In a Comment in the Aug. 4 issue of Nature — coinciding with the 20th anniversary of the public release of the World Wide Web — Oren Etzioni, Professor and Director of the Turing Center at the University of Washington, calls on the computing research community to “think outside the keyword box and improve Internet” search.

Two decades after Internet pioneer Tim Berners Lee introduced his World Wide Web project to the world using the alt.hypertext newsgroup, web search is on the cusp of a profound change — from simple document retrieval to question answering. Instead of poring over long lists of documents that contain requested keywords, users need direct answers to their questions. With sufficient scientific and financial investment, we could soon view today’s keyword searching with the same nostalgia and amusement reserved for bygone technologies such as electric typewriters and vinyl records.

 

But this transformation could be unreasonably delayed. As a community, computer scientists have underinvested in tools that can synthesize sophisticated answers to questions, and have instead focused on incremental progress in lowest-common-denominator search. The classic keyword search box exerts a powerful gravitational pull. Academics and industry researchers need to achieve the intellectual ‘escape velocity’ necessary to revolutionize search. They must invest much more in bold strategies that can achieve natural-language searching and answering, rather than providing the electronic equivalent of the index at the back of a reference book.

Read more after the jump…

Etzioni argues that we need to be more ambitious and imaginative:

Much more research has to be done to improve information-extraction systems… Their abilities need to be extended from being able to infer relations expressed by verbs to those expressed by nouns and adjectives. Information is often qualified by its source, intent and the context of previous sentences. The systems need to be able to detect those, and other, subtleties. Finally, automated methods have to be mapped to a broad set of languages, many of which pose their own idiosyncratic challenges.

 

The main obstacle to the paradigm shift from information retrieval to question answering seems to be a curious lack of ambition and imagination. Much of the research on natural language processing is focused on limited tasks, such as recovering the syntactic structure of sentences rather than trying to uncover their meaning, or on methods that do not scale to massive corpora and arbitrary topics because of their reliance on manually-annotated data, or on algorithms whose computation grows explosively with the amount of text involved.

 

General-purpose question-answering systems will be a boon to scientists searching the literature, and to the increasing number of us who access the web’s richness through a mobile phone with a tiny screen that necessitates concise responses. Without it, we risk drowning in the growing sea of information.

 

[Reprinted by permission from Macmillan Publishers Ltd: Nature 476, 25-26 (04 August 2011), copyright 2011.]

Read the full Comment here — and share your thoughts below.

(Contributed by Erwin Gianchandani, CCC Director)