Optimal Inapproximability Results for MAX-CUT and Other 2-Variable CSPs?
read more
Citations
Most Tensor Problems Are NP-Hard
Analysis of Boolean Functions
The Design of Approximation Algorithms
Most tensor problems are NP-hard
Related Papers (5)
Frequently Asked Questions (11)
Q2. What is the majority is stablest theorem?
The Majority Is Stablest theorem essentially states that the discrete cube is a good approximation of the sphere in a certain sense.
Q3. What is the main theorem of the MAX-CUT problem?
Silvestri, and Trevisan [11] have shown that the weighted and unweighted versions of MAX-CUT have equal optimal approximation factors (up to an additive o(1)) and so the authors pass freely between the two problems in this paper.
Q4. What is the obvious generalization of the Plurality function to the q-ary?
The most obvious generalization of the Majority function to the q-ary domain is the Plurality function, which on input x ∈ [q]n outputs the most common value for xi (tie-breaking is unimportant).
Q5. What is the definition of the Unique Label Cover problem?
The Unique Label Cover problem, L(V,W,E, [M ], {σv,w}(v,w)∈E) is defined as follows: Given is a bipartite graph with left side vertices V , right side vertices W , and a set of edges E.
Q6. How do the authors prove inapproximability results for boolean CSPs?
inapproximability results for boolean CSPs are obtained by encoding assignments to Label Cover variables via a binary code and then running PCP tests on the (supposed) encodings.
Q7. What is the function that is able to assume that all the supposed q-ary Long?
Their PCP verifier for Γ-MAX-2LIN(q) will be able to assume that all the supposed q-ary Long Codes fw with which it works are folded.
Q8. What is the intuition that the xi should be set to TRUE?
The authors contend that the balanced versions of 2-bit CSPs ought to be equally hard as their general versions; the intuition is that if more constraints are expected to be satisfied if xi is set to, say, 1 rather than −1, it is a “free hint” that the xi should be set to TRUE.
Q9. What is the probability that u and v lie on opposite sides of the random cut?
For an edge (u, v) in G, the probability that u and v lie on opposite sides of the random cut is proportional to the angle between xu and xv.
Q10. What is the sum of the vectors associated with the endpoints of G?
The embedding is selected such that the sum∑(u,v)∈E1 2 − 1 2 〈xu,xv〉, (1)involving the inner products of vectors associated with the endpoints of edges of G, is maximized.
Q11. What is the main result of the Majority Is Stablest theorem?
In Section 6.3 the authors mention interesting generalizations of the Majority Is Stablest theorem for q-ary functions, q > 2, which are relevant for hardness of approximation and are not resolved in full.