Questions tagged [training]

For questions about training networks, rules systems, or other AI system components.

Filter by
Sorted by
Tagged with
4
votes
1answer
99 views

Do we use validation and test sets for training a reinforcement learning agent?

I am pretty new to reinforcement learning and was working with some code for the PPO and DQN algorithms. After looking at the code, I noticed that the authors did not include any code to setup a ...
3
votes
1answer
46 views

Can I apply reparametrization trick on "any" deep neural network?

I came across the "reparametrization trick" for the first time in the following paragraph from the chapter named Vector Calculus from the test book titled Mathematics for Machine Learning ...
2
votes
0answers
25 views

How to fix high variance of the returns on a 2d env?

I'm trying to train an agent on a self-written 2d env, and it just doesn't converge to the solution. It is basically a 2d game where you have to move a small circle around the screen and try to avoid ...
0
votes
0answers
17 views

How to Tensorflow models and YOLO differ in terms of training steps?

Can anybody explain how the training steps work for the Tensorflow Object Detection algorithms available on the Tensorflow 2 Detection Model Zoo? For instance, YOLOv5 cycles through epochs. As I ...
2
votes
0answers
9 views

References for the convergence of gradient-based algorithms for training neural networks

I'm looking for some good references that give convergence results of training neural networks. I'm decently familiar with works that analyze the convergence of SGD, and, in particular, I really like ...
6
votes
1answer
119 views

How to decide a train-test split?

In almost every ML model, a train-test (or train-test-val split) is critical to assess the model's performance. However, I have always wondered what the rationale is to decide a particular train-test ...
2
votes
0answers
22 views

Should I train a neural network with data with or without a constraint?

I want to train a Neural Network (NN) using a dataset. I want to use the NN model as a prediction function in one algorithm. However, in the algorithm, any data that does not meet a specific ...
1
vote
1answer
37 views

Does the ANN's training data include the proper output for every neuron?

I was designing an Artificial Neural Network a while back, but hit a bump when I got to the backpropagation. I was having trouble making the script choose whether to add or subtract from the weights, ...
1
vote
0answers
12 views

What is the difference between supervised and unsupervised training in T5?

I know unsupervised training for T5 is like: input: He went X output: X to school Z is this equivalent to the following in a supervised manner: ...
-1
votes
1answer
26 views

What practically makes a good architecture of ANN?

When we take a look at the literature there are so many opinions. I was wondering what are some generally good practices to design an architecture, like how much depth would you prefer and how much ...
0
votes
0answers
20 views

How to assess the goodness of a text generation algorithm

Take a RNN network fed with Shakespeare and generating Shakespeare-like text. Once a model seems mathematically fine, as can be assessed by observing its loss and accuracy over training epochs, how ...
9
votes
3answers
747 views

Is it okay to use publicly available Instagram videos to train an AI?

Since I haven't found any good training data for my university project, I want to use pictures and videos from public Instagram profiles. Am I allowed to do that?
3
votes
1answer
66 views

What problem does the neural network really solve?

In the image below taken from a Youtube video, the author explains that the neural network can be used to fit a relational graph for a set of data points shown by the green line. And that this is ...
1
vote
0answers
27 views

What's the benefit for using a Kalman filter for training a neural network compared to other optimization algorithms?

I found a paper about using an Unscented Kalman Filter(UKF) for traning a neural network. The UKF filter is modified so it works for parameter estimation. Assume that we have a neural network model $\...
0
votes
1answer
22 views

Validation accuracy less than training accuracy (with no sigh of overtraining)

I am working with a deep CNN with over 100k sample data. I divided it up into 75% training, 12.5% validation and 12.5% for testing. As I train my network, the training accuracy approaches near 100% ...
1
vote
1answer
84 views

Tensorflow object detection model total loss starts out good, but suddenly explodes up to high loss numbers

I'm training a Tensorflow object detection model with approx. 7500 images of two classes, which contains approx. 10,000 classes per class. I'm using Tensorflow 2.6.0, in case that is relavent. I am ...
0
votes
0answers
30 views

How big should the dataset for retraining ssd_mobilenet_v2 be?

I have retrained ssd_mobilenet_v2 using my own dataset with 2 classes (pen or pencil), using object detection API. For my project, I expect users to select specific pencils from all pencils and ...
1
vote
1answer
26 views

Using Human Confirmation in place of a loss Function for Training

Has there been any experimentation in designing an AI to prompt a human to judge the accuracy of it's outcomes? instead of using a loss function, a human can judge the accuracy of it's estimation ...
0
votes
0answers
22 views

Adding data to training results in loss random peaks

I have succesfully trained ssd_mobilenet_v2_keras for object detection, with a dataset of about 3700 images. Now I have more images to add. I tried adding only a few images (150-300) to see what ...
2
votes
1answer
99 views

Is there any way to train a neural network without using gradients?

The only algorithm I know for updation of weights of a neural network is based on gradients. The update equation can be roughly written as $$w \leftarrow w - \nabla_{w}L$$ where $\nabla_{w}L$ is the ...
0
votes
1answer
36 views

How to re-training an AI model to have smaller input image size

I need a PyTorch Model which can do road segmentation on OAK-D camera. The model provided requires Input Image Size: 896*512, which is too big for running on OAK-D camera. Thus I need to re-training ...
0
votes
0answers
12 views

What is meant by non-convergent limit cycles?

Limit cycle is a closed curve that is isolated i.e., no other closed curve near to it. You can read the following paragraph from here If there is (such) a closed curve, the nearby trajectories must ...
2
votes
0answers
40 views

Is stability an attribute of model or training algorithm used or combination of both?

From this answer, stability is attributed to a learning algorithm A stable learning algorithm is one for which the prediction does not change much when the training data is modified slightly. At ...
0
votes
0answers
13 views

Weight for Samples on SVM (Support Vector Machine)

There is a option sample_weight in fit(X[, y, sample_weight]) function (OneClassSVM, sklearn library). If I use the option ...
2
votes
1answer
49 views

What does it mean by strong or sufficient gradient for training in this context?

It has been mentioned in the research paper titled Generative Adversarial Nets that generator need to maximize the function $\log D(G(z))$ instead of minimizing $\log(1 −D(G(z)))$ since the former ...
0
votes
1answer
44 views

Does average loss function in GAN training is just an approximation of value function and does not ensure convergence of generator and discriminator?

The value function on which convergence has been proved by the original paper of GAN is $$\min_G \max_DV(D, G) = \mathbb{E}_{x ∼ P_{data}}[\log D(x)] + \mathbb{E}_{z ∼ p_z}[log (1 - D(G(z)))]$$ and ...
0
votes
0answers
32 views

Why is the proof of convergence in the GAN paper not applicable practically?

This question is about generative adversarial networks and restricted to the research paper titled Generative Adversarial Nets. If I select a particular architecture of MLP as a generator and trained ...
0
votes
0answers
23 views

How to handle critical points during generator training?

Using an MLP as a generator introduces multiple critical points in parameter space. You can read this excerpt from the research paper titled Generative Adversarial Nets by Ian J. Goodfellow et al. In ...
1
vote
0answers
22 views

Training on the dataset in parts vs training on the whole dataset

What is the difference between these two situations? are they the same ? #1 : train a model 20 epochs on the whole dataset #2 : divide dataset into n-parts then train the model 20 epochs on each part ...
0
votes
1answer
35 views

How to calculate the gradient penalty proposed in "Improved Training of Wasserstein GANs"?

The research paper titled Improved Training of Wasserstein GANs proposed a gradient penalty in order to avoid undesired behavior due to weight clipping of the discriminator. We now propose an ...
0
votes
1answer
107 views

Mapping input vectors of variable length to output vectors of variable lengths with dummy variables

I have a general question about supervised ANNs that map inputs to outputs. It is possible to vary the length of the input and output vectors by inserting some dummy variables that will not be ...
1
vote
2answers
292 views

What is Lipschitz constraint and why it is enforced on discriminator?

The following is the abstract for the research paper titled Improved Training of Wasserstein GANs Generative Adversarial Networks (GANs) are powerful generative models, but suffer from training ...
1
vote
1answer
44 views

Can the optimal learning rate differ for different architectures?

In several courses and tutorials about neural networks, people often say that the learning rate (LR) should be the first hyper-parameter to be tuned before we tweak the others. For example, in this ...
0
votes
0answers
16 views

Feeding the output back to input in 3D CNN model

I am currently designing a Model which takes Input 3D Grid and Model Output at $t-1$. The model figure is described below I have two thoughts in training the model for above situation. Feed output $...
1
vote
0answers
20 views

Is there any existing mechanism that allows us to pass input from randomly selected layers of neural network per iteration?

Consider the following neural network with $\ell$ layers. $$i_0 \rightarrow h_1 \rightarrow h_2 \rightarrow h_3 \cdots \rightarrow h_{\ell-1} \rightarrow o_{\ell} ,$$ where $i, h, o$ stands for ...
1
vote
2answers
47 views

Train Validation Test Splitting After or Before Data Augmentation?

I have seen tutorials online saying that you should do data augmentation AFTER doing the train/val/test split. However, when I go online to read some research papers, I see numerous instances of ...
1
vote
1answer
42 views

What does it mean by "zeros the networks parameters gradients" in the context of training a neural network?

Consider the following PyTorch code ...
0
votes
0answers
4 views

How to estimate the number of training samples to train a HMM Model with Baum-Welch Algorithm

Is there a thumb rule that tells us how many data sample a HMM model needs to be trained by Baum-Welch Algorithm?
0
votes
0answers
29 views

Model not learning anything, what can be the problem?

I've trained a model for heart sound classification with transfer learning (MobileNet) on Physionet dataset, and it works fine. However, when I train it on my own dataset, it seems that it can not ...
0
votes
0answers
116 views

How to validate a trained HMM Model?

What are some approaches to validate a HMM Model ? I want to check if a trained model is good enough to put it into operation or if it does need more training. I am using the Baum Welch algorithm to ...
1
vote
0answers
26 views

Is the main difference between the logistic regression and the perceptron the activation function they use?

I went through a Stats StackExchange's post about the difference between logistic regression and perceptron, which is too long to get the key point. I'd like to consider the question in terms of the ...
2
votes
1answer
1k views

Should I continue training if the neural network attains 100% training accuracy?

I have a neural network where there are two hidden layers. Each hidden layer has 128 neurons. The input layer has 20 inputs, and the output layer has 3 outputs. I have 1 million records of data. 80% ...
3
votes
1answer
58 views

Can some of the weights be fixed during the training of a neural network?

Is it possible to exclude specific layers from the optimization? For example, let's say I have an input layer, 2 hidden layers, and the output layer. I know there is a perfect solution for my problem ...
2
votes
0answers
40 views

Identifying rotating and resizing letters with background noise

I'm trying to complete a captcha, and here is what it looks like: Between captchas the calligraphy of the letters is the same, but the letters may be resized and rotated. And the background noise (...
0
votes
0answers
16 views

Train a model for different scenario and gather performance results in a single place

Recently, I extend a master's thesis. I am now in a training phase for the model associated with it. I have access to many node GPUs. I would like to train this model on different scenarios, e.g. ...
0
votes
1answer
80 views

Is it possible to overfit a model on infinite amounts of data?

This is a theoretical question. Is it possible to overfit a model on infinite amounts of data? Let me clarify there are no duplicates. Say, we have a generator function that produces data, with the ...
0
votes
0answers
28 views

Problems while transforming a 2D Variational Autoencoder into a 1D Version

I am trying to addapt the Keras variational autoencoder (VAE) here from a 2-D input/output (matrix of a picture) to a 1-D input/output (just a vector). I thought this would be a fearly easy task, but ...
2
votes
1answer
50 views

What is meant by "stable training" of a deep learning model?

I have read it said that the "stable training" of a deep learning model is important. What is meant by "stable training" of a deep learning model?
0
votes
0answers
12 views

How exactly is masking performed in the training part of the paper "Semi-Supervised Classification with Graph Convolutional Networks"?

I am struggling to understand the training part of the paper Semi-Supervised Classification with Graph Convolutional Networks (2017) by Thomas Kipf and Max Welling. The GitHub repo is here. I do not ...
1
vote
0answers
21 views

Necessity of likelihood in training energy-based models

Lately, I've been getting into energy-based models (EBMs) through some of Yann LeCun's recent talks, where he advocates the use of non-normalized models because it allows for more flexibility in the ...

1
2 3 4 5
8