$\bf{
2.16 {}^{+0.49}_{0.26}}$

OUR EVALUATION

See the ideogram below. 

$2.6$ $\pm0.4$ 
^{ 1} 

THEO 
$2.130$ $\pm0.041$ 
^{ 2} 

LATT 
$2.27$ $\pm0.06$ $\pm0.06$ 
^{ 3} 

LATT 
$2.36$ $\pm0.24$ 
^{ 4} 

LATT 
$2.57$ $\pm0.26$ $\pm0.07$ 
^{ 5} 

LATT 
$2.24$ $\pm0.10$ $\pm0.34$ 
^{ 6} 

LATT 
$2.01$ $\pm0.14$ 
^{ 7} 

LATT 
• • We do not use the following data for averages, fits, limits, etc. • • 
$2.15$ $\pm0.03$ $\pm0.10$ 
^{ 8} 

LATT 
$1.9$ $\pm0.2$ 
^{ 9} 

LATT 
$2.01$ $\pm0.14$ 
^{ 7} 

LATT 
$2.9$ $\pm0.2$ 
^{ 10} 

THEO 
$2.9$ $\pm0.8$ 
^{ 11} 

THEO 
$3.02$ $\pm0.33$ 
^{ 12} 

LATT 
$2.7$ $\pm0.4$ 
^{ 13} 

THEO 
$1.9$ $\pm0.2$ 
^{ 14} 

LATT 
$2.8$ $\pm0.2$ 
^{ 15} 

THEO 
$1.7$ $\pm0.3$ 
^{ 16} 

LATT 
^{1} 
DOMINGUEZ 2019 determine the quark mass from a QCD finite energy sum rule for the divergence of the axial current.


^{2} 
BAZAVOV 2018 determine the quark masses using a lattice computation with staggered fermions and four active quark flavors.


^{3} 
FODOR 2016 is a lattice simulation with ${{\mathit N}_{{f}}}$ = 2 + 1 dynamical flavors and includes partially quenched QED effects.


^{4} 
CARRASCO 2014 is a lattice QCD computation of light quark masses using 2 + 1 + 1 dynamical quarks, with ${{\mathit m}_{{u}}}$ = ${{\mathit m}_{{d}}}{}\not=$ ${{\mathit m}_{{s}}}{}\not=$ ${{\mathit m}_{{c}}}$. The ${\mathit {\mathit u}}$ and ${\mathit {\mathit d}}$ quark masses are obtained separately by using the ${{\mathit K}}$ meson mass splittings and lattice results for the electromagnetic contributions.


^{5} 
AOKI 2012 is a lattice computation using 1 + 1 + 1 dynamical quark flavors.


^{6} 
BLUM 2010 determines light quark masses using a QCD plus QED lattice computation of the electromagnetic mass splittings of the lowlying hadrons. The lattice simulations use 2+1 dynamical quark flavors.


^{7} 
DAVIES 2010 and MCNEILE 2010 determine ${{\overline{\mathit m}}_{{c}}}({{\mathit \mu}})/{{\overline{\mathit m}}_{{s}}}({{\mathit \mu}}$) = $11.85$ $\pm0.16$ using a lattice computation with ${{\mathit N}_{{f}}}$ = 2 + 1 dynamical fermions of the pseudoscalar meson masses. Mass ${\mathit m}_{{{\mathit u}}}$ is obtained from this using the value of ${\mathit m}_{{{\mathit c}}}$ from ALLISON 2008 or MCNEILE 2010 and the BAZAVOV 2010 values for the light quark mass ratios, ${\mathit m}_{{{\mathit s}}}/{{\overline{\mathit m}}}$ and ${\mathit m}_{{{\mathit u}}}/{\mathit m}_{{{\mathit d}}}$.


^{8} 
DURR 2011 determine quark mass from a lattice computation of the meson spectrum using ${{\mathit N}_{{f}}}$ = 2 + 1 dynamical flavors. The lattice simulations were done at the physical quark mass, so that extrapolation in the quark mass was not needed. The individual ${\mathit m}_{{{\mathit u}}}$, ${\mathit m}_{{{\mathit d}}}$ values are obtained using the lattice determination of the average mass ${\mathit m}_{\mathrm {ud}}$ and of the ratio ${\mathit m}_{{{\mathit s}}}/{\mathit m}_{\mathrm {ud}}$ and the value of $\mathit Q$ = (${{\mathit m}^{2}}_{{{\mathit s}}}$ $−$ ${{\mathit m}^{2}}_{\mathrm {ud}}$) $/$ (${{\mathit m}^{2}}_{{{\mathit d}}}$ $−$ ${{\mathit m}^{2}}_{{{\mathit u}}}$) as determined from ${{\mathit \eta}}$ $\rightarrow$ 3 ${{\mathit \pi}}$ decays.



^{10} 
DOMINGUEZ 2009 use QCD finite energy sum rules for the twopoint function of the divergence of the axial vector current computed to order $\alpha {}^{4}_{s}$.


^{11} 
DEANDREA 2008 determine ${\mathit m}_{{{\mathit u}}}−{\mathit m}_{{{\mathit d}}}$ from ${{\mathit \eta}}$ $\rightarrow$ 3 ${{\mathit \pi}^{0}}$ , and combine with the PDG 2006 lattice average value of ${\mathit m}_{{{\mathit u}}}+{\mathit m}_{{{\mathit d}}}$ = $7.6$ $\pm1.6$ to determine ${\mathit m}_{{{\mathit u}}}$ and ${\mathit m}_{{{\mathit d}}}$.


^{12} 
BLUM 2007 determine quark masses from the pseudoscalar meson masses using a QED plus QCD lattice computation with two dynamical quark flavors.


^{13} 
JAMIN 2006 determine ${\mathit m}_{{{\mathit u}}}$(2 GeV) by combining the value of ${\mathit m}_{{{\mathit s}}}$ obtained from the spectral function for the scalar ${{\mathit K}}{{\mathit \pi}}$ form factor with other determinations of the quark mass ratios.


^{14} 
MASON 2006 extract light quark masses from a lattice simulation using staggered fermions with an improved action, and three dynamical light quark flavors with degenerate ${\mathit {\mathit u}}$ and ${\mathit {\mathit d}}$ quarks. Perturbative corrections were included at NNLO order. The quark masses ${\mathit m}_{{{\mathit u}}}$ and ${\mathit m}_{{{\mathit d}}}$ were determined from their (${\mathit m}_{{{\mathit u}}}+{\mathit m}_{{{\mathit d}}})/$2 measurement and AUBIN 2004A ${\mathit m}_{{{\mathit u}}}/{\mathit m}_{{{\mathit d}}}$ value.


^{15} 
NARISON 2006 uses sum rules for ${{\mathit e}^{+}}$ ${{\mathit e}^{}}$ $\rightarrow$ hadrons to order ${{\mathit \alpha}_{{s}}^{3}}$ to determine ${\mathit m}_{{{\mathit s}}}$ combined with other determinations of the quark mass ratios.


^{16} 
AUBIN 2004A employ a partially quenched lattice calculation of the pseudoscalar meson masses.

