Questions tagged [bptt]

For questions about the backpropagation through time (BPTT) algorithm, which is often used to find the gradient of the objective function with respect to the parameters of a recurrent neural network (RNN) when training the RNN with gradient descent.

Filter by
Sorted by
Tagged with
0
votes
0answers
18 views

How Long Can BPTT Truncated?

I wanted to ask what is, in general, the maximum value (the order of magnitude) of the number of time steps I can back-propagate in the past using TBTT (Truncated Backpropagation Through Time) in an ...
0
votes
0answers
20 views

RNN - Backpropagation through time - Gradient Calculation

I think I got it right after reading multiple resources but im still not 100%. Seems like everyone is calculating it different. Or they just shortcut explaining the calculation. (or my math skills ...
0
votes
0answers
70 views

What's wrong with my understanding of how RNNs work?

Recently, I've been trying to derive the mathematics behind various Neural Network structures. I managed to derive the MLP and tested it to be on par with a Keras implementation (Using the MNIST ...
0
votes
0answers
36 views

In LSTMs, how does the additive property enables better balancing of gradient values during backpropagation?

There are two sources that I'm using to to try and understand why LSTMs reduce the likelihood of the vanishing gradient problem associated with RNNs. Both of these sources mention the reason LSTMs are ...
3
votes
0answers
156 views

What is the time complexity for training a gated recurrent unit (GRU) neural network using back-propagation through time?

Let us assume we have a GRU network containing $H$ layers to process a training dataset with $K$ tuples, $I$ features, and $H_i$ nodes in each layer. I have a pretty basic idea how the complexity of ...