Stack Exchange Network

Stack Exchange network consists of 174 Q&A communities including Stack Overflow, the largest, most trusted online community for developers to learn, share their knowledge, and build their careers.

Visit Stack Exchange

  7. current community

    your communities

    more stack exchange communities

    company blog
Data Science

Questions tagged [autoencoder]

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

122 questions
info newest frequent votes active unanswered
0
votes
0answers
5 views

LSTM Autoencoder on Patterns of Labels

Currently, I am trying to do anomaly detection on univariate data consisting of labels. For example: [A, A, B, C] is good but [A, A, A, A] is anomalous. I'm dealing with more than just ABC. Is an LSTM ...
0
votes
0answers
11 views

Keras Error when returning model [on hold]

I have defined the following function: ...
0
votes
0answers
15 views

train stacked denoising autoencoder in matlab [closed]

I am trying the following code in matlab to reconstruct the input data using Stacked denoising Autoencoder But it did not work. And I have received the following error"The output size [1 1 12] of the ...
2
votes
0answers
33 views

Build an Autocomplete model for document titles

I want to build an autocomplete model using RNN where input is article names (documents title). X: ['Billing', 'Loan status', 'Filling loan application', 'Contact Info', ...] The article name can ...
0
votes
0answers
19 views

Controlling how much a VAE overfits

I want to make my VAE overfit to the training sample to some degree. What is the best to way to control it? Weighting the KL divergence loss term, which basically becomes beta-VAE if I'm not wrong? ...
-1
votes
0answers
9 views

Thresholded shortest path distance

In impressing DeepMind's paper Human-level performance in first-person multiplayer games with population-based deep reinforcement learning, there is an attempt to classify and extract agents' ...
0
votes
0answers
23 views

Timedistributed Autoencoder weird loss optimization

I was using a TimeDistributed layer to wrap my autoencoder in another project of mine. This gave me weird loss optimization results, so I tested it on a simple MNIST autoencoder. The code for this is ...
0
votes
0answers
76 views

Outliner detection with LSTM autoencoder

I am learning about autoencoders for outlier detection. I have searched enough and internet suggest to use LSTM autoencoders for outlier detection from multivariant time series data. I have watched ...
0
votes
1answer
18 views

KL divergence in VAE

If I understand correctly KL-divergence is relative entropy of two distributions. To calculate KL divergence of two distributions, you would need two vectors of random variables. What I do not ...
0
votes
1answer
42 views

Training multiple keras models and combining outputs to determine losses

I'm trying to predict the future states of a 1D travelling wave (square, triangle and sawtooth) using a deep learning setup in Keras. The waves are discretised in a 1024 data points. As this gives a ...
-1
votes
0answers
14 views

VAE: Calculating reconstruction loss when the true posterior is continuous

In Python / Tensorflow, I am currently struggling with how to compute the VAE reconstruction loss $ E_{q_\phi(z|x^{(i)})} [log p_{\theta} (x^{(i)}|z)] $ (which is equivalent to the negative cross ...
0
votes
0answers
14 views

variational autoencoders

If I understand correctly, kl divergence is relative entropy, which measures destination between two distributions. in vae we want to measure distribution over latent space matrix and standard normal ...
1
vote
1answer
25 views

Training 128x128 autoencoders on 512x512 images, produces strange gridline after recombining

So I'm training an autoencoder that can recreate 128x128 images, so it can recreate any images by splitting them into 128x128 patches first, running it through the autoencoder, and having them ...
1
vote
0answers
52 views

What is the mathematical definition of the latent space [closed]

I have been seeking a mathematical explanation of the latent space of a neural network. The best I have gotten is "The latent space is the space in which the data lies in the bottleneck layer." ...
-1
votes
0answers
13 views

What are some limitations of using Collaborative Deep learning for Recommender systems?

Recently I worked on a paper by Hao Wang- Collaborative Deep learning for Recommender Systems which uses a two way tightly coupled method, Collaborative filtering for Item correlation and Stacked ...
1
vote
1answer
23 views

Do anomalous input features to autoencoder result in high errors on the corresponding output features?

An autoencoder is trained by replicating each training instance to both input and output. However, when predicting for anomaly detection, will the output error be local to the same output feature(s) ...
0
votes
0answers
10 views

Encoder Decoder networks with varying image sizes

Encoder Decoder Network - Computerphile : At the very beginning of this video, Michael Pound goes on to say: So it (encoder decoder network) makes no assumptions about the size of the input the ...
0
votes
0answers
59 views

Grid search for autoencoder model

I'm using autoencoder for anomaly detection. Model's output is basically the reconstructed input based on the original input. The anomalies are then calculated using the reconstrction error. I would ...
0
votes
0answers
15 views

How to assign label to un-labeld text documents

I have a bunch of text documents and I would like to assign one or more than one label to each document. I know that autoencoders can be used in a semi-supervised setting to first cluster the ...
0
votes
1answer
207 views

Validation loss is lower than the training loss

I am using autoencoder for anomaly detection in warranty data. Architecture 1: The plot shows the training vs validation loss based on Architecture 1. As we see in the plot, validation loss is ...
0
votes
2answers
188 views

How to use Autoencoders for outlier detection on images

I have a bunch of images taken from a camera showing a pipe and would like to detect if the pipe is leaking or not. There are very few examples of leaking pipe in the data set. So considering this ...
-1
votes
0answers
83 views

Fraud Autoencoder in Keras weird metrics

I have been hammering this simple sketch for some time now, and I still cant wrap my head around what is happening. I am beginning to suspect that the error lies in the dataset, but anyway, I need ...
1
vote
1answer
31 views

Is train/test-Split in unsupervised learning of neural network necessary?

I am using autoencoder for anomaly detection in warranty data. It is unsupervised. I calculate the reconstruction error by the model and the records with high reconstruction error value is considered ...
1
vote
0answers
40 views

What is an intuitive explanation for the Importance Weighted Autoencoder?

I have been reading a paper by Burda et al. on Importance Weighted Autoencoders(IWAE) but I can't quite grasp what they mean by sampling the terms h1...hk. Do they mean you have separate models from ...
0
votes
1answer
32 views

Denoising Autoencoder Parameter Search

I have ran a hyperparameter search for a denoising autoencoder and the results suggest I should make the sizes of my hidden layers as large as possible (within the range of values I allowed it to ...
2
votes
1answer
89 views

What mu and sigma vector really mean in VAE?

In standard autoencoder, we encode data to bottleneck, then decode with using initial input as output to compute loss. We do activate matrix multiplication all over the network and if we are good, ...
3
votes
1answer
104 views

How can I prove bottleneck layer of my CNN auto encoder contain useful information?

I am using CNN autoencoder to create a state representation layer which I will later be feed into my Reinforcement Agent. So I trained my CNN autoencoder and it is giving nice state representations. ...
1
vote
0answers
18 views

Quantifying feature importances using Auto-encoders

I have a set of features(mixture of numerical and categorical), each of size n. I am embedding them into a dense lower dimensionality space using an auto-encoder. I want to know if it is possible to ...
1
vote
1answer
62 views

More weightage to a categorical feature for an Autoencoder model

I am using autoencoder for anomaly detection. I don't have any labels already and so its unsupervised. If I have categorical variables, I usually one hot encode before giving it to the model. I would ...
1
vote
0answers
73 views

Preserve colour in convolutional autoencoder

at the moment i work with convolutional autoencoder and now I'am looking for paper or methods that adresses a colour preversation. Most of the AE paper use grayscale images and loss functions such as ...
3
votes
1answer
150 views

How design a autoencoder architecture

I would like to build an autoencoder (CNN) to learn a representation of my data. I never built such a network and I have some experience in supervised learning (classification). I would like to know ...
3
votes
1answer
63 views

Fluctuating accuracy of Autoencoder

I am working on an Autoencoder in keras with the following setting: 185-86-32-2-32-86-185. The problem is that its accuracy is fluctuating, besides it gives new accuracy at every run. If this is due ...
1
vote
2answers
139 views

Cross validation for anomaly detection using autoencoder

I am using autoencoder for anomaly detection in warranty data. I don't have any ground truth labels to confirm whether the anomalies detected by the model is really an anomaly or not. Since I don't ...
1
vote
0answers
19 views

Autoencoder ambivalent about order of input data?

The problem I'm working to solve is this: Given a musician's prerecorded free-form playing. I want to analyze each of the individual notes to determine how "in-rhythm" it is. See the graph in the ...
2
votes
1answer
17 views

How do you know if your Autoencoder network is fully connected?

I am new in deep learning and I am confused about fully connected network. Is an Autoencoder with more than one hidden layers a type of deep neural network (DNN)? Is DNN always fully connected? Let's ...
2
votes
0answers
48 views

Encoder-Decoder Sequence-to-Sequence Model for Translations in Both Directions

Is it possible to use a pre-trained sequence to sequence encoder-decoder model which translates an input text in source language to an output in target language to do an inverse translation? That is, ...
2
votes
1answer
63 views

Why don't we want Autoencoders to perfectly represent their training data?

From Ian Goodfellow's Deep Learning Book: If an autoencoder succeeds in simply learning to set g(f(x)) = x everywhere, then it is not especially useful. Instead, ...
3
votes
2answers
319 views

Latent loss in variational autoencoder drowns generative loss

I'm trying to run a variational auto-encoder on the CIFAR-10 dataset, for which I've put together a simple network in TensorFlow with 4 layers in the encoder and decoder each, an encoded vector size ...
1
vote
1answer
121 views

Using DONUT algorithm with keras

I am trying to get this repo of Xu's DONUT algorithm running, however I am getting an error I do not quite understand. The readme says I should load raw_data as follows: ...
0
votes
1answer
861 views

Variational Autoencoder TIme Series

Can anyone suggest a blog where Variational Autoencoder has been used for time series forecasting?
2
votes
1answer
208 views

What is meant by 'training patch size'?

Currently I read a paper about symmetric skip connections for autoencoder (link). One experiment of them changes the the 'training patch size'. In my understanding patches are sub-boxes of an image ...
0
votes
2answers
108 views

Street address clustering?

I have a huge dataset of addresses. I have another data stream that contains addresses that I need to match against those in the original dataset. As all the addresses are user-provided, matching them ...
4
votes
1answer
2k views

Using an autoencoder for anomaly detection on categorical data

Say a dataset has 0.5% of its features continuous and 99.5% categorical (binary) with ~2400 features in total. In this dataset, each observation is 1 of 2 classes - Fraud (1) or Not Fraud (0). ...
0
votes
1answer
44 views

Training Encoder-Decoder using Decoder Outputs

I am trying to build an encoder-decoder model for a text style transfer problem. The problem is I don't have parallel data between the two styles so I need to train the model in an unsupervised ...
0
votes
0answers
168 views

Why do we use leakyRELU over RELU in Variational Autoencoders?

I have seen in multiple research papers and algorithmic implementations where leakyRELU is used for variational auto encoders. Knowing the fact that leakyRELU is used to obtain non-zero gradient for ...
2
votes
1answer
226 views

Preprocessing and dropout in Autoencoders?

I am working with autoencoders and have few confusions, I am trying different autoencoders like : ...
3
votes
1answer
517 views

Is there any clear tutorial for how to use AutoEncoders with text as input

I have a pandas dataframe that describes some fields of the register. I have used one hot encoding to encode the feature vectors that are not numbers. Finally my dataset now has 4000 rows * 4 columns. ...
1
vote
0answers
26 views

Using vector of floating points as labels for input data in a CVAE

The Conditional Variational Autoencoders (CVAE) (and other classification networks) I have come across use a one hot vector encoding for labeling categorial data sets. In my case, I do not have a ...
0
votes
1answer
47 views

Encoder-Decoder performance time

I have two encoder-decoder models. *First model: *Second model: When I check the performance of the models I get approximately the same performance time (First model ~ 42 sec, Second model ~ 40 ...
0
votes
1answer
38 views

autoEncoder as LSTM input, any benefit?

When training my LSTM, is there any incentive to pre-pass its inputs through an auto-encoder, or should I always supply as raw data as possible? I have is a small amount of training data, and it ...

1 2 3