Fast, Sample-Efficient Algorithms for Structured Phase Retrieval

Authors: Gauri Jagatap, Chinmay Hegde

NeurIPS 2017 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental In Section 5 we present numerical experiments for our algorithms. We explore the performance of the Co PRAM and Block Co PRAM on synthetic data.
Researcher Affiliation Academia Gauri jagatap Electrical and Computer Engineering Iowa State University Chinmay Hegde Electrical and Computer Engineering Iowa State University
Pseudocode Yes Algorithm 1 Co PRAM: Initialization. Algorithm 2 Co PRAM: Descent. (Also mentions Algorithm 3 and 4 in Appendix A)
Open Source Code No The paper does not provide any explicit statements about releasing source code or links to a code repository.
Open Datasets No We explore the performance of the Co PRAM and Block Co PRAM on synthetic data. The nonzero elements of the unit norm vector x R3000 are generated from N(0, 1).
Dataset Splits No We repeated each of the experiments (fixed n, s, b, m) in Figure 1 (a) and (b), for 50 and Figure 1 (c) for 200 independent Monte Carlo trials.
Hardware Specification Yes All numerical experiments were conducted using MATLAB 2016a on a computer with an Intel Xeon CPU at 3.3GHz and 8GB RAM.
Software Dependencies Yes All numerical experiments were conducted using MATLAB 2016a
Experiment Setup Yes For our simulations, we compared our algorithms Co PRAM and Block Co PRAM with Thresholded Wirtinger flow (Thresholded WF or Th WF) [22] and SPARTA [23]. The parameters for these algorithms were carefully chosen as per the description in their respective papers. We generated phase transition plots by evaluating the probability of empirical successful recovery, i.e. number of trials out of 50. For this experiment we fixed a signal of length n = 3, 000, sparsities s = 20, k = 1 for a block length of b = 20.