English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Knowledge graphs (KGs) represent world's facts in structured forms. KGcompletion exploits the existing facts in a KG to discover new ones.Translation-based embedding model (TransE) is a prominent formulation to do KGcompletion. Despite the efficiency of TransE in memory and time, it suffersfrom several limitations in encoding relation patterns such as symmetric,reflexive etc. To resolve this problem, most of the attempts have circledaround the revision of the score function of TransE i.e., proposing a morecomplicated score function such as Trans(A, D, G, H, R, etc) to mitigate thelimitations. In this paper, we tackle this problem from a differentperspective. We show that existing theories corresponding to the limitations ofTransE are inaccurate because they ignore the effect of loss function.Accordingly, we pose theoretical investigations of the main limitations ofTransE in the light of loss function. To the best of our knowledge, this hasnot been investigated so far comprehensively. We show that by a properselection of the loss function for training the TransE model, the mainlimitations of the model are mitigated. This is explained by settingupper-bound for the scores of positive samples, showing the region of truth(i.e., the region that a triple is considered positive by the model). Ourtheoretical proofs with experimental results fill the gap between thecapability of translation-based class of embedding models and the lossfunction. The theories emphasise the importance of the selection of the lossfunctions for training the models. Our experimental evaluations on differentloss functions used for training the models justify our theoretical proofs andconfirm the importance of the loss functions on the performance.

Toward Understanding The Effect Of Loss function On Then Performance Of Knowledge Graph Embedding