activity
20102018
most citedEssentially No Barriers in Neural Network Energy Landscape

177 citations · 249 across the 5 of their papers we have counts for

collaborators

5 papers

cond-mat.str-el2018★ 5 cited

Low-Energy Effective Theory at a Quantum Critical Point of the Two-Dimensional Hubbard Model: Mean-Field Analysis

Kambis Veschgini, Manfred Salmhofer

We complement previous functional renormalization group (fRG) studies of the two-dimensional Hubbard model by mean-field calculations. The focus falls on Van Hove filling and the t…

stat.ML2018★ 177 cited

Essentially No Barriers in Neural Network Energy Landscape

Felix Draxler, Kambis Veschgini, Manfred Salmhofer +1

Training neural networks involves finding minima of a high-dimensional non-convex loss function. Knowledge of the structure of this energy landscape is sparse. Relaxing from linear…

cond-mat.str-el2013★ 12 cited

Schwinger-Dyson Renormalization Group

Kambis Veschgini, Manfred Salmhofer

We use the Schwinger-Dyson equations as a starting point to derive renormalization flow equations. We show that Katanin's scheme arises as a simple truncation of these equations. W…

hep-th2010★ 27 cited

Trouble Finding the Optimal AdS/QCD

K. Veschgini, E. Megias, H. J. Pirner

In the bottom-up approach to AdS/QCD based on a five-dimensional gravity dilaton action the exponential of the dilaton field is usually identified as the strong or 't Hooft couplin…

hep-ph2010★ 28 cited

QCD-Thermodynamics using 5-dim Gravity

E. Megias, H. J. Pirner, K. Veschgini

We calculate the critical temperature and free energy of the gluon plasma using the dilaton potential arXiv:0911.0627[hep-ph] in the gravity theory of AdS/QCD. The finite temperatu…