Questions tagged [autoencoder]

Autoencoders are a type of neural network that learns a useful encoding for data in an unsupervised manner.

Filter by
Sorted by
Tagged with
3
votes
0answers
24 views

Why KL Divergence instead of Cross-entropy in VAE

I understand how KL divergence provides us with a measure of how one probability distribution is different from a second, reference probability distribution. But why are they particularly used (...
1
vote
1answer
18 views

Is window based sequencing a good idea to obtain more training data for LSTMs?

I am trying to do an unsupervised autoencoder based outlier detection for time series using LSTMs. Here, there are multiple time series, and an entire series is to be considered as an outlier. However,...
0
votes
0answers
24 views

Varitional Autoencoder not accepting batch size or validation data

The input to the VAE will be a customer vector where the index of the vector represents a product id, position i in vector x is set to one iff product id i has been purchased by the customer. For ...
0
votes
0answers
6 views

What affects blurriness of images outputted by vanilla VAE?

I know that MSE between reconstruction (m(z)) and original image (X) affects blurriness, because it averages across pixel values. I also know that variance of the Gaussian distribution p(x|z) (p(x|z) =...
0
votes
0answers
21 views

Using autoencoder for time series prediction

I was recently reading a paper on time series prediction using deep learning methods. There I found a technique named "Variational Autoencoder" to predict time series data. I understand how ...
1
vote
1answer
24 views

Training a Variational Autoencoder (VAE) for Random Number Generation

I have a complicated 20-dimensional multi-modal distribution and consider training a VAE to learn an approximation of it using 2000 samples. But particularly, with the aim to subsequently generate ...
0
votes
0answers
14 views

Training a VAE for random number generation

I have a high-dimensional multi-modal distribution of random numbers in R^n and consider training a VAE to learn the distribution. What I want to do with it succeedingly, is to sample from the latent ...
0
votes
0answers
18 views

Autoencoder with specific features in encoded layer

I want to create an autoencoder which will map date + 24 hourly data for example 25/08/2020 - [25,8,2020,23,3,11,31,43,43,23,23,45,23,11,32,43,52,34,23,12,43,52,32,54,23,54,34] to encoded layer of 3 ...
0
votes
0answers
26 views

Can autoencoder latent variables to be used as features for classification?

I did some experiments on convolutional autoencoder by increasing the size of latent variables from 64 to 128. I used 4 covolutional layers for the encoder and 4 transposed convolutional layers as the ...
0
votes
0answers
15 views

Variational Autoencoder (VAE) latent features

I'm new to DL and I'm working on VAE for biomedical images. I need to extract relevant features from ct scan. So I created first an autoencoder and after a VAE. My doubt is that I don't know from ...
1
vote
0answers
18 views

Autoencoder fails to reconstruct

I'm trying to use an autoencoder to reduce dimensionality of my features. My features are of dimension 2048. I tried to train an autoencoder to reduce the dimensionality to 50. I'm using a single ...
0
votes
0answers
9 views

How to use an encoder to do feature extraction

I'm newbie with all of data science. I have a pre-trained U-Net network from which I get its encoder. Now I have to use a picture to get its features. With the whole U-Net I do this with fit method: <...
0
votes
1answer
16 views

Up to which layer can we consider the encoder to be?

I'm trying to extract the encoder from a U-Net network. Given its architecture: And its summary: ...
0
votes
0answers
13 views

Perceptual Loss for 3D VAE

I'm working on 3D VAE implementation for biomedical images. The results are too blurring, so I'm searching to improve the performance of the network. Many people recommend the use of 'perceptual loss',...
7
votes
3answers
1k views

What is an autoencoder?

I am a student and I am studying machine learning. I am focusing on deep generative models, and in particular to autoencoders and variational autoencoders (VAE). I am trying to understand the concept, ...
0
votes
0answers
40 views

MSE loss in VAE reduces only KL divergence

I am trying to reproduce the two stage VAE from this repository in TF2 with Keras to learn MNIST digits. Unfortunately I am experiencing behavior that I can not explain to myself: As far as I ...
1
vote
1answer
22 views

Autoencoder feature extraction plateau

I am working with a large dataset (approximately 55K observations x 11K features) and trying to perform dimensionality reduction to about 150 features. So far, I tried PCA, LDA, and autoencoder. The ...
0
votes
0answers
22 views

One-Hot Encoded Matrix Inupt/Ouput for Autoencoder

I am trying to write an autoencoder to reduce the dimensionality of my genomic data. Currently, my data is in the form of a $273278 \times 1$ vector. Each element of the vector indicates whether a ...
0
votes
1answer
23 views

What do the symbols in this image mean in relation to autoencoding? (which concepts do they represent and why?)

I'm completely new to neural nets, and am trying to get a grasp of autoencoding. I understand the function of the decoder and encoder, but what does the arg min section and beyond mean? Why is it ...
0
votes
0answers
29 views

Sequence-to-Sequence Autoencoder with asymmetric length of encoding sequence

I am trying to design a sequence-to-sequence autoencoder where the encoding sequence has a length shorter than the sequence itself given that the encoding's dimensionality could be more to compensate ...
0
votes
0answers
33 views

Data augmentation for recommendation systems

I have a user-item matrix that I use to train a denoising autoencoder to predict the top-k items to recommend to the different users. The idea is to corrupt the matrix by erasing a percentage ...
0
votes
0answers
8 views

Looking for the proper algorithm to compress many lowres images of nearby locations

I have an optimization problem that I'm looking for the right algorithm to solve. What I have: A large set of low-res 360 images that were taken on a regular grid within a certain area. each of these ...
0
votes
0answers
6 views

Using AutoEncoder for Greedy Layer-Wise Pretraining [Convolutional Neural Networks]

I am trying to implement greedy layer-wise pretraining for Convolutional Neural Network binary classifier using AutoEncoders. However, I am a little bit confused regarding the logic of implementation. ...
0
votes
1answer
10 views

How can I easily retrieve the latent space encodings in tensorflow?

Here is where I'm at. I built and trained an autoencoder in tensorflow. The model summary looks like so: ...
0
votes
0answers
22 views

How to use AutoEncoders or VAE for initializing custom CNN architecture?

I would like to use AutoEncoder or VAE in order to learn set of features which I can use to initialize training procedure of a custom CNN architecture that I've build. Here is the code: ...
0
votes
0answers
25 views

Autoencoder works but VAE does not - KL divergence loss drops to zero

I have an image encoding problem, to which I have successfully applied an autoencoder. But when I try to convert this into a VAE framework by introducing KL divergence loss and sampling from a unit ...
1
vote
1answer
35 views

Why autoencoders work well for outlier detection?

Apart from the fact that they are neural networks, which usually is a reason for outperforming other algorithms, is there other reason helping auto-encoders perform well in outlier detection? I know ...
1
vote
0answers
12 views

Using Iterative Hard/Soft Thresholding in autoencoder with non linear activation

Can someone please give an intuitive explanation of the difference between the Iterative Hard Thresholding VS Iterative Soft thresholding algorithm? And if we can use these algorithms in an ...
0
votes
0answers
12 views

Can the input and output be different in a variational autoencoder?

While the autoencoder and the variational autoencoder typically decode back to the same input space that was passed in, I am working on a more generic encoder-decoder structure, where the output is a ...
2
votes
3answers
57 views

What is the best way for synthetic data generation while maintaining privacy?

For one of the projects where we are working as third party contractors, we need a way for the company to share some datasets which can be used for data science. It is not possible for the company to ...
0
votes
1answer
27 views

Cross-Validation in Anomaly Detection with Labelled Data

I am working on a project where I train anomaly detection algorithms Isolation Forest and Auto-Encoder. My data is labelled so I have the ground truth but the nature of the problem requires ...
2
votes
2answers
24 views

Is it possible to compress a sequence of numbers through an autoencoder?

Specifically: I would like compress a set of coordinates, which map to the locations of 1's in a binary image, and then decode back to the original set. For instance, for a 16x16 image, the input ...
2
votes
1answer
30 views

How to choose the good number dimension of autoencoder?

I'm using Autoencoder for feature extracting. I stuck with how to choose good number of dimension of encoder layer (latent layer). After training dataset, the model gave the latent layer (embedding ...
0
votes
0answers
44 views

Anomaly Detection over multivariate data containing Nominal and numerical predictors

I am trying to implement Anomaly Detection over a multivariate dataset having nominal and numerical predictors. Dataset has following pattern: If we consider the below sample records, category_id, ...
-1
votes
1answer
27 views

Does Anomaly Detection Algorithm works when the features are not correlated?

I am working on an Anomaly Detection Problem and the algorithm I used is an Autoencoder Multivariate Gaussian. The problem with my data is that it is unlabeled and not correlated. For example, let's ...
0
votes
2answers
26 views

MNIST data shape

In going through the different tutorials on CNN, auto encoders and so on I trained my self on the MNIST problem. The different images are stored in a 3D array which shape is (60000,28,28). In some ...
0
votes
0answers
24 views

How to pass noisy images as input and original images as labels in Keras - Autoencoders

I want to make denoising autoencoder, I've added some noise to images, and i want to use them for training, original images will be used as label. Reading the keras documentation I've found that I ...
0
votes
0answers
13 views

What's the best approach to classify labeled stockchart images , based on their curve

I'm working on an exercise where I have selected a lot of different stock chart images, showing the week closing of around 1200 companies at two different periode in time. I have stock charts from ...
0
votes
0answers
10 views

Information compression for variable input size

Is there a way to compress information of a variable input size? Autoencoder requires standardized input sizes. Although I can add masks on the cost function and add dummy features to standardize ...
0
votes
0answers
33 views

Autoencoder neural network for dimensionality reduction of many-to-many regression problem

I want to find the relationship between a continues input-output using neural networks. The input consists of time series with more than 1000 data points. The output of each input time series is the ...
0
votes
0answers
19 views

questions on multi-step air pollution prediction

I am trying to use RNN to predict the concentration of various air pollutants for the next 24 hours. The input data consists of 72 hours long and every hour owns 14 elements such as temperature, ...
0
votes
1answer
14 views

How to verify a CNN encoder works as expected?

I am using CNN as a part of kernel warping. The purpose here is to reduce input dimension (from N*M to K *1). The input data is not image data. I suspected that the CNN network might not work as I ...
1
vote
2answers
36 views

Identify the parameter causing the anomaly in a multivariate dataset

I have a payment transaction dataset with a large number of predictor variables. I am trying to build a model for anomaly detection and I have evaluated various algorithms/approaches for the same like ...
0
votes
0answers
28 views

Can I used a pre-trained auto encoder as an embedding layer in the model?

The dataset has an ordinal target variable [0,1,2]. Each observation in the dataset has multiple time series. In addition, each observation have various tabular features. The purpose is to build a ...
0
votes
0answers
21 views

Derivation of ELBO in “An Introduction to Variational Autoencoders”

On page 18 of the paper titled "An Introduction to Variational Autoencoders" there is the following equation at the top of page 18. Paper can be found here. $\log p_θ(\mathbf{x}) = \mathbb{E}_{q_{φ} (...
3
votes
1answer
103 views

VAE generates bad images. due to unbalanced loss functions?

I'm training a variational autoencoder on CelebA dataset using TensorFlow.keras The problem I'm facing is that the generated images are not diverse enough and look kinda bad. (new) Example: What I ...
2
votes
1answer
157 views

Transformer-based architectures for regression tasks

As far as I've seen, transformer-based architectures are always trained with classification tasks (one-hot text tokens for example). Are you aware of any architectures using attention and solving ...
0
votes
1answer
56 views

How to make custom callback in keras to generate sample image in VAE training?

I'm training a simple VAE model on 64*64 images and I would like to see the images generated after every epoch or every couple batches to see the progress. when I train the model I wait until the ...
0
votes
0answers
7 views

Repeat in CNN autoencoder output

As I try to train a CNN autoencoder, the decoder tends to generate repeated patterns. These could be noises or poorly recognized input images. The period of these repeat increases with the number of ...
0
votes
0answers
15 views

Autoencoder to encode features/categories of data

My question is regarding the use of autoencoders (in PyTorch). I have a tabular dataset with a categorical feature that has 10 different categories. Names of these categories are quite different - ...

1
2 3 4 5 6