Questions tagged [deep-learning]

For questions related to deep learning, which refers to a subset of machine learning methods based on artificial neural networks (ANNs) with multiple hidden layers. The adjective deep thus refers to the number of layers of the ANNs. The expression deep learning was apparently introduced (although not in the context of machine learning or ANNs) in 1986 by Rina Dechter in the paper "Learning while searching in constraint-satisfaction-problems".

Filter by
Sorted by
Tagged with
0
votes
0answers
35 views

Reference for Yoshua Bengio quote

I believe that Bengio is on record as describing Deep Learning as "A collection of methods for delivering intelligence" (or something very similar). Does anyone have an actual reference for ...
2
votes
0answers
27 views

If random rotations are included in the data augmentation process, how are the new bounding boxes calculated?

When studying bounding box-based detectors, it's not clear to me if data augmentation includes adding random rotations. If random rotations are added, how is the new bounding box calculated?
0
votes
0answers
24 views

Am I using transfer learning when I use SSD ResNet 50 model architecture?

Using Label-img, I have successfully labeled my images (dimensions 1100 x 1100 pixels), and am currently training the SSD ResNet50 model (from the TensorFlow 2 Detection Model Zoo). I downloaded the ...
3
votes
1answer
45 views

DQN layers when state space and action space are multi dimensional

I have built my own RL environment, where a state is composed of two elements: the agent's position and a matrix of 0s and 1s (1 if a user has requested a service from the agent, 0 otherwise); an ...
0
votes
0answers
13 views

Does a Siamese Network need other trainable layers after the distance layer?

I'm approaching at Siamese Networks in order to use them for Image Similarity. I found that many people use famous models like VGG or ResNet to build the vectors that will go on the distance layer in ...
1
vote
0answers
22 views

How can I improve the performance on unseen data for semantic segmentation using an auto-encoder?

I am using simple autoencoders for the task of semantic segmentation on the VOC2012 dataset. I am currently using a simple autoencoder based model. It is trained on adam optimizer with cross-entropy ...
2
votes
0answers
59 views

Why should the weight updates be proportional to input?

I'm reading the book Grokking Deep Learning. Regarding weight updates during training, it has the following code and explanation: ...
1
vote
0answers
39 views

Where can I find pre-trained agents able to play games with multiple stages like exploration, dialog, combat?

My goal is to create an ML model to be able to classify different game stages, e.g., dialog with a non-player character, exploration, combat with enemy, in-game menu etc. In order to do that, I am ...
1
vote
0answers
33 views

In the MINE paper, why is $\hat{G}_B$ biased, and how does the exponential moving average reduce the bias?

While reading the Mutual Information Neural Estimation (MINE) paper [1] I came across section 3.2 Correcting the bias from the stochastic gradients. The proposed method requires the computation of the ...
0
votes
0answers
51 views

How to determine the number of fully connected layers for a convolutional neural network?

How many fully connected layers should be added to a convolutional neural network? Does it depend on input size to the fully connected layer? If so, how do we decide? What if the input size of the ...
1
vote
0answers
27 views

When training deep learning models for object detection in images, do you need a large number of images, or a large number of training samples?

I am training a deep learning model for object detection. The consensus is that the more images that you have, the better the results will be. All the tutorials that I have seen say that more images ...
0
votes
0answers
20 views

CNN to detect presence/absense of label on images with mixed labels

Here's my problem: I work with medical image classification, and currently I have 3 classes: class A: images with lesion 1 only; and images with lesion 1 and N other lesions class B: images with 2 ...
1
vote
0answers
17 views

How does scheduled sampling for transformers work?

I was reading this paper which applies a modified version of the transformers for traffic forecasting. I am somewhat familiar with the transformer architecture and how it functions, but, in the paper, ...
1
vote
0answers
28 views

What is a “center loss”?

I have seen that a center loss is beneficial in computer vision, especially in face recognition. I have tried to understand this concept from the following material A Discriminative Feature Learning ...
1
vote
2answers
87 views

How to verify classification model trained on classification dataset on a detection dataset for classification purpose?

I am working on a problem that involves two tasks - detection and classification. There is no single dataset for both tasks. I am training two models, separate on detection dataset and another on ...
1
vote
0answers
21 views

What exactly are deep learning primitives?

I came across the concept of "deep learning primitives" from the Nvidia talk Jetson AGX Xavier New Era Autonomous Machines (on slide 44). There doesn't seem to be a lot of articles in the ...
1
vote
0answers
25 views

WGAN-GP Loss formalization

I have to write the formalization of the loss function of my network, built following the WGAN-GP model. The discriminator takes 3 consecutive images as input (such as 3 consecutive frames of a video) ...
0
votes
0answers
36 views

Is Webpage Semantic Segmentation possible nowadays?

I'm trying to do some research about semantic segmentation for webpages, in particular e-commerce webpages. I found some articles which provide some solutions based on very old dataset and those ...
1
vote
0answers
18 views

REINFORCE Agent suddenly drops. How to verify if it's due to catastrophic forgetting?

I am using the default implementations of REINFORCE, DQN and c51 available from the tf.agents repo (links). As you can see, DQN manages to improve performance while REINFORCE seems to suffer from ...
0
votes
0answers
9 views

Is there a framework or method that would help visualise inner workings of a feedforward neural network?

I wonder if there is some framework or method to help visualising inner workings of a feedforward deep neural network? What I mean by this is something similar to what is being done with CNNs where we ...
1
vote
1answer
50 views

Single-Shot Learning for Object Re-Identification

I am looking for a way to re-identify/classify/recognize x real life objects (x < 50) with a camera. Each object should be presented to the AI only once for learning and there's always only one of ...
0
votes
1answer
23 views

Should binary feature be in one or two columns in deep neural networks?

Let's assume I have a simple feedforward neural network whose input contains binary 0/1 features and output is also binary two classes. Is it better, worse or maybe totally indifferent, for every of ...
1
vote
0answers
29 views

How to use Deep Q-Network with two-dimensional input? Hands-on Machine Learning 2

I'm studying with the book Hands-on Machine Learning with Scikit-Learn, Keras and TensorFlow, and I'm trying to implement the Deep Q-Network example that can be found on Github but that the input ...
3
votes
1answer
31 views

Are there deep neural networks that have inputs connected with deeper hidden layers?

Are there any architectures of deep neural networks that connect input neurons not only with first hidden layer but also with deeper ones (red lines on the picture)? If so could you give some names ...
2
votes
0answers
70 views

What is the justification for Kaiming He initialization?

I've been trying to understand where the formulas for Xavier and Kaiming He initialization come from. My understanding is that these initialization schemes come from a desire to keep the gradients ...
1
vote
1answer
102 views

Applications of Information Theory in Machine Learning

How is information theory applied to machine learning, and in particular to deep learning, in practice? I'm more interested in concepts that yielded concrete innovations in ML, rather than theoretical ...
0
votes
0answers
48 views

Deep Learning based image restoration using multiple frames

Suppose we have a sequence of still images each of which has been contaminated by some particles(ex, dust/sand/smoke) making the images very poor in certain areas. What architecture would be best to ...
1
vote
1answer
39 views

What do we mean by “infrequent features”?

I am reading this blog post: https://ruder.io/optimizing-gradient-descent/index.html. In the section about AdaGrad, it says: It adapts the learning rate to the parameters, performing smaller updates (...
1
vote
1answer
74 views

How can I determine whether a video's frame is realistic (was recorded by a camera) or contains computer-generated graphics?

Given a video, I'm trying to classify whether it is a graphical (computer-generated) or realistic scene. For instance, if it contains computer-generated graphics, credit, moving bugs, blue screen, etc....
2
votes
2answers
233 views

How to tell a neural network that: “your i-th input is special”

Assume that I have a fully connected network that takes in a vector containing 1025 elements. First 1024 elements are related to ...
1
vote
1answer
40 views

How does vanish gradient restrict RNN to not work for long range dependencies?

I am really trying to understand deep learning models like RNN, LSTMs etc. I have gone through many tutorials of RNN and have learned that RNN cannot work for long Range dependencies, like: Consider ...
0
votes
0answers
25 views

XOR problem with bipolar representation

I am taking a course in Machine Learning and the Professor introduced us to the XOR problem. I understand the XOR problem is not linearly separable and we need to employ Neural Network for this ...
0
votes
0answers
19 views

Extracting specific information from an Invoice images

Tried to extract only specific information from the images but Couldn't, We have to automate this process using this as the format of the Images keeps on changing. LinkSample data What I have Tried: <...
0
votes
0answers
14 views

How long will it take to train SSD V1 Mobilenet image recognition algorithm?

I am training a deep learning algorithm using an NVIDIA GEFORCE RTX GPU, with 16 GB RAM. I went through my image database and generated 50,000 training samples with the Labelimg software; the images ...
1
vote
0answers
61 views

What is the computational complexity in terms of Big-O notation of a Gated Recurrent Unit Neural network?

I have been digging up of articles across the internet in context of computational complexity of GRU. Interestingly, I came across this article, http://cse.iitkgp.ac.in/~psraja/FNNs%20,RNNs%20,LSTM%...
1
vote
2answers
71 views

Is it possible to have a variable-length latent vector in an autoencoder?

I'm trying to have a simple autoencoder but with variable latent length (the network can produce variable latent lengths with respect to the complexity of the input), but I've not seen any related ...
0
votes
0answers
29 views

How did they use their dataset with VAEs?

Old Photo Restoration via Deep Latent Space Translation (https://paperswithcode.com/paper/old-photo-restoration-via-deep-latent-space) In the article, it says : "We propose to restore old photos ...
1
vote
1answer
37 views

How to define Agar.io state and action space?

I am trying to implement an AI bot for my Agar.io clone using deep neural network. However, I am struggling with the state and action space of the AI bot. Because the bot can take real number for ...
0
votes
0answers
26 views

Wasserstein GAN with gradient penality - Loss values

I have trained a WAN with gradient penalty and the loss values ​​seem to me much higher than the examples I have seen on the net. The generator receives 2 images as input and must generate a ...
2
votes
0answers
34 views

How to define loss function for Discriminator in GANs?

To train the discriminator network in GANs we set the label for the true samples as $1$ and $0$ for fake ones. Then we use binary cross-entropy loss for training. Since we set the label $1$ for true ...
1
vote
1answer
42 views

Should the range and initial values of weights and biases be adjusted to fit input and output data?

As a routine (in typical everyday tasks) of a data scientist, should they usually decide about weights and biases range and initial values as a function of which data they are planning to insert as ...
1
vote
0answers
12 views

How Restricted Boltzman Machine (RBM) generates hand-written digit?

I am reading RBMs from this paper. In Fig1 they show an example of generating hand-written digit using RBMs. This is the figure they are showing: In the learning step first we sample $h$ from $h \sim ...
0
votes
0answers
27 views

Deep Continuous Clustering algorithm - just one output cluster

I use the DCC algorithm to cluster some data. The whole algorithm is available here, but shortly it is: construct mkNN graph of the data points (the connected components of it are the clusters). ...
2
votes
1answer
60 views

How does back-propagation through time work for optimizing the weights of a bidirectional RNN?

I am aware that back-propagation through time is used for training the recurrent neural network. But I am not able to understand how this happens for the bi-directional versions of the recurrent ...
2
votes
0answers
21 views

What is the difference between text-based image retrieval and natural language object retrieval?

I'm working on creating a model that locates the object in the scene (2D image or 3D scene) using a natural language query. I came across this paper on natural language object retrieval, which ...
3
votes
1answer
93 views

What exactly is an interpretable machine learning model?

From this page in Interpretable-ml book and this article on Analytics Vidhya, it means to know what has happened inside an ML model to arrive at the result/prediction/conclusion. In linear regression, ...
0
votes
0answers
26 views

Choice of loss function for semantic segmentation

I am training a U-Net for semantic segmentation of large medical images (4096x4096px). The two classes are "too" unbalanced. The white pixels are just about 0.1% (or less) of the whole image....
1
vote
1answer
45 views

What is the status of the capsule networks?

What is the status of the capsule networks? I got an impression that capsule networks turned out not to be so useful in applications more complicated than the MNIST (at least according to this reddit ...
1
vote
0answers
22 views

Machine Learning Techniques for Objects Location/Orientation in Images

what Machine Learning tool can understand in which location and orientation a picture was taken from? That is from pictures of similar objects, say for example pictures of car interiors. So given a ...
0
votes
1answer
68 views

speech comment detection by deep speech mozilla for data set

I want to create a system so that when a human being says a word or command through a microphone, such as "shut down", the system can execute that command "shut down". I used the ...

1 2
3
4 5
29