Search Results
| Search type | Search syntax |
|---|---|
| Tags | [tag] |
| Exact | "words here" |
| Author |
user:1234 user:me (yours) |
| Score |
score:3 (3+) score:0 (none) |
| Answers |
answers:3 (3+) answers:0 (none) isaccepted:yes hasaccepted:no inquestion:1234 |
| Views | views:250 |
| Code | code:"if (foo != bar)" |
| Sections |
title:apples body:"apples oranges" |
| URL | url:"*.example.com" |
| Saves | in:saves |
| Status |
closed:yes duplicate:no migrated:no wiki:no |
| Types |
is:question is:answer |
| Exclude |
-[tag] -apples |
| For more details on advanced search visit our help page | |
Results tagged with neural-networks
Search options not deleted
user 26948
Artificial neural networks (ANNs) are a broad class of computational models loosely based on biological neural networks. They encompass feedforward NNs (including "deep" NNs), convolutional NNs, recurrent NNs, etc.
3
votes
Neural network hidden layer output
It passes on the result of the sigmoid function and does not binarize the output. The reason for this is that the neural network must (usually) be differentiable for the backpropagation training algor …
- 26.5k
1
vote
Accepted
Is it better to have one neural network per feature, or a single neural network for everything?
I think this would require some expert domain knowledge to decide which approach is better, although ultimately it is better to try both ways and see.
If you don't expect the interaction between the …
- 26.5k
1
vote
"Forward Propagation" in Neural Networks
Yes, it is typical that a loss function depends on more than just the output layer. For example with weight decay regularization, the loss is a function of all the weights in the network as well as th …
- 26.5k
0
votes
How can I use an $n$-input neural network to predict some values for $n-1$ inputs?
You could have an additional $n$ binary inputs which indicates whether each object is present or not.
Alternatively, you could have an architecture which can take in a variable number of inputs and pr …
- 26.5k
0
votes
Retrieving similar frames
Yes, you could find the nearest neighbor in your database for each test image. Standard L2 distance or cosine similarity should work well.
- 26.5k
1
vote
Accepted
Can all regression networks be reduced to one output?
Yes, there is a bijection betweeen $\mathbb{R}$ and $\mathbb{R}^n$ for any $n \geq 1$. No, it's not a good idea to take advantage of this bijection to turn a multivariate regression problem into a uni …
- 26.5k
2
votes
What are exactly restricted boltzmann machines?
A feed forward neural network is just a function $f(x; \theta)$ of some inputs and its weights. There is a clear procedure for computing the values of each unit/layer in the network from the previous …
- 26.5k
1
vote
Varriational autoencoder latent space and editing the images
In VAE we encode the image and take its statistics (mean and variance)
I wouldn't take the word "encode" too seriously here. The encoder parameterizes an approximate posterior $p(z|x)$.
1) Ho …
- 26.5k
1
vote
which layer contribute more in deep neural network
Which part of a car contributes more to its ability to transport people from point A to B? Is it the tires, engine, or transmission? With any of these, the car wouldn't move very well, so it doesn't m …
- 26.5k
2
votes
Accepted
How many neurons are in DALL-E?
There's reason to believe that number of parameters, not number of neurons, is the right metric by which to measure model size. And although number of parameters isn't comparable to neurons in the hum …
- 26.5k
7
votes
Can any one explain why dot product is used in neural network and what is the intitutive tho...
The reason we use dot products is because lots of things are lines.
One way of seeing it is that the use of dot product in a neural network originally came
from the idea of using dot product in lin …
- 26.5k
6
votes
What is the continuous depth concept in Neural ODE paper by David Duvenaud?
I wouldn't read into the term "continuous depth" too much. It's just that since the ODE allows you to evaluate the neural network at any layer (for example we could compute $h(\pi)$ to obtain the valu …
- 26.5k
5
votes
Accepted
Why is every local minimum of a neural network with one layer and $H$ hidden units in a fami...
If the activation used is an even function: $f(x)=−f(−x)$, then one could choose to "flip" a neuron by flipping the sign of all the input and the output weights associated with that neuron. The abilit …
- 26.5k
8
votes
Accepted
Capacity and expressivity of a neural network
There definitely is a lot of overlap and interchangeability in how those terms are commonly used. I think the main distinction is that expressivity is often used to talk about what classes of function …
- 26.5k
2
votes
Can neural networks figure out some unknown transform?
A neural network would be overkill for this task, since the linear transform you have described is essentially a neural network with no hidden layers and no activation function. NN's should be used wh …
- 26.5k