Sam Reid
نویسنده
چکیده
Pascal’s mechanical calculator and Babbage’s Analytic Engine set the stage for a great deal of modern computing machinery. Modern versions of these concepts were finally realized in von Neumann’s architecture in about 1950. However, the science of computing does not depend on physical implementations of these ideas, but rather on virtual machines—that is, abstractions that perform computations with a well defined semantics. In 1937, well before the advent of the stored program computer, Turing gave a complete description of a computational device which could be programmed and executed and whose properties could be inspected at runtime. This Turing Machine, and the program for the Universal Turing Machine were the first formal descriptions of virtual machines. Wolfram generalized and simplified the idea of a program, depicting a class of virtual machines known as Cellular Automata. The Java Virtual Machine was invented as a high-level conceptual machine with a great deal of internal structure, homogeneity and metadata. The Turing Machine, Cellular Automata and Java are all optimized for different aspects of programming. This paper describes each of these virtual machines, their histories and the relationships between them. The implementation of a simple logical function is described in several virtual machines, and a formalism is presented which describes transformations between machines and programs. In Appendix I, the language of mathematics is examined with respect to its machine-like properties.
منابع مشابه
A Review of Heterogeneous Ensemble Methods
Several ensemble methods have been proposed that can accommodate differing base model types. This document reviews the recent literature, and for each method, we identify (1) main contributions, (2) theoretical motivation, (3) empirical results and (4) relationships to other techniques.
متن کاملBayesian Model Averaging with Cross-Validated Models
Several variants of Bayesian Model Averaging (BMA) are described and evaluated on a model library of heterogeneous classifiers, and compared to other classifier combination methods. In particular, embedded cross-validation is investigated as a technique for reducing overfitting in BMA.
متن کاملOptimization and Neural Networks
Artificial Neural Networks are a supervised machine learning technique with a number of drawbacks. The drawbacks fall into the categories of topology selection, optimization and manual tuning. These drawbacks can be partially overcome in a recently proposed technique that reformulates the problem as a convex optimization
متن کاملDefining cachexia in a renal population.
Publisher rights © 2015 European Dialysis and Transplant Nurses Association/European Renal Care Association This is the peer reviewed version of the following article: Reid J, Noble H, Davenport A, Farrington K, Fouque D, Porter S, Seres D, Shields J, Slee A, Witham M.D., Wright M, Maxwell A.P.. ( 2015). DEFINING CACHEXIA IN A RENAL POPULATION. Journal of Renal Care 41(2), 79–80, which has been...
متن کاملTowards a theory of delusional thought: a connectionist model of an adaptive categorization disorder
Acknowledgements Many thanks to Sam Joseph for discussion, encouragement and ideas of how to put the components of the model together. Many thanks to Al Reid for discussion, encouragement and insights into mental illness. Thanks to my rst and second supervisors, Dr Mark Ellison and Dr Alex Lascarides. Lastly I acknowledge the support of my friend Ruth Cornick. \Happy the Hare in the morning, fo...
متن کامل