An 11-million Transistor Neural Network Execution Engine

Griffin, M.; Tahara, G.; Knorpp, Kurt; Pinkham, R.; Riley, Bernard

doi:10.1109/isscc.1991.689117

Cited by 54 publications

(15 citation statements)

References 1 publication

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The CNAPS2 architecture is a Single-Instruction Multiple-Data (SIMD) array ofprocessors[7J [8]. The design uses multiple processing nodes (PNs) that execute the same operation on different values.…”

Section: The Cnaps Architecturementioning

confidence: 99%

<title>Comparison of neural network classifiers for optical character recognition</title>

Baker¹,

McCartor²

1992

SPIE Proceedings

View full text Add to dashboard Cite

The recognition of handwritten characters is an important technology for document processing and for advanced user interfaces. Recent advances in Artificial Neural Network (ANN) classifiers have shown impressive pattern recognition results when using noisy data. One advantage of ANN algorithms is that they are parallel by design, which allows a natural implementation on high-speed parallel architectures.The availability of standard databases of handwritten characters permits a fair comparison between different OCR classifiers. This paper compares the classification performance of two popular ANN algorithms: Back Propagation and Learning Vector Quantization. A set of digits from the National Institute of Standards and Technology's Handwritten Database is used to test the two classifiers. Each algorithm's execution time and memory efficiency is also compared, based on an implementation for Adaptive Solutions' highly parallel CNAPS architecture. We also show that a fair comparison cannot be made between OCR research that does not use the same set of characters for testing. INTRODUCTIONHandwritten Optical Character Recognition is a problem that has attracted much interest by both academic and commercial researchers. Some commonly cited applications include handwritten insurance forms, bank checks, and tax forms. The difficulty of OCR is caused by the large variety of styles used by different people. The complexity of the problem is illustrated by the fact that a common indication of a person's identity is the shape of their handwriting. It is also noted that many people can not read their own handwriting, much less the writing of another individual. This paper describes research results on a subset of the domain of handwritten OCR. We evaluate the use of two popular artificial neural network classifiers, Back Propagation and Learning Vector Quantization (LVQ), for the use of isolated digit recognition. We first discuss some of the results of previous researchers. We then proceed by describing the ANN algorithms and the recognition results on a common data set. Finally we evaluate the ANN algorithms based on the memory requirements and training time while running on a SIMD parallel architecture. Previous workThere is a lot of activity in applying ANNs to Handwritten OCR[1][2][3]. One of the most important research results in the last few years is the findings by Martin and Pittman [3] on recognition of isolated digits by Back Propagation networks. Martin and Pittman found that one of the most significant factors that influence recognition accuracy is the size of the training set. They made an independent study in 08194-08 158/921$4.00 SPIE Vol. 1661 (1992) / 191 Downloaded From: http://proceedings.spiedigitallibrary.org/ on 06/26/2016 Terms of Use: http://spiedigitallibrary.org/ss/TermsOfUse.aspx

show abstract

Section: The Cnaps Architecturementioning

confidence: 99%

<title>Comparison of neural network classifiers for optical character recognition</title>

Baker¹,

McCartor²

1992

SPIE Proceedings

View full text Add to dashboard Cite

show abstract

“…2 The CNAPS chip has 64 PNs on each chip, and has 4 Kbytes of static RAM local to eh PN. 2 The CNAPS chip has 64 PNs on each chip, and has 4 Kbytes of static RAM local to eh PN.…”

Section: Chip Implementationmentioning

confidence: 99%

<title>Artificial neural network and image processing using the Adaptive Solutions' architecture</title>

Baker¹

1991

Image Processing Algorithms and Techniques II

View full text Add to dashboard Cite

Adaptive Solutions' CNAPS architecture is a parallel array of digital processors. This design features a Single-Instruction Multiple-Data (SIMD) stream architecture. The architecture is designed to execute on-chip learning for Artificial Neural Network (ANN) algcwithms with unprecedented performance. ANNs have shown impressive results for solving difficult image processing tasks. However, current hardware prevents many ANN solutions from being effective products. The CNAPS architecture will provide the computational power to allow real time ANN applications.Because of the high parallelism of the architecture, it is also ideal for digital image processing tasks. This architecture will allow high performance applications that combine conventional image processing methods and ANNs on the same system. This paper gives a brief introduction to the CNAPS architecture, and gives the system performance on implementations of neural network algorithms, and conventional image processing algorithms such as convolution, and 2D Fourier transforms. INTRODUCTIONArtificial Neural Network classifiers have provided robust solutions to many applications that were previously unsolved problems. There have been impressive results of ANN technology in such areas as optical character recognition, speech recognition, signal processing, and computer vision. However, along with the impressive classifier results, there has been an increased concern for pre-processing required for real-world problems. A major area ofresearch has been concerned with how to best represent input data to the ANN classifiers. Pre-processing of the inputs is especially important for image processing tasks, because of the large input space, inherent noise of image data, and the spatial variations of the target objects. For these reasons, ANN researchers are using conventional image processing techniques for pre-processing the input data.One major impediment to implementation of real time ANN classifier is the large number of computations required to run a typical algorithm. The training and execution time for neural networks takes too long on current single processor systems for real time applications. Execution time is also a problem for image processing systems. The CNAPS architecture, designed by Adaptive Solutions, is a highly parallel solution that will allow many ANN applications to run in real time.1 Although the CNAPS architecwre is designed for running and training ANNs, it can also be used to execute many image processing algonthms in real time. The flexibility of the CNAPS architecture provides a powerful platform for implementing complete image processing applications. This paper will describe the CNAPS architecture, as well as the implementation of ANN and image processing algorithms. Section 2 is an overview of the CNAPS architecture. Section 3 explains the implementation and performance of neural network algorithms. Implementation and performance of image processing algorithms is discussed in Section 4, and Section 5 summarizes the paper. THE CNAPS ARC...

show abstract

“…([Pomerlau et al (1988)], [Iwata et al (1990)], [Griffin et al (1991)] and [Wawrzynek et al (1993)]) and consequently, this leads to costly and inefficient computations. Therefore, model reduction is of paramount importance.…”

Section: Introductionmentioning

confidence: 99%

On Stability And Model Order Reduction Of Perturbed Nonlinear Neural Networks

Condon

Grahovski

2008

ECMS 2008 Proceedings Edited By: L. S. Louca, Y. Chrysanthou, Z. Oplatkova, K. Al-Begain

View full text Add to dashboard Cite

In this paper, the qualitative theory of large-scale dynamical systems is surveyed. In particular, the focus is the Hopfield Neural networks both with and without perturbations. Properties relating to asymptotic and exponential stability and instability are detailed. A model reduction technique based on balanced truncation is applied to the neural networks. Its effect on the stability properties of the networks is then examined. A numerical test illustrates some important points.

show abstract

An 11-million Transistor Neural Network Execution Engine

Cited by 54 publications

References 1 publication

<title>Comparison of neural network classifiers for optical character recognition</title>

<title>Comparison of neural network classifiers for optical character recognition</title>

<title>Artificial neural network and image processing using the Adaptive Solutions' architecture</title>

On Stability And Model Order Reduction Of Perturbed Nonlinear Neural Networks

Contact Info

Product

Resources

About