Science Among the Machines Workshop

From Santa Fe Institute Events Wiki

Revision as of 18:32, 15 March 2017 by Casey (talk | contribs)

Co-organized by Google

1965 Charleston Road
Mountain View, CA 94043

Tuesday, March 21, 2017

Science Among the Machines


This workshop expands upon some of the the basic questions raised in the topical meeting the day before to ask a series of more technical research and collaboration questions. Broadly, this meeting seeks to explore areas where the expertise in machine learning and the computational resources of Google might be of value to some of the more fundamental questions being asked by SFI complexity research. And importantly to pinpoint areas where computational power and statistical inference has proven insufficient to make progress and to discern why.

The format of the meeting will be brief presentations by the SFI and Google participants and round table discussions that address one or more of the day two objectives described below (day 1 questions are included to help with continuity from the day 1 topical meeting).

1. Day 1: What will happen to efforts at human understanding when computing platforms equal or outperform our best mechanistic sciences? What will this mean for the structure of knowledge and institutions that create, curate, and manipulate knowledge?
Day 2: Can we provide examples of machine-derived predictions or solutions where we no longer seek or require mechanistic models and theory? And are these new solutions robust and scaleable?

2. Day 1: What are the limits of insight and prediction associated with computational science? And how will the scientific revolution of the 17th century be combined with the revolution of the 21st century? Combining the best of fundamental theory with the best of prediction.
Day 2: Can we provide examples of the successful combination of fundamental/parsimonious theory with machine based-prediction that might pave a path forward? That have in other words improved both prediction and explanation.

3. Day 1: Can we map out those areas of activity, from engineering, the economy, health care, and national defense, etc, where computational and data science will continue to be transformative?
Day 2: What are novel data sets and problems of interest to SFI researchers, to include natural and social science, that might benefit from the expertise and resources of Google, and how might we collaborate to solve these problems?