From the 1 of 15 linked papers with an AI index.
15 papers
The Hypergraph Moore Bound
Afonso S. Bandeira, Dmitriy Kunisky, Petar NiziÄ-Nikolac +2
The paper proves Feige’s hypergraph Moore bound for all even uniformities (k ≥ 4) without extra polylogarithmic factors, using colored walks in a Kikuchi graph and a polynomial int…
pscaling small models: Principled warm starts and hyperparameter transfer
Yuxin Ma, Nan Chen, Mateo DÃaz +3
Modern large-scale neural networks are often trained and released in multiple sizes to accommodate diverse inference budgets. To improve efficiency, recent work has explored model…
Lehner's operator norm formulas, semidefinite programming, and spiked matrix models
Dmitriy Kunisky
Lehner (1999) derived elegant formulas for the operator norm of operators of the form $\mathfrak{X} = \mathbf{A}_0 \otimes \mathfrak{1} + \sum_{i = 1}^n \mathbf{…
A revision of Litvak's conjecture on Gaussian minima and a volumetric zone conjecture
Dmitriy Kunisky
Litvak (2018) conjectured that, for any , the quantity where is a centered Gaussian random vector is minimiz…
Universality of first-order methods on random and deterministic matrices
Nicola Gorini, Chris Jones, Dmitriy Kunisky +1
General first-order methods (GFOM) are a flexible class of iterative algorithms which update a state vector by matrix-vector multiplications and entrywise nonlinearities. A long li…
Empirical universality and non-universality of local dynamics in the Sherrington-Kirkpatrick model
Grace Liu, Dmitriy Kunisky
Several recent works have aimed to design algorithms for optimizing the Hamiltonians of spin glass models from statistical physics. While Montanari (2018) eventually gave a sophist…