Search code examples
nlpterminologysemanticssemantic-web

Difference between Semantic Web and NLP?

What exactly is the difference between Semantic Web and Natural Language Processing?

Is Semantic Web a part of Natural Language Processing?

Solution

  • These are two separate subject areas but they do overlap in some places. Because documents, regardless of their format are made up of heterogeneous syntax and semantics, the goal is to represent information that is understandable to a machine and not just a human being. This is a common goal of the Semantic Web and Natural Language Processing.

    Semantic Web

    The semantic web is based on two fundamental ideas:

    • Associating meta-information with Internet-based resources. Metadata is pieces of information about other data which can be added explicitly or implicitly.
    • The ability to reason about the meta-information. For example, a machine should be able to recognize that a picture of balloon is not an animal, even if it is shaped like one. This idea of reasoning and inference on textual data is still very experimental, however, showing considerable success. There is a range of techniques to query such information such as SPARQL, Machine Learning (a pre-annotated corpus), and other statistical techniques.

    The use of ontologies is becoming evermore important in this domain. Description Logic provides the mathematical foundation for knowledge representation systems and can be used to reason with the information.

    Natural Language Processing

    Whereas Natural Language Processing is an important and ongoing research area in theoretical computer science and artificial intelligence, it can look beyond the web and process anything from text in PDFs to speaking to your phone. Wikipedia has highlighted some of the key areas which I will expand on:

    1. Automatic (Abstractive & Extractive) Summarisation
    2. Coreference Resolution
    3. Discourse Analysis
    4. Language Translation
    5. Morphological Segmentation
    6. Named Entity Recognition
    7. Natural Language Generation
    8. Optical Character Recognition
    9. Parsing
    10. Question Answering
    11. Relationship (Semantics) Extraction
    12. Speech Segmentation
    13. Speech & Voice Recognition
    14. Topic Segmentation
    15. Word Sense Disambiguation (WSD)
    16. Information Retrieval
    17. Information Extraction
    18. Text Simplification
    19. Spelling Correction

    Both subject areas have been heavily researched into the syntactics of language, both research fields aim to understand language, notably text. However, in recent times the use of semantics has had a lot of time and investment put into it. But in essence, how to represent relationships in text and miscellaneous structures is a top priority of both fields of thought.

    Conclusion

    Semantic Web is mostly annotated with RDF, OWL, etc., whereas NLP really focuses on freeform Text.