The hallmark of the Cooley-Tukey algorithm for Fast Fourier Transform is the butterfly network, which helps reduce O(N^2) computations to O(NlogN). Butterflies are very special graphs entangled in routing [Arora], switching [Chung], shuffling [Yang], and mixing [Czumaj].
Given a weighted connected component, what is the weight of of the maximum independent set, and how many different sets have this weight? A Petersen graph GP(5, 2) with zero vertex weights has three maximal subsets by size, 0 maximum set weight but 76 distinct independent sets with this weight. Continue reading
Gus wants to open franchises of his restaurant, Los Pollos Hermanos, along Central Avenue. There are n possible locations for franchises, where location i is at mile i on Central. Each location i > 1, is thus a distance of 1 mile from the previous one. There are two rules.
- At each location, there can be at most one restaurant, and the profit of a restaurant at location i is p_i.
- Any two restaurants must be at least 2 miles apart.
Min-max heaps were introduced in [ASSS86] as an efficient way to support heap operations for both minimum and maximum values. Structurally, the min-max heap levels alternate between min-heap condition and max-heap, and hence evaluates grandchildren/grandparents during insertion or search. Min-max heaps can also be generalized to find the k-th smallest element in O(1) time.
The aim of this post is to highlight the utility of non-negative factorization (NMF) in data analysis through examples. In a different post, we’ll talk theory and implementation by thinking of NMF as constrained optimization.
Learning parts of a whole
When Lee and Seung [Lee99] re-introduced non-negative matrix factorization (NMF), they emphasized on the algorithm’s ability to learn parts of a whole. Continue reading
Fast Iterative Shrinkage-Thresholding Algorithm (FISTA) [Beck09] was one of the first algorithms to use Nesterov’s accelerated gradient descent to speed up the convergence of iterative shrinkage-thresholding (ISTA) from O(β/ε) to O(√(β/ε)). The algorithm and proof sketch for in the previous post based were based on [Bubeck14] and [Beck09] so we’ll skip them. Instead give an R implementation and results that compares ISTA and FISTA performance (in terms of number of iterations) for various step sizes (iterations increase and step size decreases). Continue reading
General first-order methods use only the values of objective/constraints and their subgradients to optimize non-smooth (including constrained) function . Such methods that are oblivious to the structure of the problem are termed black-box optimization techniques. Continue reading
A problem I was asked: suppose a file that stored many valid IPv4 addresses was corrupted such that all dot in the addresses were removed. Given such a corrupt IP, infer all possible corresponding IP addresses. For example, given 12345, all possible addresses are: 126.96.36.199, 188.8.131.52, 184.108.40.206 and 220.127.116.11. The hint that recursion should be used made the problem easier: