Expected Value Minimization in Information Theoretic Multiple Priors Models

2018 ◽  
Vol 64 (6) ◽  
pp. 3957-3974 ◽  
Author(s):  
Imre Csiszar ◽  
Thomas Breuer
Author(s):  
Ryan Ka Yau Lai ◽  
Youngah Do

This article explores a method of creating confidence bounds for information-theoretic measures in linguistics, such as entropy, Kullback-Leibler Divergence (KLD), and mutual information. We show that a useful measure of uncertainty can be derived from simple statistical principles, namely the asymptotic distribution of the maximum likelihood estimator (MLE) and the delta method. Three case studies from phonology and corpus linguistics are used to demonstrate how to apply it and examine its robustness against common violations of its assumptions in linguistics, such as insufficient sample size and non-independence of data points.


CFA Digest ◽  
2008 ◽  
Vol 38 (3) ◽  
pp. 35-36
Author(s):  
Michael Kobal
Keyword(s):  

Author(s):  
Sean Maw ◽  
Janice Miller Young ◽  
Alexis Morris

Most Canadian engineering students take a computing course in their first year that introduces them to digital computation. The Canadian Engineering Accreditation Board does not specify the language(s) that can or should be used for instruction. As a result, a variety of languages are used across Canada. This study examines which languages are used in degree-granting institutions, currently and in the recent past. It also examines why institutions have chosen the languages that they currently use. In addition to the language used in instruction, the types and hours of instruction are also analyzed. Methods of instruction and evaluation are compared, as well as the pedagogical philosophies of the different programs with respect to introductory computing. Finally, a comparison of the expected value of this course to graduates is also presented. We found a more diverse landscape for introductory computing courses than anticipated, in most respects. The guiding ethos at most institutions is skill and knowledge development, especially around problem solving in an engineering context. The methods to achieve this are quite varied, and so are the languages employed in such courses. Most programs currently use C/C++, Matlab, VB and/or Python.


Sign in / Sign up

Export Citation Format

Share Document