A neural-symbolic AI agent system for biomedical concept mapping.

From lexical regularities to axiomatic patterns for the quality assurance of biomedical terminologies and ontologies

Concept mapping has been applied in a variety of fields, including instruction, learning, curriculum development, and assessment. Because many empirical studies have proven the validity of concept mapping, a computer-based concept mapping system has been developed. The system provides two learning environments. In the 'construct-by-self' environment, the system provides students with the evaluation results and corresponding hints for feedback. The students construct concept maps by themselves with only the assistance of the feedback. In the 'construct-on- scaffold' environment, in addition to the feedback, the students receive an incomplete concept map, within which some nodes and links were set as blanks for the scaffold. A study comparing the effectiveness of the 'construct-by-self', 'construct-on-scaffold', and 'construct by paper-and- pencil' concept mapping showed that the 'construct-on-scaffold' had better effect for learning on biology. Both of the two computer-based procedures are helpful for students in completing their concept maps.

Data mining has recently drawn an increasing interest as an effective approach to generation of a concept map in an adaptive learning platform that provides students with personalized learning guidance. Although it has seen significant progresses, the data mining-based concept map generation needs to be further improved both in complexity and accuracy for wide acceptance in actual education services. This paper aims to improve the accuracy of concept map by considering both wrong-to-wrong and correct-to-correct relationships of questions, and by adopting more accurate formulas in calculation of relevance degrees between concepts. Through simulations using a set of concepts, questions, and student test records sampled from a practical courseware, we show that the proposed approach can generate a more accurate and robust concept map at an acceptable additional complexity.

The National Center for Biomedical Ontology (NCBO) is one of the National Centers for Biomedical Computing funded under the NIH Roadmap Initiative. Contributing to the national computing infrastructure, NCBO has developed BioPortal, a web portal that provides access to a library of biomedical ontologies and terminologies (http://bioportal.bioontology.org) via the NCBO Web services. BioPortal enables community participation in the evaluation and evolution of ontology content by providing features to add mappings between terms, to add comments linked to specific ontology terms and to provide ontology reviews. The NCBO Web services (http://www.bioontology.org/wiki/index.php/NCBO_REST_services) enable this functionality and provide a uniform mechanism to access ontologies from a variety of knowledge representation formats, such as Web Ontology Language (OWL) and Open Biological and Biomedical Ontologies (OBO) format. The Web services provide multi-layered access to the ontology content, from getting all terms in an ontology to retrieving metadata about a term. Users can easily incorporate the NCBO Web services into software applications to generate semantically aware applications and to facilitate structured data collection.

Over the recent years, ontologies are widely used in the biomedical domains. However, biomedical ontology heterogeneity problem hamper the cooperation between intelligent applications based on biomedical ontologies. It is crucial to establish correspondences between the heterogeneous biomedical concepts in different ontologies, which is so-called biomedical ontology matching. Approaches based on Multi-Objective Evolutionary Algorithm (MOEA), such as NSGA-II, are emerging as a new methodology to solve the ontology matching problem. In this paper, to further improve the quality of biomedical ontology alignments, a hybrid NSGA-II is proposed, which modifies the knee solutions in the Pareto front by using a local search method. Experiment utilizes two biomedical ontology matching tracks provided by Ontology Alignment Evaluation Initiative (OAEI 2017). The experimental results show that our approach outperforms the participants of OAEI 2017 and NSGA-II based ontology matching technique.

Ontology: Footstone for Strong Artificial Intelligence.

A kit-build is a framework that utilizes a concept map to visualize knowledge and automatically assess the concept map that was created by the learner by recomposing the kit. In this case study, a kit-build will be implemented in object-oriented programming learning to investigate its usefulness in increasing learning gain compared with a traditional summarizing activity. In this study, learners will practically use a kit-build concept map system to facilitate reviewing and reconstruct their knowledge of object-oriented programming learning resources. This study conducted an experiment using two groups: the summarizing group (n = 46) and the kit-build group (n = 35). The variables that were measured were the pre- and post-test scores and learning gain. To evaluate the significance of the difference between the two groups, a non-parametric ANOVA analysis was performed. The experiment result confirms that incorporating the kit-build concept map as a knowledge visualization technique could improve not only the learning performance but also the overall learning gain compared with the group that used the regular summarizing method.

Numerous studies have proved that graphic strategies, such as graphic organization and concept mapping, can facilitate improving reading comprehension. However, the question as to what graphic strategies can improve argumentative essay reading comprehension ability is not yet resolved. To determine whether graphic strategies can improve students' reading comprehension ability, we designed a computer-aided argumentative essay reading system that can construct graphic strategies. In the designed system, three approaches, namely a traditional teaching approach without graphic strategies, concept mapping, and argument mapping, are created for determining the effects of graphic strategies on students' argumentative essay reading comprehension ability. In addition, the proposed argument mapping system provides a function for helping students identify three key argumentative essay elements, namely claims, reasons, and evidence, to enable them to construct an argument map with no burden. The designed system can help students learn how to read argumentative essays easily, improving their reading comprehension ability. The experimental results from 373 sixth graders showed that the argument mapping method enhanced students' argumentative essay reading comprehension ability compared with traditional and concept mapping approaches. Statistical results revealed that between-group differences were statistically significant (p value between the experimental and Control Group 1 was 0.001 and that between the experimental and Control Group 2 was 0.013).

This paper describes the Karta-Casa project, an attempt to design an agent framework that provides a flexible, extensible and easy-to-use base for constructing various single-user and distributed groupware visualization and concept mapping applications. The structure and the development process behind Karta-CASA are detailed, and several novel aspects of the system are described. Details of several applications based on the framework are provided. Further, this paper suggests several research directions for extending the Karta-CASA system into areas that have not yet been explored by other agent visualization systems.

Ontologies have proven to be useful to enhance NLP-based applications such as information extraction.In the biomedical domain rich ontologies are available and used for semantic annotation of texts.However, most of them have either no or only few non-English concept labels and cannot be used to annotate non-English texts.Since translations need expert review, a full translation of large ontologies is often not feasible.For semantic annotation purpose, we propose to use the corpus to be annotated to identify high occurrence terms and their translations to extend respective ontology concepts.Using our approach, the translation of a subset of ontology concepts is sufficient to significantly enhance annotation coverage.For evaluation, we automatically translated RadLex ontology concepts from English into German.We show that by translating a rather small set of concepts (in our case 433), which were identified by corpus analysis, we are able to enhance the amount of annotated words from 27.36 % to 42.65 %.

The purpose of this study was to review existing principles of oro-facial pain classifications and to specify design recommendations for a new system that would reflect recent insights in biomedical classification systems, terminologies and ontologies. The study was initiated by a symposium organised by the International RDC/TMD Consortium Network in March 2013, to which the present authors contributed. The following areas are addressed: problems with current classification approaches, status of the ontological basis of pain disorders, insufficient diagnostic aids and biomarkers for pain disorders, exploratory nature of current pain terminology and classification systems, and problems with prevailing classification methods from an ontological perspective. Four recommendations for addressing these problems are as follows: (i) develop a hypothesis-driven classification structure built on principles that ensure to our best understanding an accurate description of the relations among all entities involved in oro-facial pain disorders; (ii) take into account the physiology and phenomenology of oro-facial pain disorders to adequately represent both domains including psychosocial entities in a classification system; (iii) plan at the beginning for field-testing at strategic development stages; and (iv) consider how the classification system will be implemented. Implications in relation to the specific domains of psychosocial factors and biomarkers for inclusion into an oro-facial pain classification system are described in two separate papers.

ABSTRACTToday’s scenario of manufacturing and supply chain is full of uncertainty because of numerous types of disruptions and failures such as fire, storm, machine failure, are a few names. Always supply chain disruptions present disastrous impacts, although probabilities of happening are low. In the recent robust as well as flexible supply chain has captured the focus of researchers in designing a supply chain network with consideration of disruption’s risk. In this work, an ontology-based decision support system is proposed to intensify the supply chain resilience during a disruption. The concept of semantic and ontology is adopted in developing the knowledge base for the entire supply chain network including manufacturing units. Protégé is used for defining the classes and sub-classes along with numerous types of properties and expressed in a rule-based system using semantic web rule language (SWRL). Furthermore, a mixed integer linear programming model with an objective of maximising quantified resilience to fulfil the demand. A hybrid particle swarm optimisation – differential evolution (PSO-DE) is utilised as an optimisation technique for the defined problem. In performing the study, a set of simulated data is formed and then interpreted in the ontology for an optimal selection of recovery activity.

Ontologies provide formal, machine‐readable, and human‐interpretable representations of domain knowledge. Therefore, ontologies have come into question with the development of Semantic Web technologies. People who want to use ontologies need an understanding of the ontology, but this understanding is very difficult to attain if the ontology user lacks the background knowledge necessary to comprehend the ontology or if the ontology is very large. Thus, software tools that facilitate the understanding of ontologies are needed. Ontology visualization is an important research area because visualization can help in the development, exploration, verification, and comprehension of ontologies. This paper introduces the design of a new ontology visualization tool, which differs from traditional visualization tools by providing important metrics and analytics about ontology concepts and warning the ontology developer about potential ontology design errors. The tool, called Onyx, also has advantages in terms of speed and readability. Thus, Onyx offers a suitable environment for the representation of large ontologies, especially those used in biomedical and health information systems and those that contain many terms. It is clear that these additional functionalities will increase the value of traditional ontology visualization tools during ontology exploration and evaluation.

This study examined the influence of various conflict management mechanisms embedded into computer‐supported collaborative concept mapping systems on the behaviour and learning of elementary students. Four conflict management mechanisms were compared: an assign design, in which the mapping control was designated to a particular group member; a rotate design, in which the mapping control was rotated among the group members; a give design, in which the child with mapping control could relinquish control to another group member; and an open design, in which every group member simultaneously had mapping control. Ninety‐six fifth and sixth grade Taiwanese students participated in this study. They were arranged into assign, rotate, give, or open three‐member groups to generate collaboratively shared concept maps. Student interactions, attitudes, and achievement were analysed. The results suggest that each conflict management mechanism has a different effect on the elementary students.

In this paper, we propose to add domain knowledge from the most comprehensive biomedical ontology SNOMED CT to facilitate the embedding of EMR symptoms and diagnoses for oral disease prediction. We first learn embeddings of SNOMED CT concepts by applying the TransE algorithm prevalent for representation learning of knowledge base. Secondly, the mapping from symptoms/diagnoses to biomedical concepts and the corresponding semantic relations defined in SNOMED CT are modeled mathematically. We design a neural network to train embeddings of EMR symptoms and diagnoses and ontological concepts in a coherent way, for the latter the TransE-learned vectors being used as initial values. The evaluation on real-world EMR datasets from Peking University School and Hospital Stomatology demonstrates the prediction performance improvement over embeddings solely based on EMRs. This study contributes as a first attempt to learn distributed representations of EMR symptoms and diagnoses under the constraint of embeddings of biomedical concepts from comprehensive clinical ontology. Incorporating domain knowledge can augment embedding as it reveals intrinsic correlation among symptoms and diagnoses that cannot be discovered by EMR data alone.