Topcon (Netherlands)

output

20202026

most citedWhen Do Neural Networks Outperform Kernel Methods?

73 citations

researchers with a paper here

A. Endo1 · h 2
A. J.1
A. Moerman1 · h 3
A. Montanari1 · h 81
A. P. Laguna1 · h 5
B. Ghorbani1 · h 17
D. Thoen1 · h 19
J. Baselmans1 · h 42
Jochem J A Baselmans1
K. Karatsu1 · h 4
Leon G. G. Olde Scholtenhuis1
L. Marting1 · h 3

collaborating institutions

Google (United States)US1 paper
Stanford UniversityUS1 paper
Tokyo Medical University Ibaraki Medical CenterJP1 paper
Topcon Corporation (Japan)JP1 paper
Topcon (Japan)JP1 paper
University of California, BerkeleyUS1 paper
University of TsukubaJP1 paper

Showing 2020Show all

2 papers · 1 filter

math-ph2020

Fredholm property and essential spectrum of $3 - D$ Dirac operators with regular and singular potentials

Vladimir Rabinovich

We consider the $3 - D$ Dirac operator $D_{A, Φ, Q_{s i n}}$ with variable regular magnetic and electrostatic potentials $A$ , $Φ$ and with singular…

stat.ML2020★ 73 cited

When Do Neural Networks Outperform Kernel Methods?

Behrooz Ghorbani, Song Mei, Theodor Misiakiewicz +1

For a certain scaling of the initialization of stochastic gradient descent (SGD), wide neural networks (NN) have been shown to be well approximated by reproducing kernel Hilbert sp…

Fredholm property and essential spectrum of 3−D Dirac operators with regular and singular potentials

When Do Neural Networks Outperform Kernel Methods?

Fredholm property and essential spectrum of $3 - D$ Dirac operators with regular and singular potentials