In their article "The Science of Brute Force" (Aug. 2017), Marjin J.H. Heule and Oliver Kallmann humorously asked whether a mathematician using brute force is really "a kind of barbaric monster." While applying simplistic approaches to complex domains (such as image and speech processing) is inefficient, certain specific computational problems do indeed benefit from brute force. In this regard, the mathematician who uses brute force is simply functioning as a good engineer intent on solving problems efficiently.
My primary focus during my Massachusetts General Hospital fellowship (2013–2016) was analyzing the electronic health records for 314,292 patients.2 To identify biomarkers associated with outcomes, my colleagues and I were initially interested in knowing the smoking status of all of them—current, past, or never—for our prediction models. Smoking status is typically documented in clinical narrative notes as free text, and, as reported throughout the literature, classification accuracy of current methods is poor.
No entries found