Perfect search


In Google’s 2008 annual letter, Sergey Brinn writes “Perfect search requires human-level artificial intelligence, which many of us believe is still quite distant.” This seemingly cautious statement reveals Google’s narrow focus on precision-oriented search. It is plausible that as systems get better and better at understanding the searcher’s intent, they will be more likely to identify useful documents. Sergey’s take on search is that in his childhood he “could not have imagined that today anyone would be able to research any topic in seconds.”

While this may be true for finding specific facts that are part of the fabric of our culture, this model of search breaks down for exploratory search tasks. Not only is the information being sought often distributed across multiple documents and needs to evaluated and synthesized to become useful, but also quite often the searcher cannot articulate the information need in any kind of closed form. This situation is well described in the Library Science literature over the last 30 years, and the advent of the web has not changed human nature. The situation, described, for example by Nicholas Belkin in his ASK (Anomalous States of Knowledge) model is that people will often recognize anomalies in their state of knowledge, but are “unable to specify precisely what is needed to resolve that anomaly” (Belkin et al., 1982). In some cases, for example, people may not even know the terms required to find possibly useful documents.

Google is rightfully proud of its accomplishments, but their claims of innovation (” In the past year alone we have made 359 changes to our web search”) mask the scope of their effort. Google should look more broadly at information seeking so that in their 2010 annual report they can discuss not only the diversity of devices they support, but also the diversity of information seeking tasks that benefit from their technology.

Belkin, N. J.; Oddy, R. & Brooks, H. (1982). ASK for Information Retrieval. Journal of Documentation, 38, 61-71 (part 1) & 145-164 (part 2).

1 Comment

  1. Thanks for taking them to task on this. I was also thrown by that statement, but too underwater to write a rebuttal myself, let alone such a thorough one!

Comments are closed.