The data science debate on machine learning vs domain expertise has gone on for years now, but none of the soothsayers have been able to establish the dominance of one school of thought over the other. As the debate has attempted to polarize the thinkers into two opposing viewpoints, 2012 was the year that turned this heated argument into a more thought-provoking, introspective exercise.
The controversy started with KDNuggets Poll on ML Replacing Domain Expertise, held in April 2012, where people had to share their personal views about machine learning replacing domain expertise. 55% of the poll respondents firmly expressed that in a large number of domains, machine learning can never replace domain expertise.
This poll highlights a very interesting comment from a specific respondent.
Ross Bettinger, on ML replacing Human Domain Expertise:
But I believe that, no matter how thoroughly one or more domain experts are debriefed by knowledge engineers, there will still be unknown unknowns or emerging trends that will not be sensible to a computer program. Until a true AI comes along that can autonomously adapt to new and unanticipated experiences, I vote for the human element and say that ML will not triumph over domain expertise.
On the other side of the spectrum, available information shows that Machine Learning + Big Data together can outperform domain expertise. At Strata 2012, Claudia Perlichs experiences on data mining competitions will lead you to believe that she won the competitions on movie reviews, breast cancer, and customer behavior without any prior (domain) knowledge.
However, the recent poll achieved the triggering of a more rational, more inclusive dialogue between the two polarized groups. This dialogue was perhaps the beginning of a collaborative approach between the two groups to finding participatory problem-solving methods between machine learning and domain expertise. One rejoinder to KDNuggets Poll results shows how far this debate has advanced: