This may come across as an open and opinion-based question, I definitely want to hear expert opinions on the subject, but I am also looking for references to materials that I can read deeply.

One of the ways question answering systems can be classified is by the type of data source that they use:

  1. Structured knowledge bases with ontologies (DBPedia, WikiData, Yago, etc.).

  2. Unstructured text corpora that contain the answer in natural language (Wikipedia).

  3. Hybrid systems that search for candidate answers in both structured and unstructured data sources.

From my reading, it appears as though structured knowledge bases/knowledge graphs were much more popular back in the days of the semantic web and when the first personal assistants (Siri, Alexa, Google Assistant) came onto the scene.

Are they dying out now in favor of training a deep learning model over a vast text corpus like Bert and/or Meena? Do they have a future in question answering?



You must log in to answer this question.