A Fast, Compact Approximation of the Exponential Function

By N. Schraudolph. 1999

Subfields: Optimization & Training

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Optimization & Training

1. Thomas Unterthiner
2. Sepp Hochreiter
3. Chia-Hua Ho
4. Chih-Jen Lin
5. Guo-Xun Yuan
6. Djork-Arné Clevert
7. Stefano Ermon
8. Dan Hendrycks

Citations in the corpus

• Fast and Accurate Deep Network Learning by Exponential Linear Units (ELUs) (2015, methodology): “ets are 5% slower on ImageNet. In terms of wall clock time (12.15h vs. 11.48h for 10k iterations), we expect this to improve by using a more sophisticated implementation, e.g. following the advice of Schraudolph (1999). 5 CONCLUSION In this paper we introduced the novel exponential linear units, uni”
• An improved GLMNET for l1-regularized logistic regression (2011, methodology): “5 Approximate Exponential Operations for CDN Because CDN suffers from slow exp/log operations, we tried to use the approximate exponentiation proposed by Schraudolph (1999). However, we failed to speed up CDN because of erroneous numerical results.”