4375

Автор(ы): 

Автор(ов): 

1

Параметры публикации

Тип публикации: 

Доклад

Название: 

Digital Neural Cortex

Наименование конференции: 

  • Artificial Neural Networks in Engineering Conference (ANNIE-2007, St.Louis)

Наименование источника: 

  • Proceedings of the Artificial Neural Networks in Engineering Conference (ANNIE-2007, St.Louis)

Обозначение и номер тома: 

V. 17

Город: 

  • New York

Издательство: 

  • ASME Press

Год издания: 

2007

Страницы: 

99-106
Аннотация
The ANNIE 2001 Artificial Neural Cortex paper [1] formally introduced the direct/inverse columnar index pair. The power of direct/inverse index pair was clearly shown by the Google search engine, which is based on such a pair, where direct index is a {document}/word ID and the inverse index is {word ID}/document relationship [2]. An efficient way to achieving a generalization of information is in using a hierarchy of indexes. Litespeed Ptd Ltd, Singapore, http://litespeed.com.sg, implemented such a system in 2002, which supports content-based marking of open-ended structured answers with an extremely high level of accuracy. This paper introduces an indexing formalism, which the implementation of the digital cortex is based upon. With the partitioning of the digital cortex into a hierarchy of regions, the cortical algorithm repeats itself on every level, thus serving as a homogeneous indexing tool. The paper also discusses an application of the suggested technology to the scene analyses, which is performed by the digital cortex without training sessions.

Библиографическая ссылка: 

Михайлов А.М. Digital Neural Cortex / Proceedings of the Artificial Neural Networks in Engineering Conference (ANNIE-2007, St.Louis). New York: ASME Press, 2007. V. 17. С. 99-106.