Scalable Greedy Feature Selection via Weak Submodularity

Greedy algorithms are widely used for problems in machine learning such as feature selection and set function optimization. Unfortunately, for large datasets, the running time of even greedy algorithms can be quite high. This is because for each greedy step we need to refit a model or calculate a function using the previously selected choices and the new candidate. Two algorithms that are faster approximations to the greedy forward selection were introduced recently ([Mirzasoleiman et al. 2013, 2015]). They achieve better performance by exploiting distributed computation and stochastic evaluation respectively. Both algorithms have provable performance guarantees for submodular functions. In this paper we show that divergent from previously held opinion, submodularity is not required to obtain approximation guarantees for these two algorithms. Specifically, we show that a generalized concept of weak submodularity suffices to give multiplicative approximation guarantees. Our result extends the applicability of these algorithms to a larger class of functions. Furthermore, we show that a bounded submodularity ratio can be used to provide data dependent bounds that can sometimes be tighter also for submodular functions. We empirically validate our work by showing superior performance of fast greedy approximations versus several established baselines on artificial and real datasets.

[1]  R. Tibshirani,et al.  Least angle regression , 2004, math/0406456.

[2]  Oluwasanmi Koyejo,et al.  On Prior Distributions and Approximate Inference for Structured Variables , 2014, NIPS.

[3]  Morteza Zadimoghaddam,et al.  Randomized Composable Core-sets for Distributed Submodular Maximization , 2015, STOC.

[4]  Deanna Needell,et al.  CoSaMP: Iterative signal recovery from incomplete and inaccurate samples , 2008, ArXiv.

[5]  Abhimanyu Das,et al.  Submodular meets Spectral: Greedy Algorithms for Subset Selection, Sparse Approximation and Dictionary Selection , 2011, ICML.

[6]  Yaron Singer,et al.  Maximization of Approximately Submodular Functions , 2016, NIPS.

[7]  Andreas Krause,et al.  Lazier Than Lazy Greedy , 2014, AAAI.

[8]  Sergei Vassilvitskii,et al.  Fast greedy algorithms in mapreduce and streaming , 2013, SPAA.

[9]  Aditya Bhaskara,et al.  Greedy Column Subset Selection: New Bounds and Distributed Algorithms , 2016, ICML.

[10]  M. L. Fisher,et al.  An analysis of approximations for maximizing submodular set functions—I , 1978, Math. Program..

[11]  Oluwasanmi Koyejo,et al.  Sparse Submodular Probabilistic PCA , 2015, AISTATS.

[12]  Huy L. Nguyen,et al.  The Power of Randomization: Distributed Submodular Maximization on Massive Datasets , 2015, ICML.

[13]  Aaron Roth,et al.  Constrained Non-monotone Submodular Maximization: Offline and Secretary Algorithms , 2010, WINE.

[14]  Oluwasanmi Koyejo,et al.  Examples are not enough, learn to criticize! Criticism for Interpretability , 2016, NIPS.

[15]  Andreas Krause,et al.  Submodular Dictionary Selection for Sparse Representation , 2010, ICML.

[16]  Joel A. Tropp,et al.  Greed is good: algorithmic results for sparse approximation , 2004, IEEE Transactions on Information Theory.

[17]  Andreas Krause,et al.  Distributed Submodular Maximization: Identifying Representative Elements in Massive Data , 2013, NIPS.

[18]  Alexandros G. Dimakis,et al.  Restricted Strong Convexity Implies Weak Submodularity , 2016, The Annals of Statistics.

[19]  Mohamed S. Kamel,et al.  Greedy column subset selection for large-scale data sets , 2014, Knowledge and Information Systems.