Questions tagged [loss-functions]

For questions related to the concept of loss (or cost) function in the context of machine learning.

Filter by
Sorted by
Tagged with
1
vote
1answer
702 views

Why isn't the reverse KL divergence commonly used in supervised learning?

Forward KL Divergence (also known as cross entropy loss) is a standard loss function in supervised learning problems. I understand why it is so: matching a known a trained distribution to a known ...
1
vote
1answer
52 views

Which local minima to choose according to the shape of the error surface?

The following plot shows error function output based on system weights. Two equal local minima are shown in green pointers. Note that the red dots are not related to the question. Considering the ...
1
vote
1answer
56 views

How do I calculate $max_{a′}Q(s′,a′,w−)$ when it is represented as a neural network?

Consider the following loss function $$ L(\mathbf{w}) = [(r + \gamma max_{a'} Q(s', a', \mathbf{w^-})) - Q(s, a, \mathbf{w})]^2 $$ where $Q(s, a, \mathbf{w^-})$ and $Q(s, a, \mathbf{w})$ are ...
1
vote
1answer
27 views

Training by one batch of examples, what does it mean

Say I have a batch of examples, each examples represent a state: ...
1
vote
1answer
77 views

How to understand marginal loglikelihood objective function as loss function (explanation of an article)?

I am reading article https://allenai.org/paper-appendix/emnlp2017-wt/ http://ai2-website.s3.amazonaws.com/publications/wikitables.pdf about training neural network and the loss function is mentioned ...
1
vote
0answers
31 views

Could the inputs of the mean squared-error loss function be transformed to allow larger learning rates?

In the context of a neural network $\hat{y} = f_\theta(\mathbf{x})$ with parameters $\theta$ that is trained to perform regression such that the prediction $\hat{\mathbf{y}} = [\hat{y}_1,\hat{y}_2,...,...
1
vote
0answers
16 views

Is it normal getting noise values in the error history along training iteration?

I'm giving my first steps in really learning machine learning. As an exercise in my online course, it was asked for me to code the Cost function of some neural network that should resolve the ...
1
vote
0answers
19 views

How to deal with losses on different scales in multi-task learning?

Say I'm training a model for multiple tasks by trying to minimize sum of losses $L_1 + L_2$ via gradient descent. If these losses are on a different scale, the one whose range is greater will dominate ...
1
vote
0answers
48 views

BlackOut - ICLR 2016: need help understanding the cost function derivative

In the ICLR 2016 paper BlackOut: Speeding up Recurrent Neural Network Language Models with very Large Vocabularies, on page 3, for eq. 4: $$ J_{ml}^s(\theta) = log \ p_{\theta}(w_i | s) $$ They have ...
1
vote
0answers
20 views
1
vote
0answers
37 views

How to make SAC (Soft-Actor-Critic) learn a policy?

I cannot make SAC learn a task in a certain environment. The point is that it actually sometimes finds a very good policy, but it never learns the policy in the end. I am using the SAC implementation ...
1
vote
2answers
119 views

If the training data are linearly separable, which of the following $L(w)$ has less optimum answer for $w$, when $y = w^Tx$?

I'm studying machine learning and I came into a challenging question. The answer is 2. But based on my ML notes, all of them are true. Where are the wrong points?
1
vote
0answers
17 views

What is the intuition behind equations 10, 11 and 12 of the paper “Noise2Noise: Learning Image Restoration without Clean Data”?

Can anyone help me understand these functions described in the paper Noise2Noise: Learning Image Restoration without Clean Data I have read the portion A.4 in the appendix but need a more detailed and ...
1
vote
1answer
59 views

Explanation of this L2 minimization equation

I am trying to understand the last two lines of this math notation. How Var and double summation of Cov came to the equation. The first two lines I understood something like $(a-b)^2 = a^2 -2ab +b^2$.
1
vote
1answer
63 views

Explain the difference in graphical patterns between discriminator fake loss and generator loss in GAN

In GAN (generative adversarial networks), let us take "binary cross-entropy" as the loss function for discriminator $$(overall \; loss = -\sum log(D(x_i)) -\sum log(1-D(G(z_i))) $$ $$ where \...
1
vote
0answers
28 views

What is a “center loss”?

I have seen that a center loss is beneficial in computer vision, especially in face recognition. I have tried to understand this concept from the following material A Discriminative Feature Learning ...
1
vote
0answers
24 views

Could the data augmentation lead to the model learning features which corresponds to data augmented data and not to the real data?

I am trying to train a Unet network with Synthetic data to do binary segmentation due to the fact that is is not easy to collect real data. And there is something in the training process that I do not ...
1
vote
0answers
29 views

How to find distance between 2 points when dimensions are all of different nature?

I have a dataset with four features: the x coordinate the y coordinate the velocity magnitude angle Now, I want to measure the distance between two points in the dataset, taking into account the ...
1
vote
0answers
64 views

Which loss function to choose for imbalanced datasets?

For imbalanced datasets (either in the context of computer vision or NLP), from what I learned, it is good to use a weighted log loss. However, in competitions, the people who are in top positions are ...
1
vote
0answers
40 views

Why would the loss increase on a single fixed input?

I'm training a neural network on some input data. I know that loss increasing may be related to: overfitting, if the loss increases on test data (while still decreases on training data) oscillations ...
1
vote
0answers
111 views

Classification or regression for deep Q learning

DQN implemented at https://github.com/PacktPublishing/PyTorch-1.x-Reinforcement-Learning-Cookbook/blob/master/Chapter07/chapter7/dqn.py uses the mean square error loss function for the neural network ...
1
vote
0answers
57 views

Loss Function In Units Of Bits?

Where can I find a machine learning library that implements loss functions measuring the Algorithmic Information Theoretic-friendly quantity "bits of ...
1
vote
0answers
58 views

LSTM - MAPE Loss Function gives Better Results when Data is De-Scaled before Loss Calculation

I am building an LSTM for predicting a price chart. MAPE resulted in the best loss function compared to ...
1
vote
0answers
36 views

What are some good loss functions used to minimize extreme errors in regression and time series forecasting?

I'm working on a time series forecasting task, and, in some specific cases, I don't need perfect accuracy, but the network cannot by any means miss by a lot. So, in detriment of a smaller mean error, ...
1
vote
0answers
44 views

Simplifying Log Loss

I am reading through a paper (https://www.mitpressjournals.org/doi/pdf/10.1162/0891201053630273) where they describe logloss as a ranking function and can be simplified to the margin of the training ...
1
vote
0answers
219 views

what will be the best loss function for unet to predict the each pixel values?

I'm predicting the used 9 pictures to predict the last picture so (40,40,9) -> unet -> (40,40,1) but as you see the predict picture It's not just a mask(0or 1) its float so which loss function ...
1
vote
0answers
48 views

How MSE should be appliead with multi target deep network?

I'm having a problem understanding how the MSE should be used when working with a multidimensional target, e.g 3 dimensiones. (My outputs are continuois values, not categorical) Let us say I have a ...
1
vote
1answer
92 views

How does the DQN loss from td_targets against q_values make sense?

Why td_loss is calculated from (td_targets against q_values)? Why I am lost is because: <...
1
vote
0answers
22 views

How to update edge features in a graph using a loss function?

Given a directed, edge attributed graph G, where the edge attribute is a probability value, and a particular node N (with binary features f1 and f2) in G, the algorithm that I want to implement is as ...
1
vote
0answers
27 views

Face recognition model loss not decreasing

I wrote a script to do train a Siamese Network style model for face recognition on LFW dataset but the training loss doesnt decrease at all. Probably there's a bug in my implementation. Could you ...
1
vote
0answers
42 views

How to reduce fluctuation of a neural network?

I've modeled an AlexNet neural network, with 50 epochs and a batch size of 64. I used a stochastic gradient descent optimizer with a learning rate of 0.01. I attached the train and validation loss and ...
1
vote
0answers
24 views

Keras MLP returns always loss 0.0 [closed]

I'm implementing a multilayer perceptron with Keras to predict the correct words order in a sentence. I'm using train_on_batch()because I convert each sentence in a ...
1
vote
1answer
105 views

RealNVP gives wrong probabilities

I am trying to use RealNVP with some data I have (the input size is a 1D vector of size 22). Here is the link to the RealNVP paper and here is a nice, short explanation of it (the paper is pretty long)...
1
vote
1answer
118 views

What is the difference between batches in deep Q learning and supervised learning?

How is the batch loss calculated in both DQNs and simple classifiers? From what I understood, in a classifier, a common method is that you sample a mini-batch, calculate the loss for every example, ...
1
vote
1answer
59 views

Why does PyTorch use a different formula for the cross-entropy?

In my understanding, the formula to calculate the cross-entropy is $$ H(p,q) = - \sum p_i \log(q_i) $$ But in PyTorch nn.CrossEntropyLoss is calculated using this ...
1
vote
0answers
31 views

Deduce properties of the loss functions from the training loss curves

I have two convex, smooth loss functions to minimise. During the training (a very simple model) using batch SGD (with tuned optimal learning rate for each loss function), I observe that the (log) loss ...
1
vote
0answers
42 views

Using U-NET for image semantic segmentation

I'm getting literally crazy trying to understand how U-NET works. Maybe it is very easy, but I'm stuck (and I have a terrible headache). So, I need your help. I'm going to segment MRI to find white ...
1
vote
0answers
45 views

A generalized quadratic loss and Newton iteration for Support Vector Regression, why doesn't it generalize well?

I'm comparing the results of an Newton optimizer for a modified version of SVM ( a generalized quadratic loss, similar to the one stated in: A generalized quadratic loss for SVM ) with classic SVM^...
1
vote
0answers
25 views

How would the “best function” been constructed if there are no computationally limitations?

I am reading the Wikipedia article on gradient boosting. There is written: Unfortunately, choosing the best function $h$ at each step for an arbitrary loss function $L$ is a computationally ...
1
vote
0answers
297 views

CNN clasification model loss stuck at same value

I have CNN model to classify 2 classes. (Yes or No) I use categorical_crossentropy loss and softmax activation at the end. For input I use image with all 3 channels, for output I use One hot encoded ...
1
vote
0answers
113 views

How to interpret a large variance of the loss function?

How do I interpret a large variance of a loss function? I am currently training a transformer network (using the software, but not the model from GPT-2) from scratch and my loss function looks like ...
1
vote
0answers
44 views

What are the loss functions used in teacher-student learning models?

I am not sure what are the common loss functions people usually use when training a student in a teacher-student learning model. Any insight on this is appreciated.
1
vote
1answer
90 views

LSTM text classifier shows unexpected cyclical pattern in loss

I'm training a text classifier in PyTorch and I'm experiencing an unexplainable cyclical pattern in the loss curve. The loss drops drastically at the beginning of each epoch and then starts rising ...
1
vote
0answers
174 views

Could the Jensen-Shannon divergence and Kullback-Leibler divergence be used as loss functions of non-generation problems?

If I understand correctly, the KL divergence is a measure of information loss between a ground truth distribution $P$ and a predicted distribution $Q$, and the Jensen-Shannon divergence is the mean of ...
1
vote
0answers
17 views

Limits for a bottleneck

I have some 64x64 pixels frames from a (simulated) video, with a spaceship moving on a fixed background. The spaceship moves in a straight line with constant velocity from left to right (along the x-...
1
vote
0answers
32 views

Why such a big difference in number between training error and validation error?

Question Why such a big difference between my 'Train loss' and 'Validation loss' as shown in the picture below? Is it a signal that my codes are wrong and my trained network is wrong as well? Some ...
1
vote
0answers
65 views

Unit integral condition on the output layer

I want to train a neural network on some input data from a probability distribution (say a Gaussian). The loss function would normally be $-\sum\log(f(x_i))$, where the sum is over the whole data (or ...
1
vote
0answers
35 views

Add a layer derivative in the loss function

I am writing a NN in pytorch and I want to add the derivative of the output with respect to one of the inner layers in the loss. Here is a simple example of what I mean: ...
0
votes
1answer
69 views

When calculating the cost in deep Q-learning, do we use both the input and target states?

I just finished Andrew Ngs's deep learning specialization, but RL was not covered, so I don't know the basics of RL. So, I have been having trouble understanding the cost function in deep Q-learning. ...
0
votes
1answer
74 views

Why is the derivative of the softmax layer shaped differently than the derivative of other neurons?

If the derivative is supposed to give the rate of change of a function at that point, then why is the derivative of the softmax layer (a vector) the Jacobian matrix, which has a different shape than ...