Automatic Ontology Population Task

From GM-RKB
Jump to navigation Jump to search

An Automatic Ontology Population Task is an ontology population task that is an ontology-based information extraction task.

References

2024

  • Perplexity
    • An Automatic Ontology Population Task is an ontology population task that is a closed automatic information extraction task.[2] It involves automatically extracting instances or concepts from text to populate an existing ontology, without human intervention.[1][3][4][5] The key aspects are:
      • It is an automated process of populating an ontology by identifying relevant concepts, instances, and relationships from unstructured text data.[1][3][4]
      • It applies natural language processing and information extraction techniques to analyze text and map the extracted information to the target ontology.[1][5]
      • It is a closed information extraction task, meaning the target ontology schema is known in advance and guides the extraction process.[2]
      • The goal is to semi-automatically or fully automatically extend and enrich the ontology with new knowledge acquired from text, reducing manual effort.[3][4][5]
    • Here are three examples of Automatic Ontology Population Tasks:
      1. Extracting instances of the "Person" class from news articles to populate an ontology about famous people.[1] The system would identify mentions of people's names and extract relevant information like their occupation, birthdate, etc. to create new instances in the ontology.
      2. Populating a product ontology from e-commerce websites and product descriptions.[3] The task would involve extracting the names of products, their categories, features, prices, and other attributes to instantiate the ontology classes and properties.
      3. Enriching a gene ontology by extracting information about genes, proteins, and their functions from biomedical research papers.[5] The system would map textual descriptions to ontology concepts like gene names, protein families, biological processes, etc.
    • Citations:
[1] https://www.sciencedirect.com/science/article/pii/S0167642313003419
[2] http://www.gabormelli.com/RKB/Automatic_Ontology_Population_Task
[3] http://ieeexplore.ieee.org/document/6614374/
[4] https://www.researchgate.net/figure/A-process-for-automatic-ontology-population_fig1_261468422
[5] https://www.semanticscholar.org/paper/Ontology-Population-from-Textual-Mentions%3A-Task-and-Magnini-Pianta/f0a0d1c8a2ff7e76c54a171e29e31fe73f9117c6

2016

  • http://cacm.acm.org/magazines/2016/9/206254-a-new-look-at-the-semantic-web/fulltext
    • QUOTE: Latent semantics: Obviously, there is a lot of semantics that is already on the Web, albeit mostly in text, or in data that machines cannot readily interpret. To complement formally developed ontologies, we must be able to extract latent, evidence-based models that capture the way that users structure their knowledge implicitly. We need to explore these questions: How much of the semantics can we learn automatically and what is the quality of the resulting knowledge? As ontologies are learned or enhanced automatically, what is the very meaning of "formal ontologies"? How do we develop some notion of approximate correctness? Do similar or different reasoning mechanisms apply to the ontologies that are extracted in this way? How do crowdsourcing approaches allow us to capture semantics that may be less precise but more reflective of the collective wisdom?

2010

2009

2008

2006

2003

Retrieved from "http://www.gabormelli.com/RKB/index.php?title=Automatic_Ontology_Population_Task&oldid=910621"