ARCHIVED CONTENT
You are viewing ARCHIVED CONTENT released online between 1 April 2010 and 24 August 2018 or content that has been selectively archived and is no longer active. Content in this archive is NOT UPDATED, and links may not function.By ACEDS
In his much touted recent opinion in Rio Tinto v. Vale , US magistrate Judge Andrew Peck noted that, “If [a technology-assisted review] methodology uses ‘continuous active learning’ (as opposed to simple passive learning or simple active learning), the contents of the seed set is much less significant.”
The question some practitioners may be grappling with is, why?
Traditional forms of technology-assisted review (TAR) generally start with the identification of seed sets. These samples of documents, which are chosen from the total document universe, randomly or with keywords, are coded by experts and then used as the reference point to teach the TAR tool how to recognize responsive documents in the larger data population. With some first-generation approaches, the more time and effort put into crafting a seed set, the higher the potential that the documents identified by the tool’s algorithm as responsive will resemble the documents coded as responsive in the seed set. The algorithm relies solely on these initial decisions.