The Stack Overflow podcast is back! Listen to an interview with our new CEO.
Data Science

Questions tagged [autoencoder]

Ask Question

Autoencoders are a type of neural network that learns a useful encoding for data.

180 questions
Newest
Active
Bountied
Unanswered
Filter by
Sorted by
Tagged with
0
votes
0answers
2 views

Removing layers from a convolutional encoding-decoding network

I've been reading the following paper on Style-transfer (Universal Style Transfer via Feature Transforms): https://arxiv.org/pdf/1705.08086.pdf A crucial part of the algorithm (Section 3.3) uses a ...
0
votes
1answer
12 views

Validation generator in Autoencoder returning NaN

I am trying to build a fairly simple autoencoder using Keras on the OpenImages dataset. Here is the architecture of the ae: ...
0
votes
1answer
20 views

Keras - Autoencoder different from Encoder + Decoder

I build a CNN 1d Autoencoder in Keras, following the advice in this SO question, where Encoder and Decoder are separated. My goal is to re-use the decoder, once the Autoencoder has been trained. The ...
0
votes
0answers
9 views

VAE latent space dimension with Weibull PDFs as input data

I am currently using a Variational Autoencoder to reconstruct PDFs of Weibull distributions with varying parameters (sampled uniformly from a given parameter span). The PDFs are generated by sampling ...
0
votes
1answer
21 views

why do we operate with graphical models in VAE, if there are no probabilites involved?

In the variational autoencoder, I often see graphical models e.g. $P(X|Z)$ for the decoder, but when I looked at code, I don't see any random variables, I see just deterministic network, with ...
0
votes
1answer
10 views

how to design sigma(std) vector to have values above > 0 in VAE autoencoder

I am trying to understand how to design mu and sigma vector in VAE autoencoder, as we know std value can not be larger than 0. If I use linear function, the value can be lower than zero, so I think ...
2
votes
1answer
49 views

Autoencoders for the compression of time series

I am trying to use autoencoder (simple, convolutional, LSTM) to compress time series. Here are the models I tried. Simple autoencoder: ...
0
votes
0answers
11 views

How can we use resnet in autoencoders?

I am creating an unsupervised classifier model, for which i want to use resnet 50 on a custom database and used the top layers of resnet as start point of my autoencoder. How to proceed with this. ...
0
votes
0answers
12 views

How to measure performance of a denoising autoencoder?

Is there a way to theoretically estimate (avoiding brute force) what is the minimum signal-to-noise ratio required for a denoising autoencoder (like a regular filter)? What I want to know specifically ...
0
votes
0answers
15 views

Combining multiple loss functions in the right way

I am trying to build an (variational) Autoencoder to generate fake but representative data from a generic data set with a couple of numeric and categorical columns. So far I have built functions that ...
0
votes
1answer
26 views

Autoencoder or layer-based dimensionality reduction?

I have a few TB of wide data. I want to reduce the number of features in my dataset before feeding my dataset into a classification model... or should I not? Obviously, I will want to try both ...
0
votes
1answer
12 views

How to design n-dimensional feature descriptor similar as the input image?

I am re-writing the H-Net code in Keras for cross-domain image similarity. The network architecture is described in the attached paper. I wrote the encoder and decoder parts but unable to get similar ...
0
votes
1answer
99 views

Autoencoder: using cosine distance as loss function

I'm trying to train an autoencoder (in PyTorch) to reconstruct gene profiles. At the moment I'm using the Mean Squared Error (MSE) loss for training: the model is not overfitting and both the training ...
1
vote
0answers
11 views

Autoencoders linear latent space

According to "Linear interpolation in latent space" in https://hackernoon.com/latent-space-visualization-deep-learning-bits-2-bd09a46920df and others, the latent space representation of an autoencoder ...
1
vote
0answers
27 views

Disadvantages of Mean Squared Error? [closed]

I'm using mean squared error as reconstruction error for my autoencoder. The dataset is ECG (time series) and model is conv1d. I assumed MSE as the best option for reconstruction error, but it's ...
8
votes
1answer
647 views

How does an encoder-decoder network work?

Let's say I trained an encoder-decoder network on a cat dataset using reconstruction error as loss function. The network is fully trained and the decoder is able to reconstruct good cat images. Now ...
0
votes
1answer
45 views

Fully Convolutional Variational Autoencoder

I would like to make a neural network which uses black and white images as input and outputs a colored version of it. The important thing in that process is that the size of the images must stay the ...
0
votes
0answers
23 views

What is the meaning of constraint two neural networks against each other?

I have two identical autoencoders with the same number of layers and parameters. I would like to find the similarity between two images from different domains such as an image captured from the camera ...
0
votes
1answer
64 views

Which algorithm can be used to reduce dimension of multiple time series?

In my dataset, a data point is essentially a Time series of 6 feature over a year per month so in all, it results in 6*12=72 features. I need to find class outliers so I perform dimensionality ...
0
votes
1answer
21 views

Comparison of performance of autoencoder with PCA

I am running PCA and autoencoder (2 hidden layer with relu) on a data. Both PCA and autoencoder give similar accuracy of the order 50%. I have tried different variations of autoencoder: changing ...
1
vote
0answers
8 views

How to calculate a location prior in CT's based on atlas segmentation?

I'm working myself through the paper "Anatomical Priors in Convolutional Networks for Unsupervised Biomedical Segmentation" (https://arxiv.org/abs/1903.03148) I lack to understand how the atlas ...
1
vote
0answers
16 views

How can one be assured that generative models are not memorizing dataset, and that they will generate an unique image outside of dataset?

If all GAN can do is capture the probability distribution of the dataset, then shouldn't they be similar to handing out images from the dataset? How can we verify that the images that they generate ...
0
votes
0answers
8 views

Post Training Sampling from a unit Gaussian?

During the training of a VAE, we sample Z from N(mu, sigma) ( = Q(Z|X) ) and then feed it to the decoder for reconstruction. Why do we, then, feed Z ~ N(0,1) to the decoder after the training is ...
0
votes
0answers
38 views

variational autoencoder doesn't fit with KL loss Keras

Hi I am learning the variational autoencoders and Idesigned my VAE as follow: ...
0
votes
0answers
23 views

Developing an encoder/decoder for image modification

On the project I am currently working on, my goal is to train a neural network to convert images of circles to ellipses in a way that models convolution/blurring in real imaging processes. What ...
1
vote
0answers
61 views

Optimizing parameters for CNN autoencoder based on training and validation loss

I have designed an autoencoder with a encoder and decoder consiting of 2D convolutational layers (the input are 40'000 2D images). I train the autoencoder using adam optimizer. The autoencoders has ...
0
votes
0answers
69 views

How do we overfit a CNN AutoEncoder for anomaly detection?

I have been working on an anomaly detection problem in which I need to treat the images of "street" as an anomaly. The images of "glaciers" will be treated as not-anomaly. The autoencoder which ...
0
votes
0answers
22 views

Keras / TensorFlow 2.0: are UpSampling2D() layers the inverse of Max-Pooling?

I am trying to build a Variational Autoencoder for image data. As I employ MaxPool2D() in the encoding part, I need the reverse it in the decoder. Are ...
0
votes
0answers
10 views

Is my Autoencoder architecture correct?

I am dipping my toe in deep learning, especially autoencoders. I am trying to reconstruct my dataset. The dimension of my dataset is 13000x3. Can the following architecture work? ...
0
votes
0answers
39 views

Using smote for data augmentation of a data set which has no dependent variable

I am trying to use the reconstruction error obtained using an auto encoder to do novelty detection. My data set is of size (4500,55)(Note: this data doesn't have any abnormalities.When an auto-encoder ...
1
vote
1answer
115 views

How do I get similarity with autoencoders

I have build an autoencoder to extract from a very high dimensional (200 dimensions) space a smaller but significant representation (16 dimensions). Now that I have these "encoded" vectors, I would ...
0
votes
0answers
50 views

Replication of Andrew Ng's Sparse Autoencoder

for the past three days I have been trying to replicate the results presented in Andrew Ng's sparse autoencoding lecture (https://web.stanford.edu/class/cs294a/sparseAutoencoder.pdf) however I have ...
0
votes
1answer
41 views

Autoencoder feature extraction without validation set?

I plan to use autoencoder for feature extraction, then use the latent vector for clustering. My autoencoder performs very very well on my training set (loss small and reconstructed image look very ...
0
votes
2answers
88 views

Filters in convolutional autoencoders

I have a question regarding the number of filters in a convolutional Autoencoder. As far as I have understood, as the network gets deeper, the amount of filters in the ...
1
vote
0answers
59 views

Can autoencoders take time series into account?

Here, I read the following: The first key to understanding is that HTM relies on data that streams over time (...) By contrast, conventional deep learning uses static data and is therefore time ...
0
votes
1answer
30 views

Does it make sense to train an Autoencoder for Dimensionality Reduction using Mini-Batch Gradient Descent?

I want to reduce the dimensionality of a dataset using a stacked Autoencoder. The size of the dataset and the computing power at my disposal make it very difficult to train the Network using simple, ...
0
votes
0answers
15 views

Design / Choice of Autoencoder to classify temporal pattern in images

Suppose I have a temporal stack of images of shape $m \times n \times k$ where shape of each image is $m \times n$ and $k$ represents the temporal dimension. In this context, I am trying to detect and ...
0
votes
1answer
42 views

Resize instead of transposed convolutions

I'm trying to build a decoder version of ResNet, i.e. one that goes from the prelogits layer and attempts to recreate the image. I can get it working by using transposed convolutions, but the quality ...
0
votes
0answers
17 views

Autoencoder not overfitting data after large number of epochs and small number of samples

I am training a deep autoeocoder on numerical data, with python jupyter notebook. I have 17 samples, each with 534 values, and my auto encoder has all layers to 534, but even after 5,000 epochs, the ...
0
votes
0answers
10 views

Application of Varitiational Auto Encoders in improving the generalizability of classifiers when there is limited amount of data?

Is there any prior work on the above topic? Currently, I am working on Domain Adversarial Neural Networks and merging it with VAE to improve generalizability and transfer learning. I could not find ...
0
votes
0answers
13 views

How to decode text output in autoencoders?

I have made an autoencoder for text based input, and fitted it to the data. Now I want to see the output text. Is there any way to decode the numbers to text? ...
0
votes
1answer
31 views

String indices must be integers

I was trying to encode the string values of the feature 'ProductCategory' into integer values but I got this error. Kindly help. And I would also like to ask if label-encoding this feature would not ...
1
vote
1answer
129 views

How to calculate compression ratio when using autoencoder in neural network

For example, if I use an autoencoder to compress a 1000 dimensional data set to 25 dimensions. Is the compression ratio is 40:1? Other info: The dataset contains 5000 samples. 2 million parameters ...
3
votes
1answer
174 views

What is the difference between an autoencoder and an encoder-decoder?

I want to know if there is a difference between an autoencoder and an encoder-decoder.
1
vote
2answers
85 views

Why do we use a softmax activation function in Convolutional Autoencoders?

I have been working on an image segmentation project where I have created a convolutional autoencoder. I saw this image and implemented it using Keras. At the output layer, the author has used the ...
0
votes
0answers
16 views

LSTM -Detect single point in time series

I know that LSTMs can learn dependencies for many variables across many timestamps easily. I have used LSTMs to forecast the time series. Now my Aim is to identify when the process is getting ...
2
votes
1answer
47 views

Intractability in Variational Autoencoders

I'm having difficulty understanding when integrals are intractable in variational inference problems. In a variational autoencoder with observation $x$ and latent variable $z$ we want to maximize ...
0
votes
0answers
11 views

What machine learning model to predict a sequence of instructions from a matrix

So my task involves taking an input of variable sizes convolutional filters and predicting the sequence of instructions for a processor to implement that filter. Now the sequence can be of variable ...
3
votes
2answers
84 views

How to scale outputs from AutoEncoder from multiple models?

I have a problem for which I have not been able to find any answers in my search so far. BACKGROUND I am working on an anomaly detection problem on machines utilising an auto-encoder. I am building ...
0
votes
1answer
123 views

Autoencoder gets ~0% accuracy / doesn't train at all

So I wanted to get into the topic of 'Autoencoder', and just tested how well it would work on random vectors of size 200. ...

1 2 3 4