Intel researchers think computers could use some on-the-job training.
The chipmaker"s research arm on Monday said it is offering the Open Source Machine Learning Library, a collection of software that can help computers learn from various experiences. It will offer the library to interested parties for free via the Web. Intel researchers created the library using Bayesian mathematical principles, which can help computers learn by predicting the likelihood of a person"s behavior or another occurrence by calculating how often that activity has happened in the past.
The library is intended to serve as a toolkit for others who are designing computers or robots. Although it may conjure up images from science fiction, its underlying mission is to focus on new ways of using Intel processors and, ultimately, to sell more powerful chips in larger numbers. Intel said the machine learning library can be used to boost the effectiveness of computers for consumers, businesses or industrial applications, by helping those tools to learn in those settings. "Intel wants computers to be more proactive," David Tennenhouse, Intel"s director of research, said in a statement. "To do this, they need to be able to learn from their experiences with users and the world around them. Using new statistical methods to identify key patterns, these systems will start anticipating the needs of their users and pre-computing responses to the most likely questions so that the answers will be instantly available the moment they are needed."
The library could be used in designing more interactive e-mail systems, creating a model of a person"s behavior and managing messages based on that model, for instance. The library also could be used in toys that react to children"s behaviors, computers that monitor a person"s health, or machines that help farmers monitor crops, an Intel representative said.