Viet-Anh on Software Logo

What is: Model-based Subsampling?

SourceModel-based Subsampling for Knowledge Graph Completion
Year2000
Data SourceCC BY-SA - https://paperswithcode.com

To avoid the problem caused by low-frequent entity-relation pairs, our MBS uses the estimated probabilities from a trained model θ\mathbf{\theta}' to calculate frequencies for each triplet and query. By using θ\mathbf{\theta}', the NS loss in KGE with MBS is represented as follows: \begin{align} &\ell_{mbs}(\mathbf{\theta};\mathbf{\theta}') \nonumber \ =&-\frac{1}{|D|}\sum_{(x,y) \in D} \Bigl[A_{mbs}(\mathbf{\theta}')\log(\sigma(s_{\mathbf{\theta}}(x,y)+\gamma))\nonumber\ &+\frac{1}{\nu}sum_{y_{i}\sim p_n(y_{i}|x)}^{\nu}B_{mbs}(\mathbf{\theta}')\log(\sigma(-s_{\mathbf{\theta}}(x,y_i)-\gamma))\Bigr], \end{align}