Colleague is an interactive, web-based, expert-collaborative machine learning system designed to mimic interactions between experts within a field. Currently this system is in the planning and early development stages, but the goal is to develop a system wherein researchers, who may or may not be knowledgeable about data mining, can submit a dataset, along with their own ideas and intuitions, and Colleague will use this knowledge to devise a number of new theories about the data. Once Colleague has developed a set of theories, the researcher can use his/her own knowledge about the data to provide the system with new insights to help it create more accurate theories. Comprehensibility of the theories is of paramount concern since the researcher should be able to interact with Colleague as if it is a "colleague", and not as a computer program.

The development of Colleague is supported in part by a training grant from the National Library of Medicine (Grant No. 5T15LM007359).

---

If you already have an account with Colleague, then please log in.

Otherwise, sign up for a colleague account.

If you already have an account, but have forgotten your password, click here.

If you want to explore the Colleague site without signing up for an account, please log in with the username guest and the password guest.