The Overlooked Key to Semantics and NLP


Guest Author Kathy Dahlgren, Ph.D.

We love it when we see articles written about Semantic Natural Language Processing (NLP). With each article, more attention is being focused on companies like Cognition Technologies who are building on their respective NLP foundations. What we don’t see anyone talking about, though, is the key to making NLP work. In our mind, it’s all about the Semantic Map.

Below is an example of a simple semantic map, which Cognition has mapped for all common words in the English language. The semantic mapping of the English language is the key to making NLP effective, and it’s what other NLP companies generally lack. Cognition has been building these resources for the past 23 years.

All Natural Language Processing companies are trying to solve the “relevancy issue” – meaning, how can the relevancy of the text they are processing (either in-bound à reading, or out-bound à delivering text out) be improved for the end-user? This issue is addressed both by technological adjustments to existing solutions (e.g. more sophisticated mathematical and statistical algorithms) and/or by market segmentation, such as through dataset specialization (e.g. automotive, music, videos, medical, etc.). Cognition, on the other hand, addresses the relevancy challenge by changing the NLP paradigm through its unique and complete combination of linguistic elements to optimize semantic understanding:

Morphology • The various forms of word, e.g. singular, plural, tense

Syntax • The grammatical structure, e.g. verbs, nouns

Semantics • Word and sentence meaning

Spelling • The various ways words are spelled (or misspelled)

Semantics and NLP are complex in theory and application, but we think a little education with each article written goes a long way to creating a vital marketplace for all NLP companies and technologies. A greater knowledge of all the components to NLP benefits everyone. We’d love to see sites like AltSearchEngines do write-ups on the various components to NLP like parsing, ontologies, taxonomies, and of course, semantic mapping.

Dr. Kathleen Dahlgren is the Founder and Chief Technology Officer of Cognition Technologies. Currently, she is also an adjunct professor of Linguistics at the University of California, Los Angeles. She was recently part of the Semantic Search panel at the Alternative Search Engines Day in SF with Nagaraju Bandaru of BooRah
Frank Bandach of Eeggi, Barney Pell of Powerset, and William Tunstall-Pedoe of TrueKnowledge.

3 Responses to “The Overlooked Key to Semantics and NLP”

  1. Mark Johnson Says:

    Great post, Dr. Dahlgren! Agreed that we, as semantic technology companies, need to be clearer about what makes our technology unique.

    Here at Powerset, we have a strong focus on syntax, i.e. the structure of language, but we also understand that semantics are very important. For example, if you look at a search like who did texaco buy? you’ll notice that Powerset equates buying and acquiring and knows that if you sold something to someone, then someone bought it.

    Let’s keep the dialog alive about semantic search!

    -mark, powerset product manager

  2. Stephane Says:

    Indeed thank you.

    A common question we are facing everyday is “what do you mean by semantics”. Is is the semantic-web, how about your syntax treatement (hello Powerset). We at SALSAdev have chosen to work on the semantic aspect of Information Processing: there are many very talented companies that takes care of the other components of NLP.

    So we are in great need of clarification. Please continue the dialog and papers. Although we are early in the time-to-market, NLP (and especially semantics ;-) ) will play a key role in tomorrow’s Information Systems.

    Cheers,

    _Stephane, SALSAdev founder.

  3. NLP Training Says:

    Richards best word is Syntax error hard drive crash.

Leave a Reply