Title :
Integration of Statistical Techniques in the Design Curriculum
Author :
Koushanfar, Farinaz ; Potkonjak, Miodrag
Author_Institution :
Rice Univ., Houston
Abstract :
Traditionally, design and analysis techniques have been using abstractions and models that are well captured using probabilistic models and are treated using standard algorithmic techniques. However, technology progress and application diversification, created a need for considering the impact of non-Gaussian and non-uniform uncertainty distributions that can only be addressed using non-parametric statistical techniques and robust algorithms. This trend provided an impetus for creation of a new graduate level course that aims to teach students sound statistical techniques, to enable them to obtain hands-on experience using statistical software, and to develop and apply the most effective ways to address design and manufacturing uncertainties. We summarize the syllabus of the course that integrates statistical techniques in the design curriculum, identifying the major obstacles and suggesting future improvements.
Keywords :
design engineering; educational courses; engineering education; nonparametric statistics; statistical analysis; statistical distributions; teaching; design curriculum; design uncertainties; graduate level course; manufacturing uncertainties; nonGaussian distribution; nonparametric statistical techniques; nonuniform uncertainty distributions; probabilistic models; standard algorithmic techniques; statistical software; statistical techniques; teaching; Algorithm design and analysis; Application software; Large scale integration; Microelectronics; Parametric statistics; Power system modeling; Predictive models; Semiconductor device modeling; Uncertainty; Wire;
Conference_Titel :
Microelectronic Systems Education, 2007. MSE '07. IEEE International Conference on
Conference_Location :
San Diego, CA
Print_ISBN :
0-7695-2849-X
DOI :
10.1109/MSE.2007.54