Search |  Contact |  SRI Home Do not follow this link, or your host will be blocked from this site. This is a spider trap. Do not follow this link, or your host will be blocked from this site. This is a spider trap. Do not follow this link, or your host will be blocked from this site. This is a spider trap.A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A ASRI International.  333 Ravenswood Avenue.  Menlo Park, CA 94025-3493. SRI International is a nonprofit corporation.

AIC Seminar Series

Back to the Drawing Board: The Myth of Data-Driven NLU and How to go Forward (Again)

Walid. S. SabaNeva.ai

Notice:  Hosted by Rodrigo de Salvo Braz

Date:  Thursday, November 2nd 2017 at 4:00pm

Location:  EK255 (SRI E building)  (Directions)

Webex: 

Recording available at:

https://youtu.be/fiud4HHWZ7M
   Abstract

Seminar video available at https://youtu.be/fiud4HHWZ7M Notwithstanding recent advances in machine learning (ML) and other quantitative/data-driven techniques in AI, we believe that these approaches do not scale into tasks that require reasoning at various levels beyond extensional data, and in particular, in natural language understanding.

In this talk we will (i) present evidence that quantitative/data-driven approaches to natural language (NL) are inadequate since what needs to be discovered for a proper interpretation is often not even in the data; (ii) that a quantitative/data-driven approach to NL can lead to absurd interpretations, especially in intensional contexts; and (iii) that what we call ’understanding’ actually happens in situations where there’s no statistical significance in the surface data, and where the required information is quite a distance from the surface data.

With this background we will then briefly suggest what an appropriate approach to tackling the language understanding problem might look like. In this regard, we will highlight where traditional formal semantics might have gone wrong, and, as Jerry Hobbs once noted, make the strong claim that "semantics might become very nearly trivial", if they were grounded in a strongly-typed ontology that reflects our commonsense view of the world and the way we talk about it in ordinary language.

   Bio for Walid. S. Saba

Walid Saba is the Principal AI Scientist at Neva.ai in Menlo Park, CA. He is a Co-Founder of Klangoo, where he was also the CTO for seven years, and has over 20 years of experience in information technology, holding various positions at such places as the American Institutes for Research, AT&T Bell Labs, Metlife, Nortel Networks, IBM and Cognos. He has also spent 7 years in academia teaching and supervising students in Computer Science at Carelton University, the New Jersey Institute of Technology, the University of Windsor, and the American University of Beirut. He has published over 35 technical articles, including an award winning paper that he received in Germany at KI-2008. Walid received an MSc in Computer Science from the University of Windsor, and a PhD in Computer Science which he obtained from Carleton University in 1999.

   On-line Resources