Importing Machine Learning Techniques to the Database

SHARE:

Importing Machine Learning Techniques to the Database

Faculty Advisor: Una-May O'Reilly
Mentor(s): Kalyan Veeramachaneni, Erik Hemberg
Contact e-mail: alfa-apply@csail.mit.edu
Research Area(s): Artificial Intelligence, Computer Systems
Large and continuously growing data repositories require machine learning methods that are able to quickly mine and update their models. This has led to the integration of machine learning algorithms with database languages. We are developing a methodology to integrate a non parametric, non-linear Machine Learning technique called Genetic Programming into the data base. This allows the algorithm to take advantage of fast data management routines available within the database, reconfigure the training data quickly, assimilate new updates to the data into the model development. These features reduce the model design time by orders of magnitude. You will gain experience developing Machine learning algorithms, very large databases and will work on a healthcare application.

See more projects

Return to the SuperUROP site.