Stack Exchange Network

Stack Exchange network consists of 175 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

Questions tagged [deep-learning]

a new area of Machine Learning research concerned with the technologies used for learning hierarchical representations of data, mainly done with deep neural networks (i.e. networks with two or more hidden layers), but also with some sort of Probabilistic Graphical Models.

1
vote
0answers
13 views

One hot encoding as input to recurrent neural networks

I'm trying to predict next label in a pattern based on previous labels using recurrent neural network. In total I have 100 labels Example of input pattern: ...
0
votes
0answers
3 views

Weighted MMD for InfoVAE?

I'm trying to figure out how can weighted MMD from Weighted Maximum Mean Discrepancy for Unsupervised Domain Adaptation (chapter 3. Weighted Maximum Mean Discrepancy) be adapted for InfoVAE: A ...
1
vote
0answers
11 views

Keras Binary Classification val_acc won't go past ~67; Full data and code included

I'm working on a binary classification in Keras with a Tensorflow backend. No matter how much I tweak, I can't seem to get my model past a val_acc of 67%. Is there something I'm missing, or is this ...
3
votes
1answer
34 views

Approaching a multi-class classification problem but without labels

I am working on a business problem where I have a movie description dataset. In this dataset I've columns as - Movie title, Movie plot summary, Date of Release. Now based on this information and using ...
0
votes
0answers
10 views

How can NN be thought of as p(x|z) in VAEs?

So, you can feed random input (z) to a NN and make the output random, but that will be just f(p(z)), where f(.) is the deterministic NN. The other way is to think of the NN output as parameteres of a ...
0
votes
1answer
5 views

Why the loss is nan by using linear activation function in the last layer?

I want to use neural network to solve a simple regression problem, and I try to program by myself accroding to lecture Backpropagation and Neural Networks . However, I meet loss divergence problem. ...
0
votes
1answer
57 views

*Challenge* Making an algo that learns from a book, and can answer anything about it

I recently took this challenge where I am trying to make a set of algorithms to read any particular book, understand and store the context and subsequently answer any question asked about it. In ways ...
1
vote
1answer
53 views

Philosophical question on redundancy

Suppose I implement a supervised learning version of LSTM similar to this. Namely, I have these univariate time series data: ...
1
vote
0answers
35 views

What is the best architecture for Auto-Encoder for image reconstruction?

I am trying to use Convultional Auto-Encoder for its latent space (embedding layer), specifically, I want to use the embedding for K-nearest neighbor search in the latent space (similar idea to ...
0
votes
0answers
21 views

How to build this simple generative model?

I'm watching a lecture and it shows (schematically) a model with 2D input (sampled from Gaussian) followed by 12 conv layers, and output an RGB image. That's all about it, no training at all! he ...
0
votes
0answers
22 views

What are the best ways to deal with huge datasets in deep learning? [on hold]

I have to do a deep learning project in which the size of the datasets are too big for my computer. My question is very simple: What are the best ways to deal with this kind of project? To me, the ...
0
votes
1answer
21 views

What is used for Machine Translation besides RNN

I am doing a university report and it seems that encode-decode RNN are optimal for Machine Translation. I need something else to compare it to but I can't seem to make a proper google search for it. ...
1
vote
1answer
28 views

Are CNNs applicable on structured data?

I can use CNN to classify MNIST images, but I don't know whether CNNs are applicable on iris data as well? If not, why?
1
vote
1answer
17 views

Why my CNN model is not learning?

I want to train a model to predict one's emotion from the physical signals. I have a physical signal and using it as input feature; ecg(Electrocardiography) In my dataset, there are 312 total ...
1
vote
0answers
8 views

Transfer Learning with CNN layer trainable True - Accuracy not improving

I am working on a image classification problem with 4 classes. And I am using Transfer Learning (Resnet50) to train the model. Below are the observation. Pre-trained weights are from ...
0
votes
0answers
15 views

Why is the reward fluctuating for Double Q-Learning?

I am trying to implement Double Q-Learning using neural networks from the Keras library. When I first tried Simple DQN, the graph of the reward was fluctuating a lot so, I implemented a Double DQN. ...
1
vote
0answers
18 views

my brittle PPO implementation for Cartpole, how can I make it more resilient?

I implemented the clipped objective PPO-clip as explained here: https://spinningup.openai.com/en/latest/algorithms/ppo.html Basically I used a dummy actor network to find the new action probability ...
1
vote
0answers
7 views

How do I interpret the Wavelet scattering transform frequency plane and tree diagrams?

I've been reading through this paper on "Invariant Scattering Convolution Networks" and am having a difficult time interpreting the tree diagrams and frequency plane diagrams shown respectively below. ...
0
votes
0answers
7 views

Not getting good predictions with Conv1D for vector classification [on hold]

I have a total of 54404 samples, each is a (8,1) vector. The number of classes is 8 and I want to run a keras CNN to classify them. I split the data after shuffling as follows: ...
1
vote
1answer
16 views

With a MLP (regression), is it appropriate to initialize bias in the final layer to be a value near the expected mean?

For instance, when predicting IQ in a population you would expect the mean to be 100. If you initialize the bias in the final layer you are basically giving the network a head start, telling it in ...
-3
votes
0answers
24 views

GANs for Stock Market Prediction [closed]

I need help in implementing a GAN for predicting stock market prices. I want to know, how can I implement this using Keras framework and how would I go in structuring the whole architecture. I will ...
1
vote
0answers
19 views

Predict a 100 dimensional array

I have several sentences that I transformed into vectors. With these vectors I would like to predict another vector (which represents a vector of a sentence (the answer)). Can you tell me if this ...
0
votes
1answer
16 views

What are the key differences between a MLP with lagged features and a RNN

I've been working with MLP's for a while. Whenever I assumed that the past values of a feature might be useful for predicting the future values of Y, I would just create a new column in my data frame ...
0
votes
0answers
19 views

How to do Natural Language Processing with few samples only?

I am familiar with SMOTE (Synthetic Minority Oversampling), and the Python Library imbalanced-learn, which can be used to handle ...
1
vote
1answer
23 views

Why is my MLP with 2 features is doing worse than MLP with 1 feature where the one feature is a combination of feature1*feature2?

I have programmed a MLP for a dataset (~500 rows) containing the length (L) and width (W) of an organism and the output of biomass (the organisms weight in pounds, B). ...
0
votes
0answers
10 views

How to use document and sentence embeddings in Keras?

Although there is an easy way to use Embedding layer in keras and make use of pretrained word embeddings, is there a way to use document or sentence embeddings?
-1
votes
0answers
21 views

What can I do to reduce false postive? [duplicate]

What should I do to reduce false positive? I use Mask RCNN.
0
votes
0answers
27 views

Deep learning(MLP) on multiclass classification. Model learns only one class

I am new to deep learning. I have imbalanced class data. I used one hot encoding and scaling to preprocess my data. I have used adamoptimizer as optimizer function and sparse categorical crossentropy ...
1
vote
1answer
41 views

Is this a data issue, or a model issue? A Keras binary classification model

I've been trying to create a binary classification model that predicts wether there will be a train delay based on the train and time. Here is a link to the data The issue I'm having is that my ...
0
votes
0answers
12 views

LSTM Bounded Forecast Time Series

Can anyone please provide a logical explanation as to why an LSTM produces a 'bounded' forecast when predicting over unseen time series data? This behaviour does not seem to occur when using a MLP. ...
-2
votes
0answers
25 views

I am getting 100% accuracy in 1st epoch how could it is possible? is this problem with data loading or over fitting? [closed]

(env) hareesh@hareesh-Z270M-D3H:/media/hareesh/hareesh/code_version$ python gen_v7.py Using TensorFlow backend. 101 ...
4
votes
1answer
37 views

Large action space for deep reinforcement learning

I know that in normal Deep Reinforcement Learning(DRL) scenario, we learn a deep neural network to map current states to Q values. The number of the Q values (# of outputs of the neural network) is ...
2
votes
1answer
49 views

Time Series - Models seem to not learn

I am doing my undergrad Dissertation on time series prediction, and use various models (linear /ridge regression, AR(2), Random Forest, SVR, and 4 variations of Neural Networks) to try and 'predict' (...
0
votes
0answers
14 views

How to code a CNN with batches with different input shapes

I want to make a CNN or FCN that can take grayscale images as an input and outputs a color image. It is very important to me that the size of the images can vary. I heard that I can only do this when ...
0
votes
1answer
53 views

Deep learning theory: why are hidden layers necessary?

For this question, I’ll refer to the popular YouTube video by 3Blue1Brown on deep learning applied to recognition of written numbers: https://www.youtube.com/watch?v=aircAruvnKk The video describes ...
0
votes
0answers
37 views

Neural network with different input shapes

I'm currently designing the architecture of a neural network for the colorization of grayscale images. Later on it should be able to colorize images with different sizes and different aspect ratios. I ...
1
vote
0answers
17 views

How does BERT deal with catastrophic forgetting?

In the ULMFit paper authors propose a strategy of gradual unfreezing in order to deal with catastrophic forgetting. That is, when the model starts be fine-tuned according to a downstream task, there ...
2
votes
1answer
139 views

What is fractionally-strided convolution layer?

In paper Generating High-Quality Crowd Density Maps using Contextual Pyramid CNNs, in Section 3.4, it said Since, the aim of this work is to estimate high-resolution and high-quality density maps,...
1
vote
0answers
16 views

Character segmentation using deep learning

I'm developing a character segmentation algorithm for license plate OCR. My algorithm includes two steps: segmentation and recognition. There is almost no problem for recognition thanks to CNN. My ...
2
votes
3answers
47 views

How to measure the similarity between two text documents?

Assume, I have 100 text documents, and I want to cluster those documents. The first step is the construct pairwise similarity matrix 100X100 for the documents My ...
0
votes
1answer
28 views

Is the role of the validation set in a deep learning network is only for Early Stopping?

In the "deep learning crash course" given by Leo Isikdogan in lecture 4 https://www.youtube.com/watch?v=ms-Ooh9mjiE&list=PLWKotBjTDoLj3rXBL-nEIPRN9V3a9Cx07&index=4 Overfitting, Underfitting, ...
1
vote
0answers
13 views

How can I edit log from tensorflow or plot in tensorboard?

I trained a neural network (32 epochs). I stop it. And next load weights from last epoch and I have been training for 18 epoch. Unfortunately, it started at 1 epoch and I got ugly plots. What can I ...
0
votes
0answers
20 views

Understanding notation of Goodfellow's GAN objective function

What is the meaning of $V(D,G)$? How do we get these expectation parts? I was trying to understand it following this article: Understanding Generative Adversarial Networks (D.Seita), but, after many ...
0
votes
0answers
10 views

Why is the training accuracy not increasing while implementing MemN2N?

I am trying to implement the MemN2N from here. I tried implementing the code using the Functional API of Keras. I tried changing the optimizer but, I got the same results. Below is my implementation:...
0
votes
1answer
51 views

Maximize Precision Deep Learning

For some binary image classification problems having close to 100% precision is super important and recall is much less important. What are best practices for maximizing precision? Setting the ...
0
votes
1answer
33 views

Neural Net Accuracy: Test Set vs Real World Data

Neural Net accuracy is high on test set but low on new real world image examples. Looking for advice regarding what generally causes this scenario and how to fix it. Sampling basis? Training/test ...
0
votes
0answers
11 views

How cnn feature map feed to lstm cell in tensorflow?

If possible suggest some good documentation about tensor shape.
2
votes
1answer
29 views

is it possible determine a best neural network classifier by considering only accuracy

Below, you have the accuracy plots for training and testing set for 6 different neural networks. is it possible to say, which of the following neural network classifier is better?. Having this little ...
0
votes
1answer
26 views

Policy gradient/REINFORCE algorithm with RNN: why does this converge with SGM but not Adam?

I am working on training RNN model on caption generation with REINFORCE algorithm. I adopt self-critic strategy (see paper Self-critical Sequence Training for Image Captioning) to reduce the variance. ...
0
votes
0answers
17 views

How to calculate receptive field size for -ception model?

I have a full convolution model like this: ...