Scholarly knowledge graphs through structuring scholarly communication: a review

It may state two separate claims as one.

Many comprehensive survey papers exist for knowledge graphs generally and for domain-specific knowledge graphs such as smart grids, industrial products, manufacturing and production, biomedical domain, and recommendation engines.

“In the literature, many comprehensive survey papers for knowledge graph [1], domain-specific knowledge graphs such as smart grids [2], industrial products [3], manufacturing and production [4] biomedical domain [5] and knowledge graphs with recommendation engines [6] exist”

• cites [1] Gutierrez Claudio, Sequeda Juan F (2021) Knowledge graphs. Communications of the ACM 64(3):96–104 sourceSupported
• cites [2] Wang Jian, Wang Xi, Ma Chaoqun, Kou Lei (2021) A survey on the development status and application prospects of knowledge graph in smart grids. IET Generation, Transmission & Distribution 15(3):383–407 sourceSupported
• cites [6] Qin Chuan, Zhu Hengshu, Zhuang Fuzhen, Guo Qingyu, Zhang Qi, Zhang Le, Wang Chao, Chen Enhong, Xiong Hui (2020) A survey on knowledge graph-based recommender systems. Scientia Sinica Informationis 50(7):937–956 sourceNot checked — no full text of the source

Probably not a substantial claim. The paper’s conclusions may not depend on it.

SPAR is widely used in various projects and publications such as [74].

“To fill the gap between domain-specific and semantic publishing ontology, Semantic Publishing and Referencing Ontology (SPAR) is widely used in various projects and publication such as [74].”

• cites [74] Penev Lyubomir, Dimitrova Mariya, Senderov Viktor, Zhelezov Georgi, Georgiev Teodor, Stoev Pavel, Simov Kiril (2019) OpenBiodiv: A knowledge graph for literature-extracted linked open data in biodiversity science. Publications 7(2):1–16 sourceSupported

Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

The embedding vectors contain rich information about entities and relationships and learned embeddings can be used in tasks such as entity classification and link prediction/ knowledge graph completion.

“The embedding vectors contain rich information about entities and relationships and learned embeddings can be used in tasks such as entity classification and link prediction/ knowledge graph completion [97].”

• cites [97] Rossi Andrea, Firmani Donatella, Matinata Antonio, Merialdo Paolo, Barbosa Denilson (2020) Knowledge graph embedding for link prediction: A comparative analysis. ACM Transactions on Knowledge Discovery from Data (TKDD) 15(2):1–49 sourcePartly supported

  Section 2 (p. 3) says KG embeddings represent entities and relationships, embody the graph’s semantics, and can identify new links, supporting the link-prediction portion. The passages found do not establish their use for entity classification; the classifiers described in Section 3.3 (p. 9) score facts or predict missing entities rather than classify entities. (LLM, conf=0.88)

  Experimental — accuracy not yet measured

Probably not a substantial claim. The paper’s conclusions may not depend on it.

There are various conceptual models classified into groups from representation of specific research areas to describing structure of scholarly documents, rhetorical elements and bibliographies [80, 112].

“There are various conceptual models that are classified into groups from representation of specific research areas to describing structure of the scholarly documents, rhetorical elements and bibliographies [80, 112].”

• cites [80] Angelo A. Salatino, Francesco Osborne, and Enrico Motta. Ontology extraction and usage in the scholarly knowledge domain. In Giuseppe Cota, Marilena Daquino, and Gian Luca Pozzato, editors, Applications and Practices in Ontology Design, Extraction, and Reasoning, volume 49 of Studies on the Semantic Web, pages 91–106. IOS Press, 2020 sourcePartly supported

  Section 2.1, “Ontologies of research areas” (page not available), describes models for specific research fields, including computer science, medicine, and physics. The passages do not establish the claimed grouping extending to scholarly-document structure, rhetorical elements, and bibliographies. (LLM, conf=0.86)

  Experimental — accuracy not yet measured
• cites [112] Pertsas Vayianos, Constantopoulos Panos (2017) Scholarly Ontology: modelling scholarly practices. International Journal on Digital Libraries 18(3):173–190 sourceNot checked — no full text of the source

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Nanopublications, i.e., a fine-grained, machine-interpretable, semantic and interlinked representation for article information, provenance and assertions, are presented as an RDF graph.

“To provide simplification and standardization, nanopublications [128], i.e., a fine-grained, machine-interpretable, semantic and interlinked representation for article information (sections, text, figures, tables, formula, footnote and review comments), provenance and assertions is presented as an RDF graph.”

• cites [128] Cristina Iulia Bucur, Tobias Kuhn, and Davide Ceolin. A Unified Nanopublication Model for Effective and User-Friendly Access to the Elements of Scientific Publishing. Knowledge Engineering and Knowledge Management: 22nd International Conference, EKAW 2020 Lecture Notes in Computer Science, 12387:104–119, 2020 sourceSupported

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Table metadata extraction possesses diverse challenges due to lack of standardization.

“For example, table metadata extraction [126] possess diverse challenges due to lack of standardization.”

• cites [126] Allard Oelen, Markus Stocker, and Sören Auer. Creating a Scholarly Knowledge Graph from Survey Article Tables. In: Ishita, E., Pang, N.L.S., Zhou, L. (eds) Digital Libraries at Times of Massive Societal Transition. ICADL 2020 Lecture Notes in Computer Science, 12504:373–389, 2020 sourceSupported

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Various data-driven machine learning algorithms have been widely used in scholarly knowledge graph knowledge acquisition, construction and extracting critical information from vast data sets.

“Various data-driven machine learning algorithms have been widely used in scholarly knowledge graph’s knowledge acquisition, construction and extracting critical information from vast data set.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

Challenges such as heterogeneity and scalability have been handled efficiently with the help of linked data sets of scholarly documents.

“It has been noticed that, challenges such as heterogeneity and scalability have been handled efficiently with the help of linked data sets of scholarly documents.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.

Most open scholarly graphs investigate either implicit or explicit representation, and combining them in a unified knowledge graph remains a challenge.

“It is observed that most of the open scholarly graph investigates either the implicit or explicit representation and combining them in an unified knowledge graph remains a challenge.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

A number of open scholarly infrastructures do not offer services related to bibliographic data such as article citation and need to collaborate with bibliographic databases directly to support digital libraries.

“A number of open scholarly infrastructures do not offer services related to bibliographic data such as article citation and required to collaborate with bibliographic databases directly to support digital libraries.”

It may state two separate claims as one.

The majority of work has considered several domain-specific ontologies and supplemented the data sources by providing patterns with unique instances.

“The majority of work has considered several domain-specific ontologies and supplemented the data sources by providing patterns with unique instances, as seen in Table 5.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.

Machine learning approaches have outperformed human baselines in many specific cases but do not properly integrate with prior knowledge of fine-tuned pre-trained data for interpreting the model's behavior.

“Though machine learning approaches outperformed human baselines in many specific cases but not properly integrates with prior knowledge of fine-tuned pre-trained data for interpreting the model’s behavior.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

As far as concept level extraction is concerned, a handful of studies are focused on extraction of phrases.

“As far as concept level extraction is concerned, a handful studies are focused on extraction of phrases.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

There are only a few research works on fact representation.

“There are only a few research on fact representation that have been published.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

There are only a few research works that focus on extracting relationships between items from scholarly literature.

“However, there are only a few research that focus on extracting relationships between items from scholarly literature.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Studies have used approaches to find new types from unlabeled data.

“Studies, on the other hand, have used approaches to find new types from unlabeled data.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

Each submission in the literature seems to have its own methodology (techniques and phases) and seeks to achieve separate key goals in the knowledge graph construction process.

“Although majority of the approaches are based on supervised and unsupervised learning methods, it is worth noting that each submission seems to have its own methodology (techniques and phases) and seeks to achieve separate key goals in the knowledge graph construction process.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

All the translation-based models depict entities solely on the basis of structural data, ignoring the richness of multi-source data contained in the entity's name, description, category, relationship type and prior knowledge.

“However, all the translation-based models depict entities solely on the basis of structural data, ignoring the richness of multi-source data contained in the entity’s name, description, category, relationship type and prior knowledge.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

A rule-based approach gathers key scholarly information in the form of patterns, leveraging regular expressions in title, abstract, research problems, application areas, and citation information.

“A rule-based approach gathers key scholarly information in the form of patterns, leveraging regular expressions in title, abstract, research problems, application areas, and citation information.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

The construction process incorporates top-down, bottom-up and mixed ways of building knowledge graphs.

“The construction process incorporates top-down, bottom-up and mixed way of building knowledge graphs.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

The latter produces nosier outcomes but can handle huge corpora of scientific documents.

“In comparison, the latter produces nosier outcomes but can handle huge corpora of scientific documents.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

Current approaches utilizing existing ontologies and vocabularies to annotate the context at long text level are semantically far from each other.

“These current approaches utilizing existing ontologies and vocabularies to annotate the context at long text level that are semantically far from each other.”

Probably not a substantial claim. The paper may give no reason why it holds.

Most of the existing approaches for author collaboration focus on semantic similarities using bibliographic metadata such as publication counts and citation network analysis.

“Most of the existing approaches for author collaboration focus on semantic similarities using bibliographic metadata such as publication counts and citation network analysis.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

There are some efforts where the advantages of both rule-based techniques are combined and presented the final form of reasoning to achieve the goal of completing multiple tasks.

“There are some efforts where the advantages of both techniques are combined and presented the final form of reasoning to achieve the goal of completing multiple tasks.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Various studies handle these tasks with the help of OpenIE tools where each KG is generated.

“Various studies handle these tasks with the help of OpenIE tools where each KG is generated.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

The first way to populate knowledge graph generates high-quality and validated outcomes with improved precision-recall analysis.

“The first way to populate knowledge graph generates high-quality and validated outcomes with improved precision-recall analysis.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

As a result of using extractor framework, a huge number of entities and relationships referring to the same concept is detected.

“As a result of using extractor framework, a huge number of entities and relationships referring to the same concept is detected.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

These approaches abandon relevant metadata information such as author organization and venues attended, affecting the quality of the recommendations.

“However, these approaches abandon relevant metadata information such as author organization and venues attended, affecting the quality of the recommendations.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

A wide range of studies using natural language processing techniques can be found that applied over a collection of scientific articles.

“A wide range of studies using natural language processing techniques can be found that applied over a collection of scientific articles.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

Domain-independent KGs are generic within-sentence extraction.

“However, domain-independent KGs are generic within-sentence extraction.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

An article leads to high number of entities when full-text is considered for annotation.

“In addition, an article leads to high number of entities when full-text is considered for annotation.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

Resolving syntax complexities and elimination of ambiguated text are also the part of the NLP extraction pipelines.

“Resolving syntax complexities and elimination of ambiguated text are also the part of the NLP extraction pipelines.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

NLP techniques are widely employed since most strategies rely on the popular pre-trained language model and its modifications to do the extraction task.

“Second, NLP techniques are widely employed since most strategies rely on the popular pre-trained language model and its modifications to do the extraction task.”

It may state two separate claims as one.

The first way to populate knowledge graph suffers from limited scalability issue as well as manual effort consuming.

“However, it suffers from limited scalability issue as well as manual effort consuming.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

Very few studies are using extractor framework such as OpenIE, DyGIE++ and RnnOIE to automatically extract the entities.

“In addition, very few studies are using extractor framework such as OpenIE, DyGIE++ and RnnOIE to automatically extract the entities.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

Translation-based models are the widely used in this domain.

“It has been observed that, translation-based models are the widely used in this domain.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

Techniques like deep neural networks and word2vec are employed to obtain precise syntactic analysis.

“For instance, techniques like deep neural networks and word2vec are employed to obtain precise syntactic analysis.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

For speeding up the extraction process in scientific publications, a collection of natural language processing algorithms supporting OpenIE and ontologies is used to generate an end-to-end automatic pipeline for the generation of knowledge graphs.

“For speeding up the extraction process in scientific publications, a collection of natural language processing algorithms supporting OpenIE and ontologies is used to generate an end-to-end automatic pipeline for the generation of knowledge graphs.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

In order to generate the knowledge graph, issues such as multiple entities refer to same concept, redundant relationships and generic entities are addressed.

“In order to generate the knowledge graph, issues such as multiple entities refer to same concept, redundant relationships and generic entities are addressed.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

Knowledge graph embedding models a triple of the form <Head,relation,Tail> as input, computes matching score and predict the validity of each triplet.

“The knowledge graph embedding models a triple of the form <Head,relation,Tail> as input, computes matching score and predict the validity of each triplet.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

Graph processing capacity and computational powers of graph databases is utilized with the help of graph structure.

“Graph processing capacity and computational powers of graph databases is utilized with the help of graph structure.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Open domain IE has been used to treat any noun phrase as a candidate entity and any verb phrase as a relation candidate.

“Open domain IE, on the other hand, has been used to treat any noun phrase as a candidate entity and any verb phrase as a relation candidate.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

A number of studies used Neo4j for data storage, querying and visualization considerably as compared to the native triple storage platforms.

“On the other hand, a number of studies used Neo4j for data storage, querying and visualization considerably as compared to the native triple storage platforms.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The quote only partly states this claim.

The knowledge graph construction can be classified into categories based on the method used: machine learning, NLP, pattern-based, hybrid and rule-based approaches.

“The knowledge graph construction can be classified into following categories based on the method used:First, studies that intended towards KG development utilizing machine learning techniques to leverage contextual data.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

The performances of existing models drop when they are applied as an embedding learner for entity typing in the task of scholar profiling.

“In addition, the performances of existing models drop when they are applied as an embedding learner for entity typing in the task of scholar profiling.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.

GraphDB is highly efficient in storing and accessing graph database and allows exploring RDF classes to access instances.

“GraphDB is highly efficient in storing and accessing graph database and allows exploring RDF classes to access instances.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

Knowledge graphs have been facilitated with open extraction tools such as OpenIE to feature the knowledge resources.

“In addition, Knowledge graph have been facilitated with open extraction tools such as OpenIE to feature the knowledge resources.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

This service suffers from the challenge of sparsity in data due to insufficient interactions.

“However, this service suffers from the challenge of sparsity in data due to insufficient interactions.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

Several domain-specific ontologies have been widely employed to cover all of the data in various sections of the study.

“Second, several domain-specific ontologies have been widely employed to cover all of the data in various sections of the study.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.

Pattern-based acquisition methods are utilized to acquire salient phrases from research contributions and attain phrasal granularity.

“Pattern-based acquisition methods are utilized to acquire the salient phrases from research contributions and attain phrasal granularity.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

Table 7 presents embedding methods that extract triples where paper and author are the head entity primarily whereas venue, author, field are the tail entity used to generate triple types.

“Table 7 presents embedding methods that extract triples where paper and author are the head entity primarily whereas venue, author, field are the tail entity used to generate triple types.”

Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it.

The extracted information is too generic and require further normalization to remove overlapping and redundancy of extracted information.

“However, the extracted information is too generic and require further normalization to remove overlapping and redundancy of extracted information.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.

One of the applications of Knowledge Graph Embedding models has been reported to give link predictions, which may also be viewed as a foundation for recommendation services.

“One of the applications of Knowledge Graph Embedding models has been reported to give link predictions, which may also be viewed as a foundation for recommendation services.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

In general, knowledge graph embedding model can be categorizes such as translation-distance-based model, neural network-based model and multiplicative model.

“In general, knowledge graph embedding model can be categorizes such as translation-distance-based model, neural network-based model and multiplicative model.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Evaluation and standardization becomes difficult during application on larger scale due to domain-independence and misclassification.

“Moreover, evaluation and standardization becomes difficult during application on larger scale due to domain-independence and misclassification.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

Manually curated rule generation and mappings result in gold standard data, however this curation can be skewed toward certain well-known issues and limited to the expert domain.

“Furthermore, manually curated rule generation and mappings result in gold standard data, however this curation can be skewed toward certain well-known issues and limited to the expert domain.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

Fusion of visual semantics and textual semantics have gradually emerged as new direction in knowledge graphs also.

“Fusion of visual semantics and textual semantics have gradually emerged as new direction in knowledge graphs also.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.

Studies lack coverage for important entity types (e.g., affiliations) and domains (e.g., physics).

“In addition, studies lack coverage for important entity types (e.g., affiliations) and domains (e.g., physics).”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

CNN-based models such as ConvE embedded 2D convolution leads to long training time due to numerous parameters.

“CNN-based models such as ConvE embedded 2D convolution leads to long training time due to numerous parameters.”

It may state two separate claims as one.

Few papers focus majorly on operation for retrieving and manipulations, on the other hand rest focus on storage and visualization as shown in Table 6.

“Few papers focus majorly on operation for retrieving and manipulations, on the other hand rest focus on storage and visualization as shown in Table 6.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Typically two types of directions have been used in the literature to populate the knowledge graph either by human experts or by applying linguistics techniques and machine learning approaches.

“Typically two types of directions have been used in the literature to populate the knowledge graph either by human experts or by applying linguistics techniques and machine learning approaches.”

Probably not a substantial claim.

The general construction of KG has been customized to fit various use-cases of the scholarly domain.

“Aiming to fill this gap between knowledge exploitation ways in the defined domain, the general construction of KG has been customized to fit in various use-cases of the scholarly domain.”

Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.

The basic structure of a large knowledge graph can be easily shown with a limited perspective, but portraying cross-linked sources and exploratory tasks is cumbersome.

“The basic structure of a large knowledge graph can be easily shown with a limited perspective, but portraying cross-linked sources and exploratory tasks is cumbersome.”

Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.

Neo4j query language named Cypher is easy to use as compared to GRAPHQL and various plugins are also available to extend its functionality.

“As Neo4j query language named Cypher is easy to use as compared to GRAPHQL and various plugins are also available to extend its functionality.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

In the literature, three types of embedding methods such as translational embedding method and multiplicative and deep learning-based models are used.

“In the literature, three types of embedding methods such as translational embedding method and multiplicative and deep learning-based models are used.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

Citation data is useful for quantifying bibliometric and trend analysis but offers less information about the content of the paper.

“Citation data, for example, is useful for quantifying bibliometric and trend analysis but offers less information about the content of the paper.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

Neural network models have not gained much popularity in spite of gaining recognizable performance.

“Second, Neural network models have not gained much popularity in spite of gaining recognizable performance.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

Consideration of coreference resolution is also being ignored till date during information extraction and its applications.

“Besides, consideration of coreference resolution is also being ignored till date during information extraction and its applications.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Most of the studies in this section analyzed natural language using basic extraction, mapping, tagging, and parsing technologies.

“It is worth mentioning that most of the studies in this section analyzed natural language using basic extraction, mapping, tagging, and parsing technologies.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.

Knowledge graphs not only measure semantic similarity based on information gained from large corpora but also calculate the semantic path distance between two concepts.

“Knowledge graphs not only measures semantic similarity based on the information gained from large corpora, but also calculates the semantic path distance present between two concepts.”

Probably not a substantial claim. The paper may give no reason why it holds. It may state two separate claims as one.

Knowledge graphs combine the capabilities of concept classes and their instances with the help of ontology and capture in-depth semantic relationships to find similar articles.

“Knowledge graphs combine the capabilities of concept classes and their instances with the help of ontology and capture in-depth semantic relationships to find similar article.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

The first round of searching found 2772 relevant items, and after duplicate removal the selected articles were narrowed to 1630.

“2772 items were found relevant through first round searching and after removal of duplicates, selected articles were narrowed down to 1630.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

Measurement of quality in terms of correctness and completeness is accomplished by sampling with human moderators or crowdsourcing platforms.

“Measurement of quality (correctness and completeness) is accomplished by sampling with human moderators or crowdsourcing platforms.”

• AssumptionRests on an assumption: The rapid growth of scholarly metadata has been largely contributed by humans, which generates blatant errors.

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.

To automatically construct knowledge graphs, three categories-domain-specific, domain-independent, and cross-domain information extraction-can be considered, where input text and output format are crucial.

“In order to automatically construct knowledge graphs, three categories such as domain-specific and domain-independent and cross-domain information extraction can be considered where input text and output format is crucial.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.

The usage of ontology assumes expert input, which leads to bias behavior in favor of precision and increases the cost.

“The usage of ontology assumes expert input, which leads to bias behavior in favor of precision and increases the cost.”

It may state two separate claims as one.

The combination of machine learning, deep learning and linguistic approaches makes comparing approaches complicated and normalization of the learning process even more challenging.

“This makes comparing approaches complicated, and a normalization of the learning process even more challenging.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

All knowledge graph implementations differ in architecture, technology used, and functionality, making it difficult to reach a consensus and define a knowledge graph.

“However, all these implementations differ in architecture, technology used and functionality, making it difficult to reach a consensus and define a knowledge graph.”

It may state two separate claims as one.

External data sources such as DBpedia, which extracts information from Wikipedia, are integrated in the scholarly domain to create extended relationships and support cross-domain text classification tasks.

“Usage of external data sources such as DBpedia, which extracts information from Wikipedia is integrated in scholarly domain to create extended relationships and support cross-domain text classification tasks [39].”

• cites [39] M. Ramakrishna Murty, J. V. R. Murthy, P. V. G. D. Prasad Reddy, and Suresh Chandra Satapathy. A survey of cross-domain text categorization techniques. In 1st International Conference on Recent Advances in Information Technology, RAIT 2012, Dhanbad, India, March 15-17, 2012, pages 499–504. IEEE, 2012 sourceNot checked — no full text of the source

Probably not a substantial claim.

The scholarly document-centric paradigm has been critiqued for not allowing automated knowledge processing, categorization, and reasoning.

“The scholarly document-centric paradigm, on the other hand, has been critiqued for not allowing for automated knowledge processing, categorization, and reasoning.”

Probably not a substantial claim.

Information extraction of scientific entities and connections is required for organizing scientific information into structured knowledge bases.

“As a result, Information extraction (IE) of scientific entities and connections is required for organizing scientific information into structured knowledge bases.”

• AssumptionRests on an assumption: Document and keyword-based information retrieval systems are no longer adequate to explore the insights of the scholarly domain.
• AssumptionRests on an assumption: Automatic knowledge curation from scholarly components such as text, tables, figures, algorithms, and charts is not easy due to improper structure.

Probably not a substantial claim. The paper’s conclusions may not depend on it.

The review's automatic literature search was conducted in ACM Digital Library, Springer, ScienceDirect, IEEE Xplore, and Google Scholar.

“An automatic search was conducted in this study, taking into account digital libraries such as the ACM Digital Library, Springer, ScienceDirect and IEEE xplore. In addition, Google Scholar also produced a robust base of primary literature relevant to the keywords.”

Probably not a substantial claim. The paper may give no reason why it holds.

Traditional graph-based approaches that only deal with structural analysis cannot perform effectively due to the large quantity and dimensionality of scholarly data.

“Due to the large quantity and dimensionality of scholarly data, traditional graph-based approaches that only deal with structural analysis cannot perform effectively.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

A next-level title search to meet relevance criteria filtered the articles to 527.

“Then a next level search was conducted on title to meet the relevance criteria and 527 articles were filtered.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

Recent advancements in knowledge-graph-based representation research focus on knowledge acquisition, knowledge graph creation, triple extraction, triple classification, and knowledge graph completion.

“Recent advancements in the field of knowledge graph-based representation research focuses on the knowledge acquisition, knowledge graph creation, triple extraction, triple classification, knowledge graph completion.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Domain-specific information extraction presents a specific set of scientific concepts that cannot generalize across various domains well.

“This paradigm presents specific set of scientific concepts that can not generalize across various domains well.”

• AssumptionRests on an assumption: Most scientific documents do not share a common set of concepts.

Probably not a substantial claim. The paper may give no reason why it holds.

Knowledge graphs are useful for determining semantic relatedness by taking into account the hierarchical structure of the scholarly network and all forms of semantic relationships between concepts.

“Knowledge graphs are useful for determining semantic relatedness by taking into account the hierarchical structure of the scholarly network as well as all forms of semantic relationships between concepts.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Scholarly documents can be retrieved by crawling and extracting using structural and content-based features.

“Scholarly documents can be retrieved by crawling and extracting using structural and content-based features [19].”

• cites [19] Kong Xiangjie, Jiang Huizhen, Yang Zhuo, Zhenzhen Xu, Xia Feng, Tolba Amr (2016) Exploiting publication contents and collaboration networks for collaborator recommendation. PLoS ONE 11(2):1–13 sourceNot checked yet

  Checked before — its earlier verdict is held back until a new check runs.

Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

The first-level basic search used the keywords TOPIC=(Knowledge graph) AND (Scholarly OR scientific OR literature OR Academic), with time span 2015-2021 and language English.

“For the first level basic search, we investigated for different keywords for such as “TOPIC=( Knowledge graph) AND (Scholarly OR scientific OR literature OR Academic); Time Span: 2015-2021; Language: English”.”

Probably not a substantial claim. The paper may give no reason why it holds.

Information extraction from scientific texts is a critical step in creating fine-grained scholarly knowledge graphs that characterize and connect scholarly articles.

“Information extraction (IE) from scientific texts is a critical step in creating fine-grained scholarly knowledge graphs that characterize and connect scholarly articles.”

• AssumptionRests on an assumption: Automatic knowledge curation from scholarly components such as text, tables, figures, algorithms, and charts is not easy due to improper structure.

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Generating a scholarly network is challenging due to the varied nature of scholarly data models.

“Despite the fact that valid information is easily accessed through the web and open data, generating scholarly network is challenging due to the varied nature of the scholarly data models.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

It is easy to derive entity and relations from the corpus using semi-automatic approaches.

“It is easy to derive entity and relations from the corpus using these semi-automatic approaches.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

More than 140 research articles were examined on the basis of abstract and 70 on the basis of full text.

“We examined more than 140 research articles refined on the basis of abstract and 70 on the basis of full-text of the paper.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

One feature of knowledge graphs is their emphasis on metadata, such as titles, abstracts, authors, and organization contained in research articles.

“One of the features of Knowledge graphs is their emphasis on metadata, such as titles, abstracts, authors, and organization contained in research articles.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Open Information Extraction (OIE) is a way to generate machine readable though domain-independent representation of information in the form of triples and propositions.

“Unlike conventional IE technique, Open Information Extraction (OIE) is a way to generate machine readable though domain-independent representation of information in the form of triples and proposition.”

Probably not a substantial claim. The paper may give no reason why it holds.

Hybrid reasoning, which incorporates the use of ontologies, knowledge completion methods, and schema construction, improves performance of KG creation.

“Hybrid reasoning, which incorporates the use of ontologies, knowledge completion methods, and schema construction, is critical and improves performance of KG creation.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.

The title of a scientific publication follows grammatical rules and includes scientific terminology at certain locations.

“The title of a scientific publications, for example, follows grammatical rules and includes scientific terminology at certain locations.”

• AssumptionRests on an assumption: The title of a scientific publication follows grammatical rules and includes scientific terminology at certain locations.

Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

MAG is a heterogeneous and attributed knowledge graph containing metadata of more than 242M scientific publications, including citations, authors, institutions, journals, conferences, and fields of study.

“MAG: A heterogeneous and attributed knowledge graph containing the metadata of more than 242M scientific publications, including citations, authors, institutions, journals, conferences, and fields of study.”

Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

Ontologies are essential aspects of academic knowledge networks that conceptualize scientific semantic communication.

“Ontologies are essential aspects of academic knowledge networks that conceptualize scientific semantic communication.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.

Knowledge graphs augmented by deep learning and NLP techniques better examine topological relationships and semantic meanings respectively, resulting in notable success in comprehending difficulties in the scholarly domain and retrieving relevant solutions.

“Technically, KG augmented by deep learning and NLP techniques better examines topological relationships and semantic meanings respectively, resulting in notable success in comprehending difficulties in scholarly domain and retrieving relevant solutions.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

Because the scholarly network has billions of nodes and edges, feature engineering and vector-based representation are becoming increasingly popular methods for processing raw data.

“Because the scholarly network has billions of nodes and edges, feature engineering and vector-based representation are becoming increasingly popular methods for processing raw data.”

• AssumptionRests on an assumption: The scholarly network has billions of nodes and edges.

Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

Quality evaluations of ontology are required to meet the criteria of construction of a knowledge graph.

“It is observed that quality evaluations of ontology is required to meet the criteria of construction of knowledge graph.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

The link prediction task helps to improve the recommender system's accuracy and diversity.

“Therefore, the link prediction task helps to improve the recommender system’s accuracy and diversity.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Software Ontology (SO) is used in [60] to extract software mentions by employing a neural network-based classifier in scientific documents.

“In the literature, Software Ontology (SO) is used in [60] that extracts software mentions by employing neural network-based classifier in the scientific documents.”

• cites [60] David Schindler, Benjamin Zapilko, and Frank Krüger. Investigating Software Usage in the Social Sciences: A Knowledge Graph Approach. In: , et al. The Semantic Web. ESWC 2020. Lecture Notes in Computer Science, 12123:271–286, 2020 sourceNot checked yet

  Checked before — its earlier verdict is held back until a new check runs.

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

The goal of knowledge integration is to create ontology and taxonomy to represent hierarchical structure.

“The goal of knowledge integration is to create ontology and taxonomy to represent hierarchical structure.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

In scholarly knowledge graphs, ontologies are core elements that conceptualize scientific semantic communication.

“In scholarly knowledge graphs, ontologies are the core elements that conceptualizes scientific semantic communication.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

OIE focuses on smaller but denser corpora rather than bigger and sparse corpora.

“It focuses on smaller but denser corpora rather than bigger and sparse corpora.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

Limited human supervision regarding the need for hand-crafted rules or human-labeled data set is required.

“It is crucial to highlight that limited human supervision regarding the need for hand-crafted rules or human-labeled data set is required.”

• AssumptionRests on an assumption: Limited human supervision regarding the need for hand-crafted rules or human-labeled data set is required.

Probably not a substantial claim. The paper may give no reason why it holds.

Scientific publications require automatic processing from human-readable format to machine-readable format.

“Scientific publications require automatic processing from human-readable format to machine-readable format.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.

Recommendation can be interpreted as a knowledge graph completion problem where various translational and semantic matching-based embedding methods outperformed.

“Recommendation can be interpreted as a knowledge graph completion problem where various translational and semantic matching-based embedding methods outperformed.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

In general, NLP tasks such as document summarization, fact verification and retrieval require taking huge data and pruning over different document contexts.

“In general, NLP tasks such as document summarization, fact verification and retrieval requires to take huge data and pruning need to be perform over different document contexts.”

Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

The preset entity and relationship model graph may considerably improve the building quality and application efficiency of knowledge graphs in the scholarly domain.

“The preset entity and relationship model graph may considerably improve the building quality and application efficiency of knowledge graphs in the scholarly domain.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

Use of common ontologies and general metadata from schema.org is required to ensure the quality of the knowledge graph.

“Use of common ontologies and general metadata from schema.org is required to ensure the quality of the knowledge graph.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Semantic web technologies enable the display of domain information in machine-readable ways.

“Semantic web technologies, which describe domain knowledge using diverse concepts such as ontologies, Open Information extraction (OpenIE) tools and query processing languages, enable the display of domain information in machine-readable ways.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

Rules can be predicted by applying statistical reasoning approach directly to filter, select and organize candidate tuples.

“Second, rules can be predicted by applying statistical reasoning approach directly to filter, select and organize candidate tuples as shown in Fig. 4c.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.

Knowledge graphs are integrated as an information source to improve recommendations and inherently provide more interpretability in knowledge representation.

“In the literature, knowledge graphs are integrated as an information source to improve recommendations and inherently provides more interpretability in knowledge representation.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

During construction of scholarly knowledge graphs, erroneous facts about entities and/or relationships may be collected, especially when using information from heterogeneous sources with varied properties.

“During the construction of SKG, erroneous facts about entities and/or relationships may be collected. This technique is prone to errors, especially when using information collected from data sources of heterogeneous sources with variety of properties.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

Applying a gold standard produces extremely accurate and reusable findings but is expensive to construct.

“Applying gold standard, on the other hand, produces extremely accurate and reusable findings, but it is expensive to construct.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.

OpenIE is unable to distinguish different surface forms for the same object or relation, resulting in poor aggregation performance.

“However, it is unable to distinguish different surface forms for the same object or relation, resulting in poor aggregation performance.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Knowledge graph construction incorporates standard NLP tasks such as semantic role labeling, part-of-speech tagging, and chunking to get the best system along with pre-trained model.

“Knowledge graph construction incorporates standard NLP tasks such as semantic role labeling, part-of-speech tagging, and chunking to get the best system along with pre-trained model as shown in Fig. 4b.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Semantic enrichment entails the incorporation of metadata from scientific publications from many perspectives, compared to methods focusing exclusively on keywords or feature extraction.

“Semantic enrichment entails the incorporation of metadata from scientific publications from many perspectives, as compared to the method that focuses exclusively on keywords or feature extraction.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

To evaluate correctness of classification assigned to concepts in NG-PL [52], a subset of 1000 entities was annotated by six human annotators.

“To evaluate the correctness (Is the information correct?) of the classification assigned to the concepts in NG-PL [52], subset of 1000 entities are annotated by six human annotators.”

• cites [52] Rabah A. Al-Zaidy and C. Lee Giles. Extracting semantic relations for scholarly knowledge base construction. In 12th IEEE International Conference on Semantic Computing, ICSC 2018, Laguna Hills, CA, USA, January 31 - February 2, 2018, pages 56–63. IEEE Computer Society, 2018 sourceNot checked — no full text of the source

Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

The description of various concepts and objects, as well as their relationships, is used in the formation and understanding of ontology.

“The description of various concepts and objects, as well as their relationships, is used in the formation and understanding of ontology.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

Since knowledge graphs have emerged as a technology with broad application areas, it seeks integration with standard third-party resources such as ontology and vocabularies.

“Since knowledge graphs have emerged as an technology with broad application areas, it seeks integration with standard third-party resources such as ontology and vocabularies.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

Certain filtering issues such as sparsity, diversity and cold-start have not been taken into account in scholarly knowledge graph recommendation.

“To best of our knowledge, certain filtering issues such as sparsity, diversity and cold-start have not been taken into account.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.

Named entities are used to represent general domains, and KGs are constructed through entity and relation extraction often.

“Named entities are used to represent general domains and KGs are constructed through entity and relation extraction often.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Graph-based knowledge representation involves querying and reasoning mechanisms for transparent and interpretable explanations [103].

“Graph-based knowledge representation involves with querying and reasoning mechanisms for transparent and (human and machine) interpretable explanations [103].”

• cites [103] Rui Zhang, Dimitar Hristovski, Dalton Schutte, Andrej Kastrin, and Marcelo Fiszman. Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID- 19 . The COVID-19 resource centre is hosted on Elsevier Connect , the company ’ s public news and information . (January), 2020Not checked yet

Probably not a substantial claim. The paper’s conclusions may not depend on it.

In [59], two subsets from the dataset were created to provide ratings by physicians to analyze relatedness of entities and to finalize the embedding method.

“In [59], two subset from data set are created to provide ratings by physicians to analyze relatedness of entities and to finalize embedding method.”

• cites [59] Sayantan Basu, Sinchani Chakraborty, Atif Hassan, Sana Siddique, and Ashish Anand. ERLKG: Entity Representation Learning and Knowledge Graph based association analysis of COVID-19 through mining of unstructured biomedical corpora. pages 127–137, 2020 sourceNot checked yet

  Checked before — its earlier verdict is held back until a new check runs.

Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.

As we are transitioning from big data to semantic data, KGs play an important role as critical component of semantic web.

“As we are transitioning from big data to semantic data, KGs play an important role as critical component of semantic web.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.

OIE models rely on unsupervised information extraction techniques and are pre-trained on heterogeneous datasets.

“OIE models relay on unsupervised information extraction techniques and pre-trained on heterogeneous datasets.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.

These applications are limited to the expert's domain, and because the expert knowledge base is heavily reliant on experts' experiences, it is difficult to transform it across domains.

“These applications are limited to the expert’s domain, and because the expert knowledge base is heavily reliant on experts’ experiences, it is difficult to transform it across domains.”

Probably not a substantial claim. The paper may give no reason why it holds.

Generating a natural language description from KG makes the stored information more accessible to a wider group of end users in terms of question responding and interpretability.

“These researches have been included in this section because generating a natural language description from KG makes the stored information more accessible to a wider group of end users in terms of question responding and interpretability.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

CORD-19 contains information about 63,000 research articles related to COVID-19, SARS-CoV-2 and other similar corona viruses.

“CORD-19: The COVID-19 Open Research Dataset (CORD-19) contains information about 63,000 research articles, related to COVID-19, SARS-CoV-2 and other similar corona viruses and from the Allen Institute for AI.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

Rules play a crucial role in automated reasoning and finding inferences.

“Rules plays crucial role in automated reasoning and finding inferences.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

Information extraction of scientific documents requires understanding of the full document, unlike traditional sentence-level extraction.

“Information extraction of scientific documents is different from the traditional extraction methods as the understanding of full document is required compared to sentence level extraction.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

Accuracy, AUC, Hits@k and MRR are better choices for evaluation during refinement.

“Other metrics, e.g., accuracy, area under curve (AUC), Hits@k and MRR, etc. are observed as better choice for evaluation during refinement.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

SemSur is a comprehensive ontology for capturing the content of computer science articles and representing it in a semantic and machine interpretable format.

“SemSur (Semantic Survey Ontology) is a new ontology for modeling components of research contributions in the domain of Semantic Web. It is a comprehensive ontology for capturing the content of computer science articles and represent it in a semantic and machine interpretable format.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Precision and recall are frequently used in information extraction and knowledge graph construction with machine learning methods.

“With respect to evaluation methods, precision and recall are quite frequently used in information extraction as well as knowledge graph construction with machine learning methods.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Computer Science Ontology [80] is an ontology for describing higher-level Computer Science study fields, as well as sub-topics and words that go with them.

“Similarly, Computer Science Ontology [80] is an ontology for describing higher-level Computer Science study fields, as well as the sub-topics and words that go with them.”

• cites [80] Angelo A. Salatino, Francesco Osborne, and Enrico Motta. Ontology extraction and usage in the scholarly knowledge domain. In Giuseppe Cota, Marilena Daquino, and Gian Luca Pozzato, editors, Applications and Practices in Ontology Design, Extraction, and Reasoning, volume 49 of Studies on the Semantic Web, pages 91–106. IOS Press, 2020 sourceNot checked yet

  Checked before — its earlier verdict is held back until a new check runs.

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

A mainstream technique in rule-based reasoning is to formalize the problem and obtain inferences as per predefined rules.

“A mainstream technique in rule-based reasoning is to formalize the problem and to obtain the inferences as per predefined rules.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

Manual intervention is still an essential step as it helps create gold standard data set generation for evaluation purposes.

“However, manual intervention is still an essential step as it helps create gold standard data set generation for evaluation purposes.”

• AssumptionRests on an assumption: Manual intervention is still an essential step as it helps create gold standard data set generation for evaluation purposes.

Probably not a substantial claim. The paper’s conclusions may not depend on it.

The knowledge graph's reasoning capacity is reduced by its insufficient integration of subjective information from the literature.

“However, the knowledge graph’s reasoning capacity is reduced by its insufficient integration of subjective information from the literature.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

The fundamental benefit of using openIE paradigms is that they can be simply applied to big corpora with no need for training data.

“The fundamental benefit of using openIE paradigms is that they can be simply applied to big corpora with no need for training data.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

Knowledge fusion helps in generating metadata from various data sources.

“Knowledge fusion helps in generating metadata from various data sources as well.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

FOAF ontology is used in [71, 73] to materialize implicit knowledge about the social relationships of authors and scientists.

“A Friend Of A Friend (FOAF) ontology is used in [71, 73] to materialize implicit knowledge about the social relationships of authors and scientists.”

• cites [71] Francisco Abad-Navarro, José Antonio Bernabé-Diaz, Alexander García-Castro, and Jesualdo Tomás Fernandez-Breis. Semantic publication of agricultural scientific literature using property graphs. Applied Sciences, 10(3), 2020 sourceNot checked — no full text of the source
• cites [73] Zárate Marcos, Rosales Pablo, Braun Germán, Lewis Mirtha, Fillottrani Pablo Rubén, Delrieux Claudio (2019) OceanGraph: Some Initial Steps Toward a Oceanographic Knowledge Graph. Communications in Computer and Information Science 1029:33–40 sourceNot checked — no full text of the source

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

Various scholarly knowledge graphs have built their own datasets by crawling data from various digital libraries, including Web Of Science, GRID, PharmaGKB, Dimensions among others.

“Various scholarly knowledge graphs have built their own datasets by crawling data from various digital libraries, including Web Of Science, GRID, PharmaGKB, Dimensions among others.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

Subjects and objects are also used to identify concepts and their attributes in scientific statements guided by the ontology.

“However, subjects and objects also used to identify concepts and their attributes in scientific statements guided by the ontology.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

Rules are commonly utilized to express links and dependencies between entities in datasets and to capture underlying patterns in data.

“To express links and dependencies between entities in datasets and to capture the underlying patterns in data, rules are commonly utilized.”

It may state two separate claims as one.

The domain-specific extraction process requires domain experts and annotators, which makes the extraction process costlier and limited.

“The domain-specific extraction process requires domain experts and annotators, which makes the extraction process costlier and limited.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

To understand the ability of a model's mechanism, extensively labeled data and a pre-trained model are required.

“To understand the ability of model’s mechanism, extensively labeled data and pre-trained model is required.”

Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.

Ontology allows refinement of the structure of knowledge and reduces conceptual ambiguity.

“Ontology allows refinement of structure of knowledge and reduces conceptual ambiguity.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

Studies based on knowledge extraction from unstructured data sources such as images, tables and pseudocode of an algorithm are limited to date.

“Studies based on knowledge extraction from unstructured data sources such as images, tables and pseudocode of an algorithm is limited to date.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Through tracing over a knowledge graph, an XAI system assists stakeholders in conceptually understanding the workings of associated systems to achieve explainable outcomes and interpretability.

“Through tracing over KG, the XAI system assists stakeholders in conceptually understanding the workings of associated systems in order to achieve explainable outcomes and interpretability.”

Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

The level of data integration is immature and fragmented due to redundancy till now.

“The level of data integration is immature and fragmented due to redundancy till now.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

Many studies pay attention to extraction from structured or semi-structured data sources.

“Many studies pays attention to the extraction from structured or semi-structured data sources.”

Probably not a substantial claim. It may be too vague to check against evidence. It may state two separate claims as one.

Overall, knowledge graph is an important advancement and has power to provide semantically structured information to huge scholarly domain.

“Overall, we are able to conclude that knowledge graph is an important advancement and have power to provide semantically structured information to huge scholarly domain.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.

Knowledge graphs have emerged as the illustration of many real-time applications and implemented practically to classify entities and relationships.

“In recent time, knowledge graphs have been emerged as the illustration of many real-time applications and implemented practically to classify entities and relationships.”

Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

Mediating the quality of algorithmic outputs produced by knowledge graph construction module is a challenging task.

“Mediating the quality of algorithmic outputs produced by knowledge graph construction module is a challenging task.”

Probably not a substantial claim. The paper may give no reason why it holds.

The scientific information provided in the literature does not fulfill the FAIR Data Principles.

“The scientific information provided in the literature, on the other hand, does not fulfill the FAIR Data Principles.”

Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

It is improbable that involved algorithms will be evaluated based on human designed gold standard and human annotators in prospecting years.

“It is improbable that involved algorithms will be evaluated based on human designed gold standard and human annotators in prospecting years.”

Probably not a substantial claim. It may be too vague to check against evidence.

Several aspects of applying knowledge graph concepts into specific domains remain to be explored.

“However, efforts of applying such concepts into specific domains have been made in recent years, several aspects remain to be explored.”

Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it.

FAIR guiding principle (Findable, Accessible, Interoperable, Reusable) [123] is an important scientific merit for Scientific knowledge graphs.

“FAIR guiding principle (Findable, Accessible, Interoperable, Reusable) [123] is the important scientific merit for Scientific knowledge graphs.”

• cites [123] Mark D. Wilkinson, Michel Dumontier, IJsbrand Jan Aalbersberg, Gabrielle Appleton, Myles Axton, Arie Baak, Niklas Blomberg, Jan Willem Boiten, Luiz Bonino da Silva Santos, Philip E. Bourne, Jildau Bouwman, Anthony J. Brookes, Tim Clark, Mercè Crosas, Ingrid Dillo, Olivier Dumon, Scott Edmunds, Chris T. Evelo, Richard Finkers, Alejandra Gonzalez-Beltran, Alasdair J.G. Gray, Paul Groth, Carole Goble, Jeffrey S. Grethe, Jaap Heringa, Peter A.C. t Hoen, Rob Hooft, Tobias Kuhn, Ruben Kok, Joost Kok, Scott J. Lusher, Maryann E. Martone, Albert Mons, Abel L. Packer, Bengt Persson, Philippe Rocca-SerrNot matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

Many scholarly ontologies are restricted over certain domain-specific entity sets.

“Many scholarly ontologies are restricted over certain domain-specific entity sets.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Lack of a unique normalized identity of an author entity makes the problem more challenging for services such as expert finding and collaborator search.

“Lack of a unique normalized identity of an author entity makes the problem more challenging for certain services such as expert finding and collaborator search.”

Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

Maintaining heterogeneity along with embedding in order to map links into low dimensional order is a great challenge.

“Maintaining heterogeneity along with embedding in order to map links into low dimensional order is a great challenge.”

Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

Knowledge graph in scholarly domain is the specific area in which semantic representation of literature-based discovery is presented.

“In this context, knowledge graph in scholarly domain is the specific area in which semantic representation of literature-based discovery is presented.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

Metadata used in construction of knowledge graphs suffers from data incompleteness to different degrees such as affiliation ambiguity.

“Metadata used in construction of knowledge graphs suffers from data incompleteness to different degrees such as affiliation ambiguity.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.

Extraction of structured knowledge is a challenge across the board, and one reason is data ingestion from multiple resources which makes the knowledge noisy and inconsistent.

“Extraction of structured knowledge is a challenge across the board and one of the reason for this is data ingestion from multiple resources which makes the knowledge noisy and inconsistent.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

Integration of information from heterogeneous sources can cause labor-intensive human annotations to train knowledge extraction systems.

“Integration of information from heterogeneous sources can cause labor-intensive human annotations to train knowledge extraction systems.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.

Researchers seeking relevant information have to deal with multiple data sources as well as unstructured search.

“Another issue is that researchers seeking relevant information have to deal with multiple data sources as well as unstructured search.”

Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

A unique and persistent identifier is required to identify relevant digital objects that possess human-readable label feature.

“Second, a unique and persistent identifier is required to identify the relevant digital objects that possess human-readable label feature.”