Sunday, February 17, 2008

Semantic Web and Natural Language Processing

I think there is great partnership potential between semantic web's goals and natural language processing (NLP). NLP wants machines and software to be able to read, understand, and create human language. Semantic web wants machines and software to understand what is written on web pages. So really semantic web is what you do after you have conquered NLP. You would then have literate bots "read" the internet. So you can enter questions into search engines rather than keywords.

I'm not the only one with this idea of course. Powerset is actually betting money on it. They working on a semantic search engine powered by NLP.

1 comment:

Median Joe said...

I agree that NLP-based approaches are probably the most viable way to generate semantically-tagged content (e.g. by post-annotating content). It is very much about getting the computers to read and understand to some degree what humans are writing. It seems an uphill battle in the meantime just to get authors to semantically tag the content while it is being created. Will authors rely on search engines to comprehend their writing though? Maybe authors could review/refine a search engine's first-cut at understanding their content.