Feedforward neural network: Revision history

Filter revisionsshowhide

To date:

Tag filter:

Invert selection

External tools:

For any version listed below, click on its date to view it. For more help, see Help:Page history and Help:Edit summary. (cur) = difference from current version, (prev) = difference from preceding version, m = minor edit, → = section edit, ← = automatic edit summary

curprev 06:0206:02, 13 July 2024‎ GorillaGlueMan talk contribs‎ m 21,679 bytes +43‎ →‎Timeline: Added link to attention is all you need undo

curprev 10:0410:04, 1 July 2024‎ Citation bot talk contribs‎ 21,636 bytes +13‎ Altered isbn. Added pages. Removed parameters. Some additions/deletions were parameter name changes. Upgrade ISBN10 to 13. | Use this bot. Report bugs. | Suggested by Abductive | Category:Neural network architectures | #UCB_Category 22/27 undo

curprev 00:3300:33, 24 June 2024‎ 2.121.108.30 talk‎ 21,623 bytes +4‎ Non-Linear Activation Functions: To approximate any continuous function, we need non-linear activation functions. The Universal Approximation Theorem states that a feedforward neural network with at least one hidden layer and non-linear activation functions (such as sigmoid, tanh, or ReLU) can approximate any continuous function to any desired degree of accuracy, given a sufficient number of neurons. Multiple Parallel Linear Units: The statement about multiple parallel linear units ap undo

curprev 06:3306:33, 28 April 2024‎ WikiCleanerBot talk contribs‎ m 21,619 bytes −1‎ v2.05b - Bot T20 CW#61 - Fix errors for CW project (Reference before punctuation) undo Tag: WPCleaner

curprev 15:1115:11, 24 April 2024‎ NicGambarde talk contribs‎ 21,620 bytes −191‎ →‎Multilayer perceptron: Removing the misnomer mention. 1) the original perceptron was 1 neuron, so the activation function is not the issue. 2) modern MLP not always use non-linear functions. 3) non-linear functions were used very early in the perceptron history. Anecdotal claim at best IMHO undo
curprev 15:0915:09, 24 April 2024‎ NicGambarde talk contribs‎ m 21,811 bytes +83‎ →‎Multilayer perceptron: Adding FCN synonym for MLP undo

curprev 15:0115:01, 13 January 2024‎ MrOllie talk contribs‎ 21,556 bytes −684‎ →‎Timeline: these are much, much less notable than the other entries here, removing per WP:UNDUE undo
curprev 14:5814:58, 13 January 2024‎ 117.251.46.226 talk‎ 22,240 bytes +145‎ Backpropagation algorithm for efficiently training higher order quadratic neural networks. undo Tag: Visual edit

curprev 13:3613:36, 11 August 2023‎ Quebec99 talk contribs‎ m 24,450 bytes −209‎ fixed duplicate reference undo
curprev 13:2913:29, 11 August 2023‎ Quebec99 talk contribs‎ m 24,659 bytes −69‎ References after punctuation per WP:REFPUNC and WP:PAIC undo

curprev 05:0505:05, 9 August 2023‎ DancingPhilosopher talk contribs‎ 24,728 bytes 0‎ →‎Timeline: timeline undo Tags: Mobile edit Mobile app edit Android app edit
curprev 05:0005:00, 9 August 2023‎ DancingPhilosopher talk contribs‎ 24,728 bytes −3‎ →‎Timeline: timeline undo Tags: Mobile edit Mobile app edit Android app edit
curprev 04:4404:44, 9 August 2023‎ DancingPhilosopher talk contribs‎ 24,731 bytes +657‎ →‎Timeline: timeline undo Tags: Mobile edit Mobile app edit Android app edit

curprev 05:4205:42, 28 July 2023‎ WikiCleanerBot talk contribs‎ m 19,954 bytes 0‎ v2.05b - Bot T20 CW#61 - Fix errors for CW project (Reference before punctuation) undo Tag: WPCleaner