What versus Why. Towards Computing Reality
What versus Why
Towards Computing Reality
Michael L. Brodie, Jennie Duggan
Computer Science and Artificial Intelligence Laboratory (CSAIL) | MIT
April 17, 2014
Abstract
In late 2013, a colleague, Herb Lin, Chief Scientist of the US National Academies of Science, expressed his concern about Big Data. Herb is responsible for overseeing Academy studies for scientific rigor. He was concerned about the growing use of Big Data since Big Data analytics can only suggest What has occurred within a specific probability or confidence level. Big Data analytics does not and cannot say anything about causation, or Why a phenomenon occurred. This observation has led us over the course of six months to some fairly profound observations. First, with Big Data we are moving from discrete, top-down, well understood models (data schemas and computational models) to a new world of computing that is vague, bottom-up, and model-less. In fact if we impose models on Big Data analytics, we may obscure or prevent its greatest value. These concerns are shared with DARPA’s Big Mechanism program.
Download full paper (.PDF): What versus Why.pdf
Links:
– DARPA: BIG MECHANISM SEEKS THE “WHYS” HIDDEN IN BIG DATA
– An automated framework for hypotheses generation using literature, by Vida Abedi, Ramin Zand, Mohammed Yeasin and Fazle Elahi Faisal