Most models of neural associative memory have used networks with broad connectivity. However, from both a neurobiological viewpoint and an implementation perspective, it is logical to minimize the length of inter-neural connections and consider networks whose connectivity is predominantly local. The `small-world networksa model described recently by Watts and Strogatz provides an interesting ap...

A novel architecture f o r a Multiple-Valued Ferroelectric Content-Addressable Memory (FCAM) is proposed. An FCAM employs ferroelectric capacitors as storage elements to provide a nonvolatile content-addressable memory. A 2-bit search operation is performed by a simultaneous access to a 4-level polarization, hence reducing the total number of search operations to h a y Four FCAM structures are ...

This paper presents a pipelined match-line and a hierarchical search-line architecture to reduce power in contentaddressable memories (CAM). The overall power reduction is 60%, with 29% contributed by the pipelined match-lines and 31% contributed by the hierarchical search-lines. This proposed architecture is employed in the design of a 1024×144 bit ternary CAM, achieving 7 ns search cycle time...

Translation Look-aside Buffer (TLB), which is small Content Addressable Memory (CAM) structure used to translate virtual addresses to physical addresses, can consume significant energy in some architectures. In addition, its power density is high, due to its small area. Consequently, reducing power consumption of TLB is important for both high-end and low-end systems. While a large TLB might be...

This paper presents a Variable-sized Kohonen Feature Map Probabilistic Associative Memory for Sequential Patterns (VKFMPAM-SP) which can realize successive (additional) learning for sequential patterns. In the proposed model, since the connection weights of the neurons which are centers of the area corresponding to the stored data are fixed, new patterns can be memorized without destroying memo...

memory S. Dey Manasi, M. M. Al Rashid, J. Atulasimha, S. Bandyopadhyay, and A. R. Trivedi Department of Electrical and Computer Engineering, University of Illinois at Chicago, Chicago, IL 60607 USA Department of Mechanical and Nuclear Engineering, Virginia Commonwealth University USA, Richmond, VA 23284, USA Department of Mechanical and Nuclear Engineering, Virginia Commonwealth University USA,...

The performance characteristics of five variants of the Hopfield network are examined. Two performance metrics are used: memory capacity, and a measure of the size of basins of attraction. We find that the post-training adjustment of processor thresholds has, at best, little or no effect on performance, and at worst a significant negative effect. The adoption of a local learning rule can, howev...

One o f t h e r e m a r k a b l e t h i n g s a b o u t human i n f o r m a t i o n p r o c e s s i n g i s t h e way w e s t o r e and r e t r i e v e i n f o r m a t i o n . The human a s s o c i a t i v e memory i s a d e v i c e t h a t has been o n l y r a t h e r c r u d e l y a p p r o x i m a t e d w i t h c o m p u t e r s i n f l e x i b i l i t y , s e l f m o d i f i a b i l i t y ,...

We investigate the memory structure and retrieval of the brain and propose a hybrid neural network of addressable and content-addressable memory which is a special database model and can memorize and retrieve any piece of information (a binary pattern) both addressably and content-addressably. The architecture of this hybrid neural network is hierarchical and takes the form of a tree of slabs w...