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.

Filter by
Sorted by
Tagged with
1
vote
1answer
270 views
7
votes
2answers
369 views

How does attention mechanism learn?

I know how to build an attention in neural networks. But I don’t understand how attention layers learn the weights that pay attention to some specific embedding. I have this question because I’m ...
0
votes
0answers
17 views

Deep reinforcement learning - action is represented by elements of tensor

I am trying to design reinforcement learning algorithm. My action and state space are continuous. Action, which I would like to take can be represented by a matrix, lets say of dimension $n \times n$. ...
2
votes
2answers
69 views

How can we create an label, value detector?

I am trying to implement an text detector using MaskRCNN such that the model detects the label and value as shown in the image below. Detecting the same is easier for fields like page date and order ...
25
votes
4answers
30k views

What is the difference between Gradient Descent and Stochastic Gradient Descent?

What is the difference between Gradient Descent and Stochastic Gradient Descent? I am not very familiar with these, can you describe the difference with a short example?
2
votes
2answers
25 views

Could GANs be used to augment data?

I want to use GAN for data augmentation but I am confuse what are the pros. and cons. of data augmentation using GAN or why we use data augmentation using GAN compared to other data augmentation ...
-1
votes
1answer
17 views

validation accuracy and loss increase

I am training a generic LSTM based autoencoder to get the sentence embeddings, the bleu score is the accuracy metric. The model is coded to output the same number of tokens as the length of labels, ...
1
vote
1answer
30 views

What are x variable and y variable in word2vec model if it is supervised learning

What are x variable and y variable in word2vec model if it is supervised learning. In both the flavours- CBOW and skip-gram model. Though some blogs have explained it as unsupervised learning. ...
1
vote
0answers
33 views

Anomaly Detection/Novelty detection

I have a data-set that has over 6 million normal data and around 50 anomaly data.Those anomaly data is identified by manually(monitoring the user`s activity over camera and identify). I need to ...
0
votes
1answer
105 views

Training the document page layout and classifying good/bad layouts

I have a use case where I am supposed to get the coordinates of each block element in a page (whether its paragraph, image, table) where I train a model to understand how they are placed in a given ...
5
votes
1answer
290 views

The Bias-Variance Trade-Off

I'm reading "An Introduction to Statistical Learning: With Applications in R". In the Paragraph 2.2.2 The Bias-Variance Trade-Off, the authors say: I'm not able to understand why the bias tends to ...
1
vote
1answer
15 views

Accuracy Testing & Training

I have 102 observations. I made standard scale for dataset. I have found the accuracy training and accuracy testing values, but training score is 1.00 and testing score is -217.541. I have run with ...
1
vote
0answers
20 views

How are the Convolution kernels learned?

As I went through the basics of machine learning, I failed to understand how do the Convolutional layers in a CNN learn the convolution kernels. After looking at first few tutorials, I thought the ...
3
votes
1answer
25 views

How can calculate Efficiency for predictive models based on accuracy or error over time?

I was wondering if I could express the efficiency of prognostic models according to their accuracy(error, e.g. MAPE or MSE) over time [sec]. So let's imagine I have the following results for different ...
0
votes
0answers
15 views

Cross layer parameter sharing in ALBERT Model

I am reading the paper "ALBERT: LITE BERT FOR SELF-SUPERVISED LEARNING OF LANGUAGE REPRESENTATIONS". ALBERT uses cross layer parameter sharing to improve the model performance. I don't understand how ...
1
vote
1answer
14 views

Multi layer full connected neural network in tensor flow

I am reading about deep learning using tensor flow using book Tensor flow for deep learning by Bharath RamSundar et. Here author implemented as below in chapter 5 ...
2
votes
2answers
43 views

Detect malicious GIFs

I was reading this article talking about a form of targeted internet bullying which involves sending flashing images via Twitter to people with epilepsy. I was wondering whether there is a way to ...
-1
votes
0answers
21 views

RuntimeError: CUDA out of memory. Tried to allocate 20.00 MiB

I'm trying to make a GAN for generating pictures using this code: https://github.com/abhinav3/Udacity-DCGAN-FaceGeneration/blob/master/dlnd_face_generation.ipynb I'm doing this learning on GPU. The ...
3
votes
1answer
40 views

Alternatives to regression to decide weights in an expression

I have a use case in which I am required to predict variable $y$ which depends on five variables $x_i$. Consider something like $$ y=w_1 x_1+ w_2 x_2+ w_3 x_3+ w_4 x_4+ w_5 x_5.$$ This expression ...
2
votes
1answer
47 views

Performance of CNN based deep models with number of classes

How does a given deep cnn model performance vary with number of classes in tasks such as classification, object detection segmentation? For example mobilenet v2 gives around 90% accuracy on ...
2
votes
0answers
15 views

Approach to classify blocks of time series

I am wondering if there exists an approach to classify blocks of time series, and not specifically individual time series. If so, can you point me out papers/articles/tutorials where these type of ...
2
votes
0answers
17 views

What doese 'v' mean in GoogLeNet?

In GoogLeNet (this link), there is 'v' notation in Figure3 like '1X1+1(v)'. I don't know the meaning of 'v'. Also, I understood 's' as stride. But, I don't know the reason why plus operation is used ...
1
vote
0answers
25 views

How to estimate and/or determine how much data is enough to train a model?

To train a nice supervised algorithm (for instance, a dependency parser, a parts-of-speech tagger or NER) data is essential, but how many samples are necessary or enough? From what kind of ...
1
vote
0answers
11 views

how to update a face recongition ( e.g facenet)

I'm searching for a face recognition that detects new faces not just faces getting from the datasets, for example, facenet is a project that can detect and recognize a face from labels. I would like ...
1
vote
2answers
32 views

Does it make sense to do train test split when trainning GANS?

For normal supervised learning the dataset is split in train and test (let's keep it simple). Generative Adversarial Networks are unsupervised learning but there is a supervised loss function in the ...
1
vote
1answer
27 views

Why would one use Deep Learning on non-local datas?

I understand the using of deep-learning for datas that have "local" datas, for example images/videos/texts, as the convolutionals layers reduce the amount of dimensions. However, I saw that some ...
0
votes
2answers
510 views

Deep learning: Training from COCO/imagenet weights or from a class-specific trained weights when classify 2 similar objects

I have a question about the possible outcome of a trained model. Imagine that I would like to classify 2 different models of Ferrari and the dataset of these 2 models is small (for example, a few ...
2
votes
1answer
66 views

How to properly apply CrossValidation and/or split the dataset?

I have a particular problem and do not really now how to properly validate my experiments in this scenario. There is one big data set with 100.000 samples, 99.000 y=0, 1.000 y=1 Each sample has 1.000 ...
1
vote
1answer
24 views

Finding sequence combinations that impact target variable the most

One can create a time series model to predict a target variable. What I need to do is find the input combinations and sequences that impact the target variable the most. In this case, the input data ...
1
vote
0answers
14 views

I don't know local spatial correlation meaning

In GoogLeNet papaer, 'local sparse correlation' is mentioned. But, I don't know the meaning. Also, I don't understand that statement " In the lower layers (the ones close to the input) correlated ...
2
votes
1answer
140 views

Are Denoising Autoencoders for anomaly detection on structured data?

Can denoising autoencoders be used for anomaly detection on structured data? I know I can use denoising autoencoders for anomaly detection on images, but I don't know if they can do it for structured ...
3
votes
1answer
36 views

Risk prediction vs classification model

I am working on a binary classification model. Currently, when I use scikit logistic regression, it outputs binary values like 0s and 1s. However, I understand, from online reading, that it outputs ...
3
votes
1answer
24 views

How to quantitatively evaluate raw neural network activations?

Below are the activations for 2 different predictions. These predictions are for different labels/classes. They are being run through a dense Keras NN (96.6% accurate) with 2 hidden layers and adam ...
3
votes
1answer
23 views

Is it wrong to use Glorot Initialization with ReLu Activation?

I'm reading that keras' default initialization is glorot_uniform. However, all of the tutorials I see are using relu ...
3
votes
1answer
263 views

Explanation of how DeepExplainer works to obtain SHAP values in simple terms

I have been using DeepExplainer (DE) to obtain the approximate SHAP values for my MLP model. I am following https://github.com/slundberg/shap and DE's performance is very high in terms of computation ...
1
vote
0answers
24 views

What is the difference between layer degradation and overfitting?

I've read the ResNet paper (https://arxiv.org/pdf/1512.03385.pdf) They determined the layer degradation as that model with less layers learn quicker than with more. It can be visiable on plots below: ...
2
votes
2answers
313 views

Auto ML vs Manual ML for a project

I recently was introduced to a AUTO ML library based on genetic programming called tpot. Thanks to @Noah Weber. I have few ...
6
votes
0answers
66 views

What are the good parameter ranges for BERT hyperparameters while finetuning it on a very small dataset?

I need to finetune BERT model (from the huggingface repository) on a sentence classification task. However, my dataset is really small. I have 12K sentences and only 10% of them are from positive ...
3
votes
1answer
29 views

Why is my validation loss going up while my validation accuracy also goes up?

I'm wondering if anyone can help me understand this. I've been playing around with training a CNN for the cifar10 dataset. I'm using tensorflow, and a CNN with 12 conv layers and 1 dense layer before ...
3
votes
1answer
33 views

How do we define a linearly separable problem?

When we talk about Perceptrons, we say that they are limited for approximating functions that are linearly separable, while Neural Networks that use non-linear transformations are not. I am having ...
1
vote
0answers
11 views

Hierarchical method for handling large class classification [closed]

I am looking for an ML algorithm which is appropriate for the following situation: I am trying to pick out specific invoice numbers from remittance advice images which contain lots of irrelevant ...
2
votes
1answer
34 views

Predict votes of future comments [closed]

I have a database with a lot of comments, each comment has a vote, a vote can be positive or negative. ex : -2, -5, -90, +45, +20... So based on this training dataset I want to predict votes of ...
5
votes
5answers
157 views

How can we extract fields from images?

I am making an document parser which extracts data fields from the documents and store them in a structured way. Each field in my dataset is horizontal which is easy to extract. But the model fails ...
0
votes
1answer
57 views

Quasi-linearity in deep learning regression problems (sports betting)

I’m attempting to build a sports betting model that aims to predict final scores for games. I’ve had some promising early results for US college football just by using linear regression to form team ...
-1
votes
0answers
9 views

How to interpret the target and output validation method [closed]

I have just run a machine learning model and am validating using Target and Output method. Please can someone tell me how I can interpret the graph
0
votes
0answers
35 views

How to approach new project in Medical Image analysis?

I am working on a project in medical image analysis - Breast cancer detection. And since I am the one who has proposed this project I was wondering, what would be the (data science) steps its ...
0
votes
0answers
16 views

Generic object detection - unspecified list of classes and high accuracy

As a part of a small project, I would like to create tags for a set of pictures (posters). I know that if I want to recognize a lot of objects I need to have a model that was trained on a large ...
2
votes
2answers
918 views

Neural network approach to the cocktail party effect

Imagine you have 2 people at 2 different microphones but in the same room. Each microphone is going to pick up some sound from the other person. Is there a good neural network based approach to ...