ملخص
It is known that greedy methods perform well for maximizing monotone submodular functions. At the same time, such methods perform poorly in the face of non-monotonicity. In this paper, we show—arguably, surprisingly—that invoking the classical greedy algorithm O(√k)-times leads to the (currently) fastest deterministic algorithm, called REPEATEDGREEDY, for maximizing a general submodular function subject to k-independent system constraints. REPEATEDGREEDY achieves (1 + O(1/√k))k approximation using O(nr√k) function evaluations (here, n and r denote the size of the ground set and the maximum size of a feasible solution, respectively). We then show that by a careful sampling procedure, we can run the greedy algorithm only once and obtain the (currently) fastest randomized algorithm, called SAMPLEGREEDY, for maximizing a submodular function subject to k-extendible system constraints (a subclass of k-independent system constrains). SAMPLEGREEDY achieves (k + 3)-approximation with only O(nr/k) function evaluations. Finally, we derive an almost matching lower bound, and show that no polynomial time algorithm can have an approximation ratio smaller than k +1/2 - ε. To further support our theoretical results, we compare the performance of REPEATEDGREEDY and SAMPLEGREEDY with prior art in a concrete application (movie recommendation). We consistently observe that while SAMPLEGREEDY achieves practically the same utility as the best baseline, it performs at least two orders of magnitude faster.
| اللغة الأصلية | الإنجليزيّة |
|---|---|
| الصفحات (من إلى) | 758-784 |
| عدد الصفحات | 27 |
| دورية | Proceedings of Machine Learning Research |
| مستوى الصوت | 65 |
| حالة النشر | نُشِر - 2017 |
| الحدث | 30th Conference on Learning Theory, COLT 2017 - Amsterdam, هولندا المدة: ٧ يوليو ٢٠١٧ → ١٠ يوليو ٢٠١٧ |
ملاحظة ببليوغرافية
Publisher Copyright:© 2017 M. Feldman, C. Harshaw & A. Karbasi.
بصمة
أدرس بدقة موضوعات البحث “Greed Is Good: Near-Optimal Submodular Maximization via Greedy Optimization'. فهما يشكلان معًا بصمة فريدة.قم بذكر هذا
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver