PITTSBURGH—Carnegie Mellon University's School of Computer Science (SCS) is extending its leadership in the burgeoning field of machine learning by creating the nation's first Machine Learning Department.
The new department designation for what was formerly known as the Center for Automated Learning and Discovery (CALD) reflects the importance of machine learning in such growing areas as data mining and sensor networks, as well as a commitment by the university to continue its pioneering efforts in the field.
Tom M. Mitchell, the Fredkin Professor of Artificial Intelligence and Learning, heads the department, one of six within SCS. Mitchell founded CALD in 1997 together with Stephen E. Fienberg, the Maurice Falk University Professor of Statistics and Social Science. The new department is the first to offer a Ph.D. in the field of machine learning. There are currently 21 students in the program, which was established in 2003.
"Machine learning has emerged as a distinct scientific discipline and one that is uniquely suited to Carnegie Mellon's strengths," said Carnegie Mellon Provost and Senior Vice President Mark S. Kamlet. "Drawing on both computer science and statistics, machine learning is another example of how the university's tradition of multi-disciplinary research produces important scientific insights."
The roots of machine learning extend back almost 50 years, when a few researchers began to explore whether computers could learn to play games. It became established as a field as they addressed the broader question of whether it was possible to develop software that could learn from experience in order to improveits performance.
"Even 20 years ago, machine learning had few commercial applications," said Mitchell. "But now it is the method of choice for an important niche of software applications."
In speech recognition systems, for instance, machine learning has proven to be vital not only for initially training the system to understand the spoken word, but also for customizing each system to respond to the speech patterns of individual users. Other sensor systems, such as computer vision, also make use of machine learning techniques. "The niche where machine learning will be used is growing rapidly as applications grow in complexity and as we develop more accurate learning algorithms," Mitchell said.
Machine learning is the key to data mining. It enables sifting through vast databases of information that would overwhelm a human analyst. Data mining techniques are used by grocery chains to analyze the purchasing habits of their customers and by astrophysicists to recognize new cosmic phenomena within sky survey data. Mitchell, working with D.O. Hebb Professor of Psychology Marcel Just, is using machine learning techniques to analyze functional brain imaging scans to better understand how the brain processes words and sentences. Automated methods for analyzing the activity of thousands of genes associated with various diseases are also being developed with the help of machine learning.
Carnegie Mellon is recognized as a leader in machine learning, with seven faculty members on the editorial boards of the field's top journals and with significant machine learning research occurring throughout the School of Computer Science.
"The transition from the Center for Automated Learning and Discovery to the Machine Learning Department recognizes the emergence of machine learning as a rigorous academic discipline, as well as the excellence of the work by Tom Mitchell and the rest of our faculty," said Randal E. Bryant, dean of the School of Computer Science. "Machine learning has grown and flourished at Carnegie Mellon, drawing ideas from and contributing to many parts of the university. It has fostered especially strong ties between the School of Computer Science and the Statistics Department, providing computer scientists with more rigorous mathematical tools, and statisticians with new challenges and opportunities."
Carnegie Mellon will host the International Conference on Machine Learning June 25-29.
Byron Spice | 412-268-9068 | email@example.com