1 ∣ Θ) = t, r: hyperparameters, validation set で tuning Data: A s t r o P h , C o n d M a t , G r Q c , H e p P h (social networks, undirected) 正例負例の扱い: 実験1 同様 Loss: cross entropy e + 1 (d(u,v)−r)/t 1 21
slide Facebook Research just published an awesome paper on learning hierarchical representations (2017‑06‑14) [R] Poincaré Embeddings for Learning Hierarchical Representations : MachineLearning 28