Figure 7 - available via license: Creative Commons Attribution 4.0 International
Content may be subject to copyright.
Source publication
Over the last several years, in parallel with the general global advancement in mobile technology and a rise in social media network content consumption, multimedia content production and reproduction has increased exponentially. Therefore, enabled by the rapid recent advancements in deep learning technology, research on scene graph generation is b...
Contexts in source publication
Context 1
... the recall of the proposed method was lower than that of the G F1-Score was higher. An example of the proposed method is shown in Figure 7. The first image sho triple <Subject, Predicate, Object> obtained through the LSTM. ...
Similar publications
In recent times, the detection of hate-speech, offensive, or abusive language in online media has become an important topic in NLP research due to the exponential growth of social media and the propagation of such messages, as well as their impact. Misogyny detection, even though it plays an important part in hate-speech detection, has not received...
Citations
The knowledge extraction helps to create knowledge from structured and unstructured sources and the extracted knowledge need to be in a machine-understandable format for effective inferencing. Ontology learning has become a complex process because of the huge collection of data and its structural complexity. Ontology learning plays a major role in representing the extracted knowledge in various formats such as eXtensible Markup Language (XML), Resource Description Format Schema (RDFS), and Web Ontology Language (OWL) for easy understanding, querying, enhancing usability, and proving unique identification. It simplifies ontology development through various techniques for automatic or semi-automatic knowledge extraction. Deep learning enhances the reasoning efficiency by converting the ontology represented knowledge into logical specification. This paper summarizes the importance of deep learning in ontology development and also discusses the various applications of deep ontological learning. 1 Introduction In the modern digitalized world, the volume of data is getting increasing day by day due to the usage of enormous software applications. Earlier, data was generated only from limited resources like the business world for promoting their product, scientific research works, etc. The COVID-19 pandemic made an enormous change in everyday life and became a major cause for the generation of voluminous data, especially in the medical field. In 2025, it was estimated that the worldwide data hits 175 zettabytes, denotes 61% compound annual growth rate, 51% of the data will be in data centers and 49% will be in the public cloud [1].The data growth necessitates the storage and representation of various knowledge sources such as corpus, dictionaries, relational databases, web, etc., for feature extraction and to discover the novel hidden patterns in the form of knowledge. Natural Language Processing (NLP) is the widely used technique for extracting knowledge from huge unstructured text and social media [2]. The influence of ontology, a conceptual knowledge base specification further enhances NLP-based prediction by focusing on conceptual context to derive more complex inferences and rules [3]. This lead to the generation of ontologies in multiple domains for reuse. The Bioportal is one of the widely used resources for biomedical ontologies [4]. Ontology learning helps to learn and extract the domain-specific concepts automatically by simplifying the ontology development process through various methods and techniques. ASIUM, Text-to-Onto, TextStorm/Clouds, SYNDIKATE, OntoLearn, CRCTOL, and OntoGain are some of the ontology learning systems for extracting semantic knowledge [5]. Semantic web standards such as RDFS, OWL, SPARQL assist ontologies by exchanging data on the web, representing semantic context, and effective retrieval of data [6]. The ontology combined with deep learning algorithms improve the learning model efficiency in terms of similarity-based semantic prediction, to infer more complex heterogeneous features by deriving complex logical rules [7]. The deep learning model with ontological background discovers meaningful, hidden, and valuable unknown patterns to guide decision making. This paper discusses various ontology learning techniques, deep learning frameworks, and ontology-based deep learning models used in numerous applications. 2 Types of Ontology Ontologies are classified into two categories based on expressiveness and structure, and content and application. The lightweight and heavyweight ontologies are based on the expressiveness and structure. The knowledge representation ontology, top-level ontology, linguistic ontology, domain ontology, and application ontologies are
In the era of big data, China’s shipping industry is developing rapidly, but the level of shipping information management is still at a low level, which becomes a hindering force for shipping industry to advance to a higher level. In order to solve this drawback, this paper combines the technology of semantic web and shipping information, constructs the shipping RDF semantic model and realizes the shipping semantic information retrieval to achieve the purpose of improving the efficiency of shipping information management.KeywordsSemantic WebRDF ModelShipping Information
In recent years, complex machine learning and deep learning algorithms have been used in prognostics and health management applications to monitor the state of industrial systems in real-time and predict failures in advance. However, the performance of these algorithms depends on the size of the collected labeled datasets. This study proposes a semi-weakly supervised learning method that creates label functions using a small amount of data and, consequently, combines weak supervision and semi-supervision methods to expand the labeled dataset for training. The proposed model training method creates labels automatically in the life prediction dataset of rolling element bearings, which were manually labeled by expert groups. The performance of the deep neural network is continuously improved through experiments, and a performance improvement similar to that of manually labeled training data is presented. Future research applications include meta-learner experiments (e.g., use of pre-trained word embedding or other model architectures) and more appropriate selection methods and definitions for the basic learner.
