Add new comment

I'm curious to see if the developers "tested" their algorithm "in the past" to see if it would accurately identify current, disruptive technologies. For example, if the algorithm ran in 2005 (and filtered only those academic articles and patents prior to 2006), would it have identified something as "disruptive" as the smartphone? Armed with this information, researchers could conduct further split a/b testing to identify characteristics that could improve accuracy for certain fields/industries of interest.