined an ad hoc agent’s behavior based on non-monotonic logical reasoning with prior commonsense domain knowledge and models learned from limited examples to predict the behavior of other agents. This paper describes KAT, a knowledge-driven architecture for AHT that substantially expands our prior architecture’s capabilities to support: (a) online selection, adaptation, and learning of the behavior prediction models; and (b) collaboration with teammates in the presence of partial observability and limited communication. We illustrate and experimentally evaluate KAT’s capabilities in two simulated benchmark domains for multiagent collaboration: Fort Attack and Half Field Offense. We show that KAT’s performance is better than a purely knowledge-driven baseline, and comparable with or better than a state of the art data-driven baseline, particularly in the presence of limited training data, partial observability, and changes in team composition.

A novel framework to improve motion planning of robotic systems through semantic knowledge-based reasoning

The need to improve motion planning techniques for manipulator robots, and new effective strategies to manipulate different objects to perform more complex tasks, is crucial for various real-world applications where robots cooperate with humans. This paper proposes a novel framework that aims to improve the motion planning of a robotic agent (a manipulator robot) through semantic knowledge-based reasoning. The Semantic Web Rule Language (SWRL) was used to infer new knowledge based on the known environment and the robotic system. Ontological knowledge, e.g., semantic maps, were generated through a deep neural network, trained to detect and classify objects in the environment where the robotic agent performs. Manipulation constraints were deduced, and the environment corresponding to the agent’s manipulation workspace was created so the planner could interpret it to generate a collision-free path. For reasoning with the ontology, different SPARQL queries were used. The proposed framework was implemented and validated in a real experimental setup, using the planning framework ROSPlan to perform the planning tasks. The proposed framework proved to be a promising strategy to improve motion planning of robotics systems, showing the benefits of artificial intelligence, for knowledge representation and reasoning in robotics.

Effects of different video- or text-based reflection stimuli on pre-service teachers’ emotions, immersion, cognitive load and knowledge-based reasoning

To reflect on ones own or other teachers teaching practices as well as on different reflection stimuli (e.g. videos or texts) can evoke different emotions and entail varying degrees of cognitive load and knowledge-based reasoning for pre-service teachers. Participants in this study were distributed to four groups (“memory”; "video+memory”; “video other”, and “text other”) and reflected online on classroom management facets. Pre-service teachers in the “video+memory” group reported significantly more immersion and enjoyment while reflecting than the other three groups. All groups showed a moderate level of germane load and there were no differences between the groups. Participants in the video+memory and memory group showed higher levels of knowledge-based reasoning than those who reflected on other teachers’ classroom practice. Enjoyment was a significant positive predictor of both immersion and germane load, whereas boredom was a negative predictor. Implications for future research and reflection processes during teacher education are discussed.

AI for life: Trends in artificial intelligence for biotechnology

Due to popular successes (e.g., ChatGPT) Artificial Intelligence (AI) is on everyone's lips today. When advances in biotechnology are combined with advances in AI unprecedented new potential solutions become available. This can help with many global problems and contribute to important Sustainability Development Goals. Current examples include Food Security, Health and Well-being, Clean Water, Clean Energy, Responsible Consumption and Production, Climate Action, Life below Water, or protect, restore and promote sustainable use of terrestrial ecosystems, sustainably manage forests, combat desertification, and halt and reverse land degradation and halt biodiversity loss. AI is ubiquitous in the life sciences today. Topics include a wide range from machine learning and Big Data analytics, knowledge discovery and data mining, biomedical ontologies, knowledge-based reasoning, natural language processing, decision support and reasoning under uncertainty, temporal and spatial representation and inference, and methodological aspects of explainable AI (XAI) with applications of biotechnology. In this pre-Editorial paper, we provide an overview of open research issues and challenges for each of the topics addressed in this special issue. Potential authors can directly use this as a guideline for developing their paper.

Development of a Knowledge-Based Expert System for Diagnosing Post-Harvest Diseases of Apple

Post-harvest diseases are one of the main causes of economical losses in the apple fruit production sector. Therefore, this paper presents an application of a knowledge-based expert system to diagnose post-harvest diseases of apple. Specifically, we detail the process of domain knowledge elicitation for constructing a Bayesian network reasoning system. We describe the developed expert system, dubbed BN-DSSApple, and the diagnostic mechanism given the evidence provided by the user, as well as a likelihood evidence method, learned from the estimated consensus of users’ and expert’s interactions, to effectively transfer the performance of the model to different cohorts of users. Finally, we detail a novel technique for explaining the provided diagnosis, thus increasing the trust in the system. We evaluate BN-DSSApple with three different types of user studies, involving real diseased apples, where the ground truth of the target instances was established by microbiological and DNA analysis. The experiments demonstrate the performance differences in the knowledge-based reasoning mechanism due to heterogeneous users interacting with the system under various conditions and the capability of the likelihood-based method to improve the diagnostic performance in different environments.

Intelligent architecture of space unmanned system with learning and reasoning capabilities

Spacecraft represented by satellites, manned spaceships, and deep space probes are typical unmanned systems that are developing from automation to autonomy. To improve the adaptability and intelligent autonomy level of spacecraft to unknown space environments and complex space tasks, an intelligent architecture of space unmanned system with learning and reasoning capabilities is designed and the architectural components and functions are proposed. Then, the operating mechanism of learning and reasoning capabilities in the architecture are analyzed, and the solutions for key technologies, such as the construction, update, and expansion methods of the action and knowledge libraries, are proposed. Finally, through an example of the extraterrestrial surveying and sampling task, the process of new knowledge generation, knowledge-based reasoning, and autonomous task execution under this intelligent architecture, as well as the interaction between the knowledge library, environment, task, and action library, are explained.

Professional vision in the classroom: Teachers’ knowledge-based reasoning explaining their visual focus of attention to students

This study investigated Grade 2 teachers' (N = 50) professional vision through eye-tracking methodology and retrospective think-aloud interviews. The study examined the extent to which teachers' knowledge-based reasoning explains their visual focus of attention to whole class and individual students. We found that teachers' descriptions of students' social relations and emotions associated positively with teachers' visual focus of attention to the whole class. Teachers' descriptions of teacher-related information/elaboration and pedagogy linked negatively with teachers' visual focus of attention to individual students. The findings suggest that teachers' visual focus of attention to students may vary depending on the reasons that guide teachers’ attention.

The Refined Consensus Model of Pedagogical Content Knowledge (PCK): Detecting Filters between the Realms of PCK

In this article, we analyse potential filters that moderate the transformation process between the realms of PCK defined in the refined consensus model of pedagogical content knowledge. We tested 58 preservice biology teachers in a 15-week one-group pretest/post-test design. To identify filters between collective PCK (cPCK) and personal PCK (pPCK), we set up moderation models with pretest pPCK as an independent variable, post-test pPCK as a dependent variable, and motivational orientations or professional values as moderator variables. To identify filters between pPCK and enacted PCK (ePCK), we set up moderation models with post-test pPCK as an independent variable, ePCK as a dependent variable, and noticing or knowledge-based reasoning as moderator variables. We did this specifically with a focus on language in biology education. We found that only the variable knowledge-based reasoning had a role as a filter. It moderates the transformation process between pPCK and ePCK (moderation analysis: F(3,19) = 10.40, p < 0.001, predicting 25.72% of the variance). In future studies, other filters should be identified.

A Data-Knowledge Hybrid Driven Method for Gas Turbine Gas Path Diagnosis

Gas path fault diagnosis of a gas turbine is a complex task involving field data analysis and knowledge-based reasoning. In this paper, a data-knowledge hybrid driven method for gas path fault diagnosis is proposed by integrating a physical model-based gas path analysis (GPA) method with a fault diagnosis ontology model. Firstly, a physical model-based GPA method is used to extract the fault features from the field data. Secondly, a virtual distance mapping algorithm is developed to map the GPA result to a specific fault feature criteria individual described in the ontology model. Finally, a fault diagnosis ontology model is built to support the automatic reasoning of the maintenance strategy from the mapped fault feature criteria individual. To enhance the ability of selecting a proper maintenance strategy, the ontology model represents more abundant knowledge from several sources, such as fault criteria analysis, physical structure analysis, FMECA (failure mode, effects, and criticality analysis), and the maintenance logic decision tool. The availability of the proposed hybrid driven method is verified by the field fault data from a real GE LM2500 PLUS gas turbine unit. The results indicate that the hybrid driven method is effective in detecting the path fault in advance. Furthermore, diversified fault information, such as fault effects, fault criticality, fault consequence, and fault detectability, could be provided to support selecting a proper maintenance strategy. It is proven that the data-knowledge hybrid driven method can improve the capability of the gas path fault detection, fault analysis, and maintenance strategy selection.

MICAR: multi-inhabitant context-aware activity recognition in home environments

The sensor-based recognition of Activities of Daily Living (ADLs) in smart-home environments enables several important applications, including the continuous monitoring of fragile subjects in their homes for healthcare systems. The majority of the approaches in the literature assume that only one resident is living in the home. Multi-inhabitant ADLs recognition is significantly more challenging, and only a limited effort has been devoted to address this setting by the research community. One of the major open problems is called data association, which is correctly associating each environmental sensor event (e.g., the opening of a fridge door) with the inhabitant that actually triggered it. Moreover, existing multi-inhabitant approaches rely on supervised learning, assuming a high availability of labeled data. However, collecting a comprehensive training set of ADLs (especially in multiple-residents settings) is prohibitive. In this work, we propose MICAR: a novel multi-inhabitant ADLs recognition approach that combines semi-supervised learning and knowledge-based reasoning. Data association is performed by semantic reasoning, combining high-level context information (e.g., residents’ postures and semantic locations) with triggered sensor events. The personalized stream of sensor events is processed by an incremental classifier, that is initialized with a limited amount of labeled ADLs. A novel cache-based active learning strategy is adopted to continuously improve the classifier. Our results on a dataset where up to 4 subjects perform ADLs at the same time show that MICAR reliably recognizes individual and joint activities while triggering a significantly low number of active learning queries.

Knowledge Service Technology for Supporting Intelligent Product Design

In this paper, a knowledge service method that supports the intelligent design of products is investigated. The proposed method provides the solutions to computational problems and reasoning and decision-making problems in the field of intelligent design. The requirement analysis of a knowledge-based intelligent design system integrates design knowledge into case-based reasoning activities through scheme analysis, scheme evaluation, and scheme adjustment, thus achieving knowledge-based intelligent reasoning and decision-making. During the similarity matching, a new hybrid similarity measurement method is proposed to calculate the similarity of crisp and fuzzy sets. This method integrates the fuzzy set similarity theory based on the traditional similarity measurement method. A method of attribute level classification is proposed to assign weight coefficients. The attributes are divided into the primary matching and auxiliary matching levels according to the decisiveness of case matching, and the set of weight coefficients is continuously and dynamically updated through case-based reasoning learning. Then, the weighted global similarity measure is used to obtain the set of similar cases from the case database. Finally, a design example of a computer numerical control tool holder product is studied to present the practicability and effectiveness of the proposed method.

Knowledge Engineering Framework for IoT Robotics Applied to Smart Healthcare and Emotional Well-Being.

Social companion robots are getting more attention to assist elderly people to stay independent at home and to decrease their social isolation. When developing solutions, one remaining challenge is to design the right applications that are usable by elderly people. For this purpose, co-creation methodologies involving multiple stakeholders and a multidisciplinary researcher team (e.g., elderly people, medical professionals, and computer scientists such as roboticists or IoT engineers) are designed within the ACCRA (Agile Co-Creation of Robots for Ageing) project. This paper will address this research question: How can Internet of Robotic Things (IoRT) technology and co-creation methodologies help to design emotional-based robotic applications? This is supported by the ACCRA project that develops advanced social robots to support active and healthy ageing, co-created by various stakeholders such as ageing people and physicians. We demonstra this with three robots, Buddy, ASTRO, and RoboHon, used for daily life, mobility, and conversation. The three robots understand and convey emotions in real-time using the Internet of Things and Artificial Intelligence technologies (e.g., knowledge-based reasoning).

Science Teacher Noticing via Video Annotation: Links between Complexity and Knowledge-Based Reasoning

ABSTRACT This study adds to the construct of science teacher noticing through an analysis that integrates concepts from pedagogical content knowledge (PCK) frameworks. It examines 929 instances of science teacher noticing made via video annotation tools (VATs) over the course of 38 early-career, secondary science teachers’ participation in a two-year professional learning experience. This study asked specifically about: 1) interacting elements of a classroom (students, practice, content) teachers noticed in videos of their own practice, 2) patterns in knowledge-based reasoning evidenced in noticing, and 3) the role of content knowledge in such knowledge-based reasoning. Using mixed methods analysis, we found that the majority of annotations exhibited a singular focus (i.e. students, teacher practice, or content) and tended to focus on students and/or practice, with less attention to content. However, we also identified a statistically significant relationship between complexity of noticing (i.e. attention to multiple foci) and the presence of knowledge-based reasoning. Further qualitative analysis of the subset of annotations that attended to content suggested that such attention to content could be an important link within teacher noticing to connect practice to student thinking. This study highlights the value of scaffolding attention to content and to multiple foci within science teacher noticing. It also demonstrates the potential for VATs to facilitate noticing.