The Phase Transition for Recovering a Random
Hypergraph from its Edge Data
Name
yao-ajyao-meng-eecs-2025-thesis.pdf
Description
Thesis PDF
Size
681.77 KB
Format
Adobe PDF
Checksum (MD5)
37449b64a087127f45d1d047ef0456ef
Author(s)
Yao, Andrew
Advisor(s)
Polyanskiy, Yury
Date Issued
May 2025
Publisher
Massachusetts Institute of Technology
Abstract
The weighted projection of a hypergraph is the weighted undirected graph with the same vertex set and edge weight equal to the number of hyperedges that contain the edge; the projection is the unweighted graph with the same vertex set and edge set consisting of edges with weight at least one. For d ≥ 3, after observing the unweighted and weighted projection of a random d-uniform hypergraph that is sampled using a generalization of the Erdős–Rényi random model, we study the recovery of a fraction of the hyperedges and the entire hypergraph. For both cases, we show that there is a sharp phase transition in the feasibility of recovery based on the density of the hypergraph, with recovery possible only when the hypergraph is sufficiently sparse. Particularly, we resolve numerous conjectures from [5]. Furthermore, we display an efficient algorithm that is optimal for both exact and partial recovery. We also analyze the phase transition for exact recovery by exhibiting a regime of probabilities that is below the exact recovery threshold by a polylogarithmic factor for which exact recovery is possible.
MIT Department
Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science
Terms of Use
In Copyright - Educational Use Permitted
Copyright retained by author(s)
Persistent DSpace Link
