Markov Logic Network ใฏใๅคๆฐ \(x_1 .. x_n \in \textrm{Bool}\) ใซๅฐฑใใฆๆใ็ซใกใใใ ใจๆใ
ๅฝ้ก \(P_1(x_1 .. x_n) .. P_m(x_1 .. x_n)\) ใใ็ทฉใๆจๅฎใ่กใใขใใซใ
็ซใฆใๅฝ้กใฎไธญใซใฏ็็พใใใใฎใใใฃใฆใใใใ
ไพใใฐๆงๅ่ชฌใจๆงๆช่ชฌใฏไธก็ซใใชใใ
ใใใงใ\(m\)ๅใฎๅฝ้กใซๅฏพใใฆ\(m\)ๅใฎ้ใฟ\(\{w_1 .. w_m\}\)ใ็จๆใใใ
ใใฎๅคใฏ่จ็ทดใใผใฟใใๅญฆ็ฟใใใ
ใใฎใขใใซใซใใใๅญฆ็ฟใจใฏใๆฌกใฎใใใชๆๅคงๅใซ้ใใชใใ
\[ maximize \sum w_i q_i(x) \]
ใใ ใใใใงใ\(q_i(x) = \textrm{if} ~ P(x) ~ \textrm{then} ~ 1 ~ \textrm{else} ~ 0\)ใงใใใ
ใใใใใใใ\(\{0, 1\}\) ใใใชใใฆ \(\{-1, +1\}\) ใชใฎใใใ
ใใจใ้ใฟใฏๅ
จใฆๆญฃใงๅใ1ใ ใจใใใใใๆกไปถใใใใฎใใใ
ๆจๅฎใฎๆใใๅ
ใปใฉใฎๆๅคงๅใใๅญฆ็ฟใงๅพใ้ใฟ\(\{w_i\}\)ใ็จใใฆ่กใใ
ใใใ็ทฉใๆจ่ซใใใจ่จใฃใๅฟใงใใใ