An Extended Kernel for Generalized Multiple-Instance Learning

The multiple-instance learning (MIL) model has been successful in areas such as drug discovery and content-based image-retrieval. Recently, this model was generalized and a corresponding kernel was introduced to learn generalized MIL concepts with a support vector machine. While this kernel enjoyed empirical success, it has limitations in its representation. We extend this kernel by enriching its representation and empirically evaluate our new kernel on data from content-based image retrieval, biological sequence analysis, and drug discovery. We found that our new kernel generalized noticeably better than the old one in content-based image retrieval and biological sequence analysis and was slightly better or even with the old kernel in the other applications, showing that an SVM using this kernel does not overfit despite its richer representation.

School of Natural Resource Sciences

Good afternoon. Thank you for asking me to be with you today; I welcome this opportunity to talk with you, to hear what you're thinking, and to participate with you in an exchange of ideas. I've been asked to make some prepared remarks this afternoon, and when those are ended I welcome your ideas, your questions, and your comments. I look forward to more discussions with you at other times, as well. I do have open hours Friday afternoons, and encourage faculty and staff to stop by to visit, to share an idea, to ask a question, or to just visit. Sometimes duties call me away, as they will for the next few Fridays, but I try to be available as many Friday afternoons as possible for drop-in visitors. I hope you'll be among them. As a big believer in the land grant university, with our mission of teaching, research, and extension education, I also am a big believer that it takes all of us, working together, to fulfill that mission.