New answers tagged deep-learning
0
Yes, in the case of the Gaussian, you have two distinct layers (so weights and biases), one for the mean and the other for the variance, as the equations are telling us.
The mean is calculated with the weights $\mathbf{W}_{4}$ and bias $\mathbf{b}_{4}$ from $\mathbf{h}$ as follows
$$\boldsymbol{\mu} =\mathbf{W}_{4} \mathbf{h}+\mathbf{b}_{4},$$
while the ...
0
I think if you want to resize and reduce your matrix size, you can use one of the dimension reduction techniques.
Here there is a link that may be helpful to you.
0
The concepts of trace and tensor also appear in other contexts outside of machine learning (ML), like quantum computing, so an answer to your question may be given independently of ML, but that may not be useful, as these concepts may be defined and implemented differently in the context of ML, which seems to be the case.
The concept of trace, in mathematics,...
0
As you observed, convolutions are "shift/translation equivariant". This is extremely useful and beneficial for image/video/audio processing where this "symmetry" exists in the underlying domain.
This is not the case in your settings. Each card from each suit carries a different meaning. You actually want to have a different (trainable) ...
-1
They are not referring to the same stride used in CNN. In CNN the stride tells you how many pixels you move on between the calculations. The stride of your reference is about optimizing hardware storage usage.
Example:
If you had a 2x2 array, you would need to store 4 physical elements. In the case that the second row of the array is just the duplicate of ...
1
I am currently working on a similar problem. I think your approach is good. As for setting the parameter lambda, since you are using deep neural networks, you can make it a learnable parameter, instead of a hyperparameter you set. This way, as the two losses fluctuate over your training iterations/epochs, the model will be able to adjust the lambda parameter ...
0
My understanding from this paper https://www.microsoft.com/en-us/research/wp-content/uploads/2016/02/TASLP2339736-proof.pdf was that the filters are used on each input channels (i.e input feature map in the paper) separately and the result is summed, as described in eq. 8. Here they use different filters for each channel, but you could totally use the same ...
0
How about you make a dataset using patches from your image, and train a CNN model with that dataset?
That is, if you want to train a neural net, a dataset with 11 images for each class is too small and thus is prone to overfitting.
However, since your image is high resolution and the printers can be classified by just using the zoomed in images, you can ...
1
A quick search about reinforcement learning applied to video games will lead you to countless tutorials that describe exactly what you're asking for.
With images the way to go is usually deep reinforcement learning. A convolutional neural network (or any other deep learning architecture) is used to process the image and compress it to a latent vector used as ...
0
Another view on this topic: think about the derivative of the MSE with respect to the inputs. You will need to apply the chain rule:
$$\frac{dMSE}{dx_i}=\sum_i2(y_i-f(x_i))\cdot(-1)\cdot\frac{df(x_i)}{dx_i}$$
which only resembles the derivative of the parabola when, as mentioned in nbro's answer, $f(x_i)=c, \forall\ i$ or, more generally, when they $f(x)$ is ...
0
[I wanted it to be a comment but it's too long :)]
I don't think it's a good approach to split point of views into a group of 12 angles. The main purpose of using neural net is to have model that is able to generalize the data. That means the perfect model will be able to recognize an object in every orientation. Your task is to create a model that will be ...
1
Positional Encodings in Transformers exist to give the model some information about the position of the embedding. This makes sense in fields like NLP or Time Series Data, since the position(order) matters in this case.
However, since you say that order of the data is not relevant in your use case, positional encoding would not be necessary.
1
Ideal advise is to feed the raw data to the neural networks to let neural networks make its own inference
Considering you have expert knowledge that $X1/X2$ has effect on $Y$ , here the new feature ($X1/X2$) is referred to as a derived feature
However, there are few advantages which can help you consider of using derived features like $X1/X2$ in your case
...
4
Adding something to nbro answer, from my personal experience there are also some hints that can quickly tell you if you're dealing with a good machine learning paper, i.e. worth to read in its entirety or not.
In random order:
Clear contribution description: machine learning and artificial intelligence in general are both broad fields. A paper can be about ...
2
You are refering to the first and very important step in a machine learning process called data preprocessing. Refering to this article, inside data preprocessing there are many smaller processes that deal with features: feature extraction, feature selection, feature aggregration and feature encoding to name a few.
The idea of creating new features out of ...
2
I have some experience reading research papers. However, in my view, there is no single answer to this question (apart from this answer I am giving you, i.e. "it depends"). The answer to your question depends on
your background knowledge/education
If you don't know much about the specific topic in the paper, you may need to study at least briefly ...
1
A different, variable reward structure might help. You could try a combination of airspeed, pitch, roll and whether it is hovering in the air or not in each timestep as a representation for the reward.
Maybe airspeed should, in expectation, contribute up to 30% of the reward, pitch up to 15%, roll up to 15% and being in the air up to 40%. This would ...
2
Language models produce a probability distribution over a set of words. You determine the next word by sampling from this distribution. So, determining the next word is stochastic even though the distribution is the same given the initial prompt.
1
Lifelong learning and MLOps are indeed complementary.
Lifelong learning (LL) can be defined as the set of learning algorithms and models that can deal with more and more data and/or tasks without forgetting (completely) the previous one and (usually) without fully retraining the model with all data that you have available now. So, in LL, we attempt to mimic ...
0
we are recognizing the disease, not the person.
If you're training a computer vision model with only images and no auxiliary information then a randomized sampling should be enough to prevent the model from over fitting on x-ray scans taken on the same person.
If images from the same person exist in both subsets, the problem will be easy, and not reasonable, ...
0
Model and Objective function are playing together. If you can have an objective function that somehow exclude this relation that you have in your mind, then the model can focus on learning to predict based on other information. Then you have trained the model, but if your downstream task should consider that at the end, you could manually add this relation ...
1
Training using only 20 timesteps at a time is far too small, especially when the goal will ultimately consist of episodes of length 6000. You definitely need to increase that substantially and that will probably solve your problem immediately. You might try something like simulate 5 episodes and then train on all timesteps in those 5 episodes.
If that still ...
Top 50 recent answers are included