منابع مشابه
Turing Computability with Neural Nets
This paper shows the existence of a finite neural network, made up of sigmoidal nenrons, which simulates a nniversal Turing machine. It is composed of less then lo5 synchronously evolving processors, interconnected linearly. High-order connections are not required.
متن کاملComputability with Neural Nets
This paper shows the existence of a nite neural network, made up of sigmoidal neurons, which simulates a universal Turing machine. It is composed of less than 10 5 synchronously evolving processors, interconnected linearly. High-order connections are not required.
متن کاملTuring Universality of Neural Nets (Revisited)
We show how to use recursive function theory to prove Turing universality of finite analog recurrent neural nets, with a piecewise linear sigmoid function as activation function. We emphasize the modular construction of nets within nets, a relevant issue from the software engineering point of view.
متن کاملParallel Feedback Turing Computability
In contrast to most kinds of computability studied in mathematical logic, feedback computability has a non-degenerate notion of parallelism. Here we study parallelism for the most basic kind of feedback, namely that of Turing computability. We investigate several different possible definitions of parallelism in this context, with an eye toward specifying what is so computable. For the determini...
متن کاملNon-Turing Computers and Non-Turing Computability*
Various examples of Malament-Hogarth (hereafter, M-H) spacetimes are given in Hogarth (1992), but the following artificial example from Earman and Norton (1993) is perhaps the simplest. Start with Minkowski spacetime (R4, η) and choose a scalar field Ω on M such that Ω=1 outside a compact set C⊂M and Ω tends rapidly to infinity as a point r∈C is approached. The spacetime (R4-r, Ω2η), depicted i...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
ژورنال
عنوان ژورنال: Applied Mathematics Letters
سال: 1991
ISSN: 0893-9659
DOI: 10.1016/0893-9659(91)90080-f