New answers tagged

0 votes

Reduce false positives having imbalanced data

One issue with SMOTE is overgeneralization: in a nutshell, it creates new points that are in the line joining two existing points of the minority class. This helps in some situations, but it might not ...
SeaEngineering's user avatar
2 votes

I build my first neural network! What's next?

In addition to the @brewmaster321's answer, in my opinion the hyper-parameters you need to care about are: Learning rate: especially if you use SGD as optimizer, which does not adapt the LR unlike ...
Luca Anzalone's user avatar
1 vote

I build my first neural network! What's next?

Welcome to the forum @Tim, and congratulations on your first neural network. If you're already using Keras, you can use the Keras tuner on your network instead of using a separate framework. The Keras ...
brewmaster321's user avatar
0 votes

How do we get output layer in skip-gram?

We have 2 embedding matrices(U,V) that are learnt during word2vec training. U has shape (vocab_size, dimensions) V has shape (dimension, vocab_size) For any given ...
lateBloomer's user avatar
0 votes

Appropriate input size for nn.Embedding

nn.embedding layer is a sytactic sugar equivalent to one hot vector+ linear layer. Suppose you have 2 distinct variables. and you want your model to learn their ...
lateBloomer's user avatar
0 votes

Feature importance in neural networks

As a solution, you can train single layer linear perceptron to find weights of each feature. After that iteratively drop the most useless feature (with min weight) and test combinations of remained ...
Vadym's user avatar
  • 1
0 votes

How to improve accuracy on a single class out of 3 classes in model

Looking at the numbers: Your classifier is quite good (or even excellent) on distinguishing class 0 and class 1. Contrary to that, class 2 is either recognized as class 0 or class 1. I would expect ...
Broele's user avatar
  • 1,147
0 votes
Accepted

Reduce false positives having imbalanced data

First of all, there will always be a trade-off between precision_score and recall_score. So you must choose a satisfying ...
Harshad Patil's user avatar
0 votes

How to improve accuracy on a single class out of 3 classes in model

Try creating one model for one class. Basically you can call it as class detector, it works as following: One model will predict class 0. One model will predict 1 and another one for class 2. Then ...
Harshad Patil's user avatar
1 vote

Reduce false positives having imbalanced data

Is this tabular data that you are exploring? If so, how many different models have you compared the results with? What is the variance of your precision and recall for both class given different ...
FabC's user avatar
  • 31
1 vote
Accepted

This simple python Feed forward Neural Network isn't learning. What am I doing wrong?

There are a couple of issues, but I think the main one is your training loop. An epoch is generally one pass through the entire dataset, and then backprop is performed, notwithstanding mini-batch or ...
brewmaster321's user avatar
0 votes

Why might a neural network consistently underestimate its target?

It seems like you are dealing with a zero inflated lognormal (ZILN) distributed target. In this case standard error loss / metrics are not correctly capturing the error structure. I suggest to take ...
Georg M. Goerg's user avatar
0 votes

What is the relationship between the accuracy and the loss in deep learning?

Ayúdenme a aclarar esta duda por favor Después de ver este comentario: ------si sus datos están entre 0 y 1, una pérdida de 0,5 es enorme, pero si sus datos están entre 0 y 255, un error de 0,5 es ...
jhazmin tapia's user avatar
0 votes

Why does cost function on a neural network increase?

I had similar problems and for me the reason was that my regularization parameter lambda was too high. I removed regularization completely and set it to 0 (that would be a good starting point). I was ...
Flavius Miron's user avatar
0 votes

Using conformal predictors to estimate uncertainty?

What you are looking for is calibrated probabilities, the name of Conformal Prediction method is Venn-ABERS predictors. There are several tutorials on my repo. https://github.com/valeman/awesome-...
valeman's user avatar
  • 11
3 votes
Accepted

Why my validation loss and accuracy decays over epochs?

You are experiencing a lot of overfitting on the training set here. I would go back and see if there are any inherent issues with the data (scaling? class imbalance? etc.) before diving into modeling. ...
NLP from scratch's user avatar
1 vote

What Model to Choose for a NN with a Very Wide Output Layer?

I've managed to find the solution! A model that can do it is a transformer. Here's a paper on such a model doing just this job: Predicting Properties of Quantum Systems with Conditional Generative ...
avpol's user avatar
  • 11
0 votes

How to optimize transposed convolution?

I suggest using a bilinear interpolation followed by a convolution instead of deconvolution. Deconvolution is prone to checkerboard artifacts. It may also helps in terms of execution speed.
Lelouch's user avatar
  • 131
1 vote

Tensorflow diagram for attention mechanism

This is a diagram of the attention layer that appears in the English version of the Tensorflow Transformer tutorial (other languages do not have this figure). The ingoing arrows are inputs to the ...
noe's user avatar
  • 22.3k
0 votes

Why are discriminative models denoted as $P\left(Y|X\right)$?

Actually it is not misleading at all. Without loss of the generality, we can refer to generative and discriminative classifiers. Both models predict $Y$ on the basis of $X$ in other words they have to ...
dobrowol's user avatar

Top 50 recent answers are included