Is Submodularity Testable?

@article{Comandur2012IsST,
  title={Is Submodularity Testable?},
  author={Seshadhri Comandur and Jan Vondr{\'a}k},
  journal={Algorithmica},
  year={2012},
  volume={69},
  pages={1-25}
}
We initiate the study of property testing of submodularity on the boolean hypercube. Submodular functions come up in a variety of applications in combinatorial optimization. For a vast range of algorithms, the existence of an oracle to a submodular function is assumed. But how does one check if this oracle indeed represents a submodular function?Consider a function f:{0,1}n→ℝ. The distance to submodularity is the minimum fraction of values of f that need to be modified to make f submodular. If… Expand

Paper Mentions

On the Approximation of Submodular Functions
TLDR
A large number of prior works imply that monotone submodular functions can be approximated by coverage functions with a factor between $O(\sqrt{n} \log n)$ and $\Omega(n^{1/3} /\log^2 n) $ and this work proves both upper and lower bounds on such approximations. Expand
Optimal bounds on approximation of submodular and XOS functions by juntas
TLDR
This work investigates the approximability of several classes of real-valued functions by functions of a small number of variables (juntas), and shows that 2Ω(1/e) variables are necessary even for XOS functions, and provides learning algorithms over the uniform distribution. Expand
A pr 2 01 3 On the Approximation of Submodular Functions
Submodular functions are a fundamental object of study in combinatorial optimization, economics, machine learning, etc. and exhibit a rich combinatorial structure. Many subclasses of submodularExpand
Learning submodular functions
TLDR
This paper considers PAC-style learning of submodular functions in a distributional setting and uses lossless expanders to construct a new family of matroids which can take wildly varying rank values on superpolynomially many sets; no such construction was previously known. Expand
Testing Real-Valued Modularity and Submodularity
We study the question of testing whether a function f : {0, 1} → R is modular/submodular or ε-far from it (with respect to Hamming distance). We provide two results: First, it is possible to testExpand
Optimal Bounds on Approximation of Submodular and XOS Functions by Juntas
  • V. Feldman, J. Vondrák
  • Mathematics, Computer Science
  • 2013 IEEE 54th Annual Symposium on Foundations of Computer Science
  • 2013
TLDR
This work investigates the approximability of several classes of real-valued functions by functions of a small number of variables (juntas) and shows that 2<sup>Ω(1/ϵ)</sup> variables are necessary even for XOS functions, and provides learning algorithms over the uniform distribution. Expand
Testing properties of boolean functions
Given oracle access to some boolean function f, how many queries do we need to test whether f is linear? Or monotone? Or whether its output is completely determined by a small number of the inputExpand
Testing Submodularity and Other Properties of Valuation Functions
We show that for any constant $\epsilon > 0$ and $p \ge 1$, it is possible to distinguish functions $f : \{0,1\}^n \to [0,1]$ that are submodular from those that are $\epsilon$-far from everyExpand
Testing Coverage Functions
TLDR
An algorithm is demonstrated which makes O(m|U|) queries to an oracle of a coverage function and completely reconstructs it, giving a polytime tester for succinct coverage functions for which |U| is polynomially bounded in m. Expand
Learning pseudo-Boolean k-DNF and submodular functions
TLDR
It is proved that any submodular function f: 0,1}^n -> {0,1,...,k} can be represented as a pseudo-Boolean 2k-DNF formula, and it is shown that an analog of Hastad's switching lemma holds if all constants associated with the terms of the formula are bounded. Expand
...
1
2
3
4
5
...

References

SHOWING 1-10 OF 62 REFERENCES
Approximating submodular functions everywhere
TLDR
The problem of approximating a non-negative, monotone, submodular function f on a ground set of size n everywhere is considered, after only poly(n) oracle queries, and it is shown that no algorithm can achieve a factor better than Ω(√n/log n), even for rank functions of a matroid. Expand
An analysis of approximations for maximizing submodular set functions—I
TLDR
It is shown that a “greedy” heuristic always produces a solution whose value is at least 1 −[(K − 1/K]K times the optimal value, which can be achieved for eachK and has a limiting value of (e − 1)/e, where e is the base of the natural logarithm. Expand
Optimal bounds for monotonicity and Lipschitz testing over the hypercube
TLDR
A very general theorem is proved showing that edge testers work for a class of “bounded-derivative” properties, which contains both monotonicity and Lipschitz, and it is proved that the edge tester only requires O(n/e) samples (regardless of R). Expand
Monotonicity testing over general poset domains
TLDR
It is shown that in its most general setting, testing that Boolean functions are close to monotone is equivalent, with respect to the number of required queries, to several other testing problems in logic and graph theory. Expand
Learning submodular functions
TLDR
This paper considers PAC-style learning of submodular functions in a distributional setting and uses lossless expanders to construct a new family of matroids which can take wildly varying rank values on superpolynomially many sets; no such construction was previously known. Expand
On Testing Convexity and Submodularity
Submodular and convex functions play an important role in many applications, and in particular in combinatorial optimization. Here we study two special cases: convexity in one dimension andExpand
Testing Coverage Functions
TLDR
An algorithm is demonstrated which makes O(m|U|) queries to an oracle of a coverage function and completely reconstructs it, giving a polytime tester for succinct coverage functions for which |U| is polynomially bounded in m. Expand
On the strength of comparisons in property testing
  • E. Fischer
  • Computer Science, Mathematics
  • Inf. Comput.
  • 2001
TLDR
It is proved that an e-test for a property of integer sequences, such as the property of the sequence being a monotone non-decreasing sequence, cannot perform less queries than the best e- test which uses only comparisons between the queried values. Expand
A combinatorial, strongly polynomial-time algorithm for minimizing submodular functions
This paper presents a combinatorial polynomial-time algorithm for minimizing submodular functions, answering an open question posed in 1981 by Grotschel, Lov asz, and Schrijver. The algorithm employsExpand
A combinatorial strongly polynomial algorithm for minimizing submodular functions
This paper presents a combinatorial polynomial-time algorithm for minimizing submodular functions, answering an open question posed in 1981 by Grötschel, Lovász, and Schrijver. The algorithm employsExpand
...
1
2
3
4
5
...