Generalized Approximate Survey Propagation for High-Dimensional Estimation
Authors: Carlo Lucibello, Luca Saglietti, Yue Lu
ICML 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | In Fig. 1 (Top and Middle), we show the probability of a perfect recovery and convergence times of GASP for the realvalued phase retrieval problem, for different sampling ratios α and values of the symmetry-breaking parameter m, with λ = 0. |
| Researcher Affiliation | Collaboration | 1Microsoft Research New England, Cambridge, MA 02142, USA 2Italian Institute for Genomic Medicine, Turin, Italy 3John A. Paulson School of Engineering and Applied Sciences, Harvard University, Cambridge, MA 02138, USA 4Bocconi Institute for Data Science and Analytics, Bocconi University, Milan, Italy. |
| Pseudocode | Yes | Algorithm 1 GASP(m) for MAP |
| Open Source Code | No | The paper does not contain any statement about releasing source code, providing a repository link, or making code available in supplementary materials. |
| Open Datasets | No | The paper specifies how data is generated synthetically (e.g., 'x0 N(0, IN)', 'F µ i N(0, 1/N)') but does not refer to a publicly available or open dataset with access information (link, DOI, formal citation). |
| Dataset Splits | No | The paper conducts experiments on synthetically generated data and mentions sample sizes (e.g., 'N = 10^3, averaged over 100 samples'), but it does not specify explicit training, validation, or test dataset splits (e.g., percentages, counts, or references to standard splits). |
| Hardware Specification | No | The paper does not provide specific details about the hardware (e.g., GPU/CPU models, memory) used for running the experiments. |
| Software Dependencies | No | The paper does not provide specific software dependency details, such as library names with version numbers, needed to replicate the experiments. |
| Experiment Setup | Yes | The initial condition is given by V t=0 0 = V t=0 1 = 1 and ˆxt=0 N(0, IN). ... Therefore we initialize SE with ρt=0 = 0.1, and qt=0 = 1 + (ρt=0)2. ... Among the many possible continuation schedules for λ... in this paper we choose a simple two-stage approach: first we run GASP till convergence with a given value of λ > 0, then we set λ = 0 in the successive iterations. |