Turing Award honors learning theory, parallel computing work

March 11, 2011 | Source: Ars Technica

This year’s Turing Award was given to Leslie G. Valiant of Harvard University.

Each year, the Association for Computing Machinery honors a computer scientist for his or her contributions to the field. The prize comes with $250,000 and is named in honor of the British mathematician Alan Turing.

Valiant is cited for his contributions to a number of fields, including machine learning, computational complexity, and parallel and distributed computing.

Valiant’s approach to machine learning, called the probably approximately correct (PAC) model, provides ways of determining whether a machine learning algorithm provides sufficient information for it to make any accurate predictions at all.

Valiant also showed that, even when the well-known P vs. NP decision problem is easy, the associated counting problem can still be hard.

And he demonstrated that there is a solution to parallel processing that, mathematically, has to work at avoiding congestion, preventing bottlenecks on a few key routes.