%0 Conference Paper %1 Hendrycks2016 %A Hendrycks, Dan %A Gimpel, Kevin %D 2016 %K cs.LG %T Gaussian Error Linear Units (GELUs) %X We propose the Gaussian Error Linear Unit (GELU), a high-performing neural network activation function. The GELU nonlinearity is the expected transformation of a stochastic regularizer which randomly applies the identity or zero map to a neuron's input. The GELU nonlinearity weights inputs by their magnitude, rather than gates inputs by their sign as in ReLUs. We perform an empirical evaluation of the GELU nonlinearity against the ReLU and ELU activations and find performance improvements across all considered computer vision, natural language processing, and speech tasks.

@inproceedings{Hendrycks2016, abstract = {We propose the Gaussian Error Linear Unit (GELU), a high-performing neural network activation function. The GELU nonlinearity is the expected transformation of a stochastic regularizer which randomly applies the identity or zero map to a neuron's input. The GELU nonlinearity weights inputs by their magnitude, rather than gates inputs by their sign as in ReLUs. We perform an empirical evaluation of the GELU nonlinearity against the ReLU and ELU activations and find performance improvements across all considered computer vision, natural language processing, and speech tasks.}, added-at = {2020-10-15T14:36:56.000+0200}, author = {Hendrycks, Dan and Gimpel, Kevin}, biburl = {https://www.bibsonomy.org/bibtex/2e4ef1a028428ce25fb3d16bfb9128d91/annakrause}, eprint = {http://arxiv.org/abs/1606.08415v3}, eprintclass = {cs.LG}, eprinttype = {arXiv}, file = {:http\://arxiv.org/pdf/1606.08415v3:PDF}, groups = {TODO}, interhash = {078d9aa80b94ec084327a83005890ea3}, intrahash = {e4ef1a028428ce25fb3d16bfb9128d91}, keywords = {cs.LG}, timestamp = {2020-10-15T14:44:42.000+0200}, title = {{Gaussian Error Linear Units (GELUs)}}, year = 2016 }