We propose a deep learning method to build an AdS/QCD model from the data of hadron spectra. A major problem generic models is that large ambiguity allowed for bulk gravity metric with which QCD observables are holographically calculated. adopt experimentally measured spectra $\ensuremath{\rho}$ and ${a}_{2}$ mesons as training data, perform supervised machine determines concretely dilaton profile model. Our (DL) architecture based on AdS/DL correspondence [K. Hashimoto, S. Sugishita, A....

We derive an explicit form of the dilaton potential in improved holographic QCD (IHQCD) from experimental data $\ensuremath{\rho}$ meson spectrum. For this purpose, we make use emergent bulk geometry obtained by deep learning hadronic work Akutagawa et al. [Phys. Rev. D 102, 026020 (2020)]. Requiring that is a solution IHQCD derives corresponding backward. This determines action data-driven way, which enables us at same time to ensure proposal consistent gravity. Furthermore, find resulting...

Abstract We derive an explicit form of the dilaton potential in improved holographic QCD (IHQCD) from lattice data chiral condensate as a function quark mass. This establishes data-driven modeling QCD—machine-learning QCD. The consists two steps for solving inverse problems. first problem is to find emergent bulk geometry consistent with simulation at boundary. solve this refinement neural ordinary differential equation, machine-learning technique. second gravity action such that its...

We present our preliminary results on the machine learning estimation of $\text{Tr} \, M^{-n}$ from other observables with gradient boosting decision tree regression, where $M$ is Dirac operator. Ordinarily, obtained by linear CG solver for stochastic sources which needs considerable computational cost. Hence, we explore possibility cost reduction trace adoption algorithm. also discuss effects bias and its correction.

We present our preliminary results on the machine learning estimation of $\text{Tr} \, M^{-n}$ from other observables with gradient boosting decision tree regression, where $M$ is Dirac operator. Ordinarily, obtained by linear CG solver for stochastic sources which needs considerable computational cost. Hence, we explore possibility cost reduction trace adoption algorithm. also discuss effects bias and its correction.

A bstract We study 1 /N corrections to a Wilson loop in holographic duality. Extending the AdS/CFT correspondence beyond large N limit is an important but subtle issue, as it needs quantum gravity side. To find physical property of corrected geometry near-horizon black 0-branes previously obtained by Hyakutake, we evaluate Euclidean string worldsheet hanging down geometry, which corresponds rectangular SU( ) mechanics with 16 supercharges at finite temperature . that potential energy defined...

We derive an explicit form of the dilaton potential in improved holographic QCD (IHQCD) from lattice data chiral condensate as a function quark mass. This establishes data-driven modeling -- machine learning QCD. The consists two steps for solving inverse problems. first problem is to find emergent bulk geometry consistent with simulation at boundary. solve this refinement neural ordinary differential equation, technique. second gravity action such that its solution geometry. non-zero...