The National Library of Medicine has provided Medical Subject Headings (MeSH) for more than 60 years, with the first official list being published in 1954 (https://www.nlm.nih.gov/mesh/intro_preface.html#pref_hist). This has been an invaluable resource and a critical part of search strategies since computerization of literature began. But, as we discuss in a new whitepaper, new technologies such as those now implemented in Qinsight™ go further and are more flexible and timely than MeSH.
Artificial Intelligence (AI), when combined with Computational Linguistics and Computational Statistics, provides a robust framework for discovering meaning and intent from textual content, without having to rely solely on predefined subject headings.
These advanced techniques may still utilize underlying ontologies, but these ontologies need not, and should not, be limited to the categories covered by MeSH. Quertle’s ontology, for example, includes actions (verbs) and other non-technical terms as a means to enhance the “understanding” derived by the AI engine.
Ontologies, whether MeSH or Quertle’s own AI-tuned ontology, are never perfect. For example, you can suggest changes to MeSH to NIH at any time. Generally, MeSH updates are released once a year, such that your suggestion, if accepted, will appear the following year. But, existing articles are not generally re-indexed, so your suggestion will not help with finding old articles, only newly published ones. On the other hand, if you make a suggestion to Quertle, the change will usually appear within 2 weeks and will apply to all past articles. Advanced learning systems, such as AI, really do have their advantages!
Happy Discovery -Jeff Saffer