Ithaca College

Professor Doug Turnbull (Computer Science) was recently awarded an NSF Information & Intelligent Systems grant for $100,000. The grant focuses on using user preference information that is collected from his MegsRadio.fm personalized radio app to improve media recommendation algorithms.   

This grant will primarily be used to support undergraduate research at Ithaca College over the next three years.  Details can be found at:

https://www.nsf.gov/awardsearch/showAward?AWD_ID=1615679

This is a collaborative grant with Professor Thorsten Joachims who was awarded $400,000 to support graduate research at Cornell University.

DETAILS:

About MegsRadio.fm:

MegsRadio.fm is a locally-focused personalized Internet radio system. You can download it for free from the Apple App Store or check it out online:

http://megradio.fm
http://megsradio.fm/about.html
https://itunes.apple.com/us/app/megsradio/id1021359050?mt=8
https://www.facebook.com/megsradiofm

https://www.ithaca.edu/intercom/article.php?story=20160725212814103#.WADT7JMrJE4

MegsRadio.fm is currently being developed by a large team of undergraduate students from Ithaca College and Cornell. 

NSF Grant Abstract:

Log data is one of the most ubiquitous forms of data available, as it can be recorded from a variety of systems (e.g., search engines, recommender systems, ad placement platforms) at little cost. Making huge amounts of log data accessible to learning algorithms provides the potential to acquire knowledge at unprecedented scale. Furthermore, the ability to learn from log data can enable effective machine learning even in systems where manual labeling of training data is not economically viable. Log data, however, provides only partial information -- "contextual-bandit feedback" -- limited to the particular actions taken by the system. The feedback for all the other actions the system could have taken is typically not known. This makes learning from log data fundamentally different from traditional supervised learning, where "correct" predictions together with a loss function provide full-information feedback.

This project tackles the problem of Batch Learning from Bandit Feedback (BLBF) by developing principled learning methods and algorithms that can be trained with logs containing contextual-bandit feedback. First, the project develops the learning theory of BLBF, especially with respect to understanding the use and design of counterfactual risk estimators for BLBF. Second, the project derives new learning methods for BLBF. Past work has already demonstrated that Conditional Random Fields can be trained in the BLBF setting, and the project derives BLBF analogs of other learning methods as well. Third, the project derives scalable training algorithms for these BLBF methods to enable large-scale applications. And, finally, the project validates the methods with real-world data from operational systems.