Multi-source transfer learning incurs an intrinsic adaptation cost that can exceed one, with phase transitions separating regimes where bias-agnostic estimators match oracle performance from those where they cannot.
Community detection and stochastic block models: re cent developments
5 Pith papers cite this work. Polarity classification is still indexing.
5
Pith papers citing it
citation-role summary
method 1
citation-polarity summary
verdicts
UNVERDICTED 5roles
method 1polarities
use method 1representative citing papers
First efficient sum-of-squares algorithms recover exact and approximate overlapping planted cliques in dense random intersection graphs for k ≫ √(n log n), with robustness to noise, monotone adversaries, and optimal edge corruptions.
Sharp conditions for exact recovery of general planted subgraphs in ER graphs are given by the minimal maximum subgraph density, with matching bounds, a spectral algorithm, and computational hardness results via low-degree polynomials.
Transfer learning from informative source networks improves target DCMM estimation accuracy by enlarging the eigenvalue gap of the connection probability matrix, with algorithms to avoid negative transfer.
Tutorial on the standard linear model with an outline of the authors' proof that replica-symmetric formulas for its phase transitions in mutual information and MMSE are exact.
citing papers explorer
-
The Statistical Cost of Adaptation in Multi-Source Transfer Learning
Multi-source transfer learning incurs an intrinsic adaptation cost that can exceed one, with phase transitions separating regimes where bias-agnostic estimators match oracle performance from those where they cannot.
-
Robust Algorithms for Finding Cliques in Random Intersection Graphs via Sum-of-Squares
First efficient sum-of-squares algorithms recover exact and approximate overlapping planted cliques in dense random intersection graphs for k ≫ √(n log n), with robustness to noise, monotone adversaries, and optimal edge corruptions.
-
Recovery of Planted Subgraphs
Sharp conditions for exact recovery of general planted subgraphs in ER graphs are given by the minimal maximum subgraph density, with matching bounds, a spectral algorithm, and computational hardness results via low-degree polynomials.
-
Transfer Learning for Degree-Corrected Mixed Membership Network Models
Transfer learning from informative source networks improves target DCMM estimation accuracy by enlarging the eigenvalue gap of the connection probability matrix, with algorithms to avoid negative transfer.
-
Understanding Phase Transitions via Mutual Information and MMSE
Tutorial on the standard linear model with an outline of the authors' proof that replica-symmetric formulas for its phase transitions in mutual information and MMSE are exact.