Bale, T.A. (1998). Modular Connectionist Architectures and the Learning of Quantification Skills. Unpublished doctoral thesis. Guildford, UK: University of Surrey.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....be obtained through a process of subitization or estimation [9] Of specific interest here is the way in which the internal magnitude representation of number interacts with number articulation. We simulate the linkage of the internal representation of number and its phonetic representation [1] [3]. We use number magnitude as the primary vector, using a 1 dimensional SOM with 66 neurons as a number line. Input to the SOM is formed as a 66 dimensional binary vector where each magnitude is identified with a block of three digits. The collateral vector is formed as a number articulation, with ....

Bale, T.A. (1998). Modular Connectionist Architectures and the Learning of Quantification Skills. Unpublished doctoral thesis. Guildford, UK: University of Surrey.

....in turn, to argue about the potential localisation of what and where tasks in the animal brain. The details related to the architecture 1 , and training and testing of MASCOT, together with the relevant psychological background, have been discussed by one of the authors (TB) in her thesis (Bale, 1998). 1 Architectural details can also be found in Bale and Ahmad (1998) ....

Bale, T. A. (1998) Modular Connectionist Architectures and the Learning of Quantification Skills. (Unpublished PhD Dissertation, University of Surrey, Guildford, Surrey.) Bale, T. A. & Ahmad, K. (1998) Simulation of quantification abilities using a modular neural network approach. Neural Computing and Applications. (Submitted).

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC