similarity Laplacian Eigenmaps Wi,j= If connected 0 otherwise Review in semi-supervised learning:If x1 and x2 are close in a high density region,1 and 2 are probably the same. t=∑co )+ As a regularization term s=2∑0-'=yy L:(R+U)x(R+U)matrix S evaluates how smooth your label is Graph Laplacian L=D-WLaplacian Eigenmaps • Review in semi-supervised learning: If 𝑥 1 and 𝑥 2 are close in a high density region, 𝑦 ො 1 and 𝑦 ො 2 are probably the same. 𝐿 = ෍ 𝑥 𝑟 𝐶 𝑦 𝑟 , 𝑦 ො 𝑟 +𝜆𝑆 As a regularization term = 𝒚 𝑇 𝑆 = 𝐿𝒚 1 2 ෍ 𝑖,𝑗 𝑤𝑖,𝑗 𝑦 𝑖 − 𝑦 𝑗 2 L: (R+U) x (R+U) matrix Graph Laplacian 𝐿 = 𝐷 − 𝑊 S evaluates how smooth your label is 𝑤𝑖,𝑗 = 0 similarity If connected otherwise