|
Abstract:
|
This talk examines the role of statistics in the cyberinfrastructure proposed by the CS community to integrate computer technologies into the very fabric of science and beyond. Areas of science and text mining are reviewed to motivate the challenges that we face for us to join the cyberinfrastructure design and taking advantage of the parallelism trend in computing. They include developing Exploratory Data Analysis (EDA) for massive data sets, using non-traditional mathematical results for insights in high dimensions, and frameworks for streaming data analysis, data fusion, and taking into account computation, transmission and compression constraints. We conclude that, for the healthy existence of our field, computer technologies must be further integrated into statistics, and statistical thinking must be integrated into computer technologies or into the making of the cyberinfrastructure.
|
