Abstract
Three key points are made in this paper. The first is that “iconic” states arise naturally in all recursive neural nets which are trained on world states. The second point is that this leads to a new paradigm of cognitive representation. Language is seen as a vehicle for retrieving iconic representations in a recursive system — this is dubbed the “iconic hypothesis”. The third point is that while a similar idea has been presented as a “symbol grounding problem” (Harnad, 1990) the “iconic hypothesis” goes further in suggesting that a recursive neural system can operate in both a symbolic fashion and use grounded internal states. To illustrate these points we introduce an architectural concept (the Neural State Machine Model — NSMM) which allows a clear formalisation of the concept of iconic representations. Examples of the application of this concept to representation of visuo-linguistic data are given.
Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
