There has been a surge of momentum for deep representationlearning in hyperbolic spaces due to their capacity of modeling data like knowledge graphs or synonym hierarchies, possessing hierarchical structure . Hyperbolic deep neural networks potentially lead to drastically compact models with much more physical interpretability than its counterpart in Euclidean space . This paper presents acoherent and comprehensivereview of the literature around the neural components in the construction ofhyperbolic neuralnetworks . It also presents current applications around various machine learning tasks on several publicly availabledatasets, together with insightful observations and identifying openquestions and promising future directions .

Author(s) : Wei Peng, Tuomas Varanka, Abdelrahman Mostafa, Henglin Shi, Guoying Zhao

Links : PDF - Abstract

Code :

https://github.com/HazyResearch/hgcn


Coursera

Keywords : deep - hyperbolic - neural - presents - networks -

Leave a Reply

Your email address will not be published. Required fields are marked *