RealTime IT News

Intel's Computers Learn From Your Mistakes

How many times have you wished your computer knew what you were thinking?

Working at the chip level, Intel is the latest company to produce software that taps into the "predictive" computing realm. The chip making giant Monday said its new tools for its processors could help developers build computers that "learn" from their experiences.

Through its Open Source Machine Learning Library (OpenML), Intel released open-source graphical model algorithms and software. The toolbox includes C++ source code for all of the library's functions and a royalty-free redistribution license.

The software is based on "Bayesian" mathematical principles, which hold that the probability of future events can be calculated by studying their prior frequency. Because Bayesian models are based on data collected from experience, the more data obtained the better the predictions, and if the data changes, the results correct themselves.

In one example, Intel said researchers used source code to help create an audio/video speech recognition system that allows computers to detect a speaker's face with a video camera and track his mouth movements. In essence, the computer was learning to "read lips".

"Intel wants computers to be more proactive," Intel vice president and director of research David Tennenhouse 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 key to educating computers, Intel said, is faster microprocessors and improvements in the graphical mathematical models to allow real-time computer "machine learning" algorithms to run on standard PCs.

"OpenML is certain to drive an explosion of machine learning-based applications such as toys that respond to a child's movements and networks of wireless sensors that will enhance our safety, productivity and stewardship of our environment," Tennenhouse said.

The Santa Clara, Calif.-based chipmaker is one of several companies applying Bayesian principles to their software. Companies like Roaring Penguin, Barracuda Networks, and Gordano use the algorithms for anti-spam software and firewall protection products. Google is also using Bayesian spam filters in its search engine.

Intel said its software could also be used for creating a model of a person's behavior to decide how best to deal with spam. The chipmaker said it also envisions the technology as aiding in the operation of huge databases of gene studies to spot promising proteins for new drugs.

In related news, Intel also released new compilers this week for building applications that run on servers, desktops and mobile systems, such as laptops, mobile phones and personal digital assistants (PDAs) based on Intel processors.

Version 8.0 adds in Intel C++ and Fortran Compilers for Windows and Linux as well as Intel C++ Compilers for Windows CE .NET. The compilers are designed for systems running Itanium 2, Xeon, Pentium 4, XScale for mobile phones and PDAs and the Intel Pentium M (the main component of Intel's Centrino technology).

The compilers also support the forthcoming Intel 90-nanometer processor code-named "Prescott," as well as Intel's Hyper-Threading technology and OpenMP 2.0, an industry specification for using high-level directives to thread applications.

New tools offered in the compilers include the Intel Code Coverage Tool (which presents a complete picture of how much application logic is exercised), and Intel Test Prioritization Tool (which ensures the right code is exercised).

Several models of the new Intel compilers are available now with prices ranging from $399 to $1,499.