Jeff A. Bilmes's Publications

• Sorted by Date • Classified by Publication Type • Classified by Research Category • Sorted by First Author Last Name • Classified by Author Last Name •

Efficient Interactive Maximization of BP and Weakly Submodular Objectives

Adhyyan Narang, Omid Sadeghi, Lillian J. Ratliff, Maryam Fazel, and Jeff Bilmes. Efficient Interactive Maximization of BP and Weakly Submodular Objectives. In Uncertainty in Artificial Intelligence (UAI), AUAI, Barcelona, Spain, July 2024.

Download

(unavailable)

Abstract

In the context of online interactive machine learning with combinatorial objectives, we extend purely submodular prior work to more general non-submodular objectives. This includes: (1) those that are additively decomposable into a sum of two terms (a monotone submodular and monotone supermodular term, known as a BP decomposition); and (2) those that are only weakly submodular. In both cases, this allows representing not only competitive (submodular) but also complementary (supermodular) relationships between objects, enhancing this setting to a broader range of applications (e.g., movie recommendations, medical treatments, etc.) where this is beneficial. In the two-term case, moreover, we study not only the more typical monolithic feedback approach but also a novel framework where feedback is available separately for each term. With real-world practicality and scalability in mind, we integrate \Nystrom sketching techniques to significantly improve the computational complexity, including for the purely submodular case. In the Gaussian process contextual bandits setting, we show sub-linear theoretical regret bounds in all cases. We also empirically show good applicability to recommendation systems and data subset selection.

@InProceedings{narang2024uai-online-sub-sup,
   Author = {Adhyyan Narang and Omid Sadeghi and Lillian J. Ratliff and Maryam Fazel and Jeff Bilmes},
   Title = {Efficient Interactive Maximization of BP and Weakly Submodular Objectives},
   booktitle =  {Uncertainty in Artificial Intelligence (UAI)},
   year = 	 {2024},
   address = 	 {Barcelona, Spain},
   month = 	 {July},
   publisher = {AUAI},
   abstract = {In the context of online interactive machine learning with combinatorial objectives, we extend purely submodular prior work to more general non-submodular objectives. This includes: (1) those that are additively decomposable into a sum of two terms (a monotone submodular and monotone supermodular term, known as a BP decomposition); and (2) those that are only weakly submodular. In both cases, this allows representing not only competitive (submodular) but also complementary (supermodular) relationships between objects, enhancing this setting to a broader range of applications (e.g., movie recommendations, medical treatments, etc.) where this is beneficial. In the two-term case, moreover, we study not only the more typical monolithic feedback approach but also a novel framework where feedback is available separately for each term. With real-world practicality and scalability in mind, we integrate \Nystrom{} sketching techniques to significantly improve the computational complexity, including for the purely submodular case. In the Gaussian process contextual bandits setting, we show sub-linear theoretical regret bounds in all cases. We also empirically show good applicability to recommendation systems and data subset selection.},
}

Share

Generated by bib2html.pl (written by Patrick Riley ) on Mon Oct 14, 2024 00:38:45