This Article 
 Bibliographic References 
 Add to: 
Algorithmic Transformations for Neural Computing and Performance of Supervised Learning on a Dataflow Machine
July 1992 (vol. 18 no. 7)
pp. 613-623

Reprogrammable dataflow neural classifiers are proposed as an alternative to traditional implementations. In general, these classifiers are based on functional languages, neural-dataflow transformations, dataflow algorithmic transformations, and dataflow multiprocessors. An experimental approach is used to investigate the performance of a large-scale fine-grained dataflow classifier architecture. In this study, the functional descriptions of high level data dependency of a supervised learning algorithm are transformed into a machine executable low-level dataflow graph. The tagged token dataflow algorithmic transformation is applied to exploit the parallelism. Dataflow neural classifiers are used to implement the learning algorithm. No attempt is made to optimize the granularity of the high-level language programming blocks to balance the computation and communication. The proposed classifier architecture is more versatile than other existing architectures. Performance results show the effectiveness of dataflow neural classifiers.

[1] R. P. Lippman, "An introduction to computing with neural nets,"IEEE ASSP Msg., vol. 4, pp. 4-22, 1987.
[2] J. Backus, "Function level computing,"Spectrum, pp. 22-27, Aug. 1982.
[3] R. S. Nikhil, "Id: A language with implicit parallelism,"Computation Structures Group Memo 305, MIT, Feb. 1990.
[4] S. Herathet al., "Dataflow computing models for high performance neural computing machines," inProc. Int. Conf. Computing and Computation, May 1989.
[5] I.G. Smotroff, "Dataflow architectures: Flexible platform for neural network simulation,"Advances in Neural Information Processing Systems, D.S. Touretzky, Ed. San Mateo, CA: Morgan Kaufmann, 1990.
[6] J. B. Dennis, "Data flow computation," NATO ASI Series, vol. F14,Control Flow and Data Flow: Concepts of Distributed Programming, M. Broy, Ed. New York: Springer-Verlag, 1985, pp. 346-397.
[7] R.S. Nikhil Arvind, "Executing a program on the MIT tagged-token dataflow architecture,"IEEE Trans. Computers, vol. 39, Mar. 1990.
[8] J. Gurd and I. Watson, "Preliminary evaluation of a prototype dataflow computer,"IFIP, 1983, pp. 545-551.
[9] J. Herathet al., "Dataflow computing models, languages, and machines for intelligence computations,"IEEE Trans. Software Eng., vol. 14, pp. 1805-1828, Dec. 1988.
[10] E. Gelenbe,Multiprocessor Performance. New York: Wiley, 1989.
[11] T. Yuba, T. Shimada, Y. Yamaguchi, K. Hiraki, and S. Sakai, "Dataflow computer development in Japan," in Proc. Int. Conf. on Supercomput. ACM Press, 1990, pp. 140-146.
[12] J.-L. Gaudiot and L. Bic,Data-Flow Systems. Englewood Cliffs, NJ: Prentice-Hall, 1991.
[13] Y. Kodamaet al., "Evaluation of the EM-4 highly parallel computer using a game tree searching problems," inProc. Fifth Generation Computer Systems, 1992.
[14] B. Wah and V. Ramamoorthy,Computers for Artificial Intelligence Processing, New York: Wiley, 1990.
[15] I. Radivojevic and J. Herath, "Executing DSP applications in a dataflow environment,"IEEE Trans. Software Eng., pp. 1028-1042, Oct. 1991.
[16] E. Gelenbe, "Random neural networks with negative and positive signals and product solutions,"Neural Computation, pp. 502-510, 1989.

Index Terms:
algorithmic transformations; reprogrammable dataflow neural classifiers; neural computing; performance; supervised learning; dataflow machine; functional languages; neural-dataflow transformations; dataflow algorithmic transformations; dataflow multiprocessors; high level data dependency; machine executable low-level dataflow graph; tagged token dataflow algorithmic transformation; granularity; computerized pattern recognition; learning systems; neural nets; parallel architectures; performance evaluation
S.T. Kim, K. Suwunboriruksa, S. Herath, A. Jayasumana, J. Herath, "Algorithmic Transformations for Neural Computing and Performance of Supervised Learning on a Dataflow Machine," IEEE Transactions on Software Engineering, vol. 18, no. 7, pp. 613-623, July 1992, doi:10.1109/32.148479
Usage of this product signifies your acceptance of the Terms of Use.