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
0
votes
0answers
18 views
-1
votes
0answers
14 views

Help with Bayesian Deep Learning

A Crosspost from StackOverflow I am trying to use TensorFlow Probability to implement Bayesian Deep Learning with dense layers. I have trained a model on my dataset with normal dense layers in ...
0
votes
0answers
8 views

Where does BERT fit in the Machine Learning Hierarchy?

I am a newbie in the machine learning world and I need guidance from the professionals. I am trying to make a hierarchy starting from machine learning, then to deep learning and to BERT. I have read ...
1
vote
0answers
20 views

How can realize the evaluation/validation of unsupervised models through unlabeled data?

I'm researching anomaly detection, which is nothing else than outliers detection on a set of time-series web servers access log data or network traffic. Recently I re-faced to following fundamental ...
0
votes
0answers
7 views

I have no access to gpu due to usage limits?

I start running my code using google colab I first set the execution to GPU and then I run my code for a training task using keras !after 1 hour I got a message saying I can't use GPU due to usage ...
0
votes
1answer
13 views

ModuleNotFoundError: No module named 'deeplab'

I am trying to run Tensorflows DeepLab using my own dataset, while following: heaversm tutorial on GitHub/Youtube. P.S I am running Windows. I want to run the model_test.py script (on my windows ...
0
votes
0answers
6 views

Satellite Change Segmentation using Unet

Hi StackExchange community I am working on to train a Unet for satellite change segmentation. My dataset consists of images(before change),images(after change) and the corresponding change ...
1
vote
0answers
4 views

CheXnet model training and validation loss stuck

I am training a CheXnet like model from scratch. I followed paper (https://arxiv.org/pdf/1711.05225.pdf) and built the model similar. But I am unable to train it, during end-to-end training the ...
0
votes
0answers
9 views

Pseudo code for cross-validation to find layer-specific learning rate? [closed]

https://arxiv.org/pdf/1510.04609.pdf How does one implement the idea in the paper attached?
0
votes
0answers
14 views

What are the implications of training a model with head data only

Let's assume that a model M1 was trained based on a given training-set D1. Now let's assume we take 70% of the head data from D1 to create D2. If we train a model M2 which is based on D2 what are the ...
0
votes
0answers
7 views

Why do we do small number of iterations per epoch and large number of epochs rather than one epoch iterate on entire batch until small gradient?

Why do we do many epochs and one iteration per epoch rather than one epoch and iterate while gradient is not within tolerance in Keras deep learning or by default?
0
votes
0answers
31 views

Keras Functional API Model not learning

I've got a very strange error with keras functional API. I'm trying to fine tune a VGG16 model, using the keras functional model. The problem is that tha model does not learn, loss is stuck at 1.6094. ...
0
votes
0answers
13 views

I think a learning rate schedule would be counter-productive with AdaBelief. Am I wrong?

I am inclined to believe the concept of a learning rate schedule is overcome by the improvements of Adabelief over Adam. My code is on Github; please check it out and attempt to replicate my results, ...
1
vote
1answer
32 views

Causal inference VS Active learning?

Imagine we have some lists of features that are changing in time. Each row of the list corresponds to a sample (Change in space). I would like to know whether machine learning is able to determine the ...
0
votes
0answers
13 views

Periodical loss increase in the learning curve

I am training a transformers-based machine translation (NMT) model. The size of the parallel corpus is 4.5 million sentence pairs in two languages. What I am observing in the learning curve is that ...
0
votes
1answer
14 views

Cause of periodic jumps in loss function

I might be missing something obvious as I am new to machine learning. I am training an SSD Inception V2 for detecting buildings from satellite images. I use the Tensorflow Object Detection API. I am ...
0
votes
0answers
21 views

Vanishing Gradients and Batch Normalization [closed]

I am working with a VGG network which suffers from vanishing gradients. The gradient flow plots show 0 gradients until the last layer which ramps it up. I have implemented Batch Normalization to solve ...
3
votes
1answer
108 views

Attention for time-series in neural networks

Neural networks in many domains (audio, video, image text/NLP) can achieve great results. In particular in NLP using a mechanism named attention (transformer, BERT) have achieved astonishing results - ...
-3
votes
0answers
14 views

“Memory Error” - deploying in AWS elastic beanstalk free tier [closed]

While trying to deploy flask image classification model on elastic beanstalk, I am getting this memory error. Is it because of the limited size of 512 MB provided for uploading source code? I face ...
1
vote
1answer
20 views

Predicting Multiple Values Values Using Time Series Forecasting

I want to illustrate my question with the following example: I have a wholesale company through which I sell 200 products: P1,P2,P3 .... P200 to a 1000 customers ...
0
votes
0answers
5 views

How to store and query biometric data for an authentication system?

I am trying to design, and hopefully implement, an authentication system which centers around the use of biometric images. I plan to use different machine learning and deep learning techniques to help ...
0
votes
0answers
7 views

Estimating coefficients of AR(n) process with deep learning

Can a deep learning model estimate parameters of an AR(n) process given a time series generated by such process? What kind of deep-learning model would you recommend? Note: I am very new to machine ...
4
votes
1answer
33 views

Is there an inherent recency bias in deep learning?

When working with very large models within Deep Learning, training often takes long and requires small batch sizes due to memory restrictions. Usually, we are left with a model checkpoint after ...
0
votes
0answers
4 views

Problem with weights having the same shape[0] during forward propagation and preventing the dot product from working

I'm trying to implement forward propagation in my neural network but it doesn't seem to work and I suspect it is because the number of neurones from one layer to the next are not the same. ...
1
vote
1answer
25 views

Dying leaky ReLU

I am trying to train a deep neural network but I am having dying ReLU problem. I am using leaky Relu but still have the same problem. Isn't leaky relu supposed to not have such problems?
0
votes
0answers
6 views

Is class discrimination is possible in Class Activation Mapping(CAM)?

I worked with Gradient Weighted class activation mappings(Grad-CAM) to understand and implement interpretability in Deep neural networks. I can also switch to a particular class(by selecting the ...
1
vote
0answers
15 views

Preprocess problem Faster RCNN in tensorflow object detection API

I've wrapped meta-architecture with the code below: ...
1
vote
1answer
22 views

How to train a neural network on multiple objectives?

I have a multi-class neural network classifier that has K classes(products). For every row, only one of the classes will be 1 at a time. Now, this approach works fine if I have only 1 objective to ...
0
votes
1answer
36 views

How can I fixed the filter and Kernel Size of a CNN?

I have created 4 x 4 2d images from a signal. Now, I want to feed this data to a Convolutional neural Network. How I can choose the nubmber of ...
0
votes
1answer
15 views

Sneakers representation learning

I am trying to make a model which would take an image of shoes as an input and output a meaningful N-dimensional embedding of the shoes, so that they could be searchable/comparable/clustered and used ...
1
vote
0answers
32 views

For short sentences(max length 10 ), which Name entity recognition algorithm is good?

My Training data look like this . I have to recognize 4 class for each sentence. Any algorithm , which have some learning parameters Means not rule based approach . So which method is good for my ...
0
votes
1answer
25 views

Why this TensorFlow Transformer model has Linear output instead of Softmax?

I am checking this official TensorFlow tutorial on a Transformer model for Portuguese-English translation. I am quite surprised that when the Transformer is created, their final output is a Dense ...
0
votes
0answers
11 views

Training on compressed video, testing on uncompressed images, performance degradation?

In my application, I have to collect training data from a single camera. At test-time, the camera frames will be fed live to the network, without being saved in a lossy format in between. Now I wonder ...
0
votes
1answer
26 views

PyTorch cross_entropy with 3D data (RNN/LSTM)

I am working on LSTMs and I want to compute cross_entropy loss, given X and y. X.shape: (batch_size, time_steps, number_of_classes) y.shape: (batch_size, time_steps) y contains the ground truth ...
-1
votes
1answer
24 views

Can we identify that an academic dataset was used for commercial purpose [closed]

There are many datasets released on the internet. Authors of many of these datasets state that the datasets are strictly for academic usage and not for commercial purposes. Although some datasets are ...
0
votes
0answers
8 views

How to train an LSTM model with data that has multiple input rows per day but only one row of label/ground-truth (output) data per day

I am doing a sleep data science experiment and I need a model that outputs multiple columns sleep quality measurement values (that are decimal numbers) for each input. For training, I collected data ...
0
votes
0answers
18 views

What sort of models work for unsupervised reinforcement learning, or is deep learning the way?

I'm setting out on an adventure to automate the statuses of the lights around my home. The lights should have different brightness in the range [0, 100] depending on some factors, which I have boiled ...
2
votes
2answers
34 views

How to compute score and predict for outcome after N days

Let's say I have a medical dataset/EHR dataset that is retrospective and longitudinal in nature. Meaning one person has multiple measurements across multiple time points (in the past). I did post here ...
1
vote
2answers
21 views

Can I create a layer with multiple rnn cell ? [question about a paper]

I am trying to implement https://dl.acm.org/doi/pdf/10.1145/3269206.3271794 . Structure: As it said: In particular, we integrate the embedding vectors learned from each individual recurrent encoder ...
0
votes
0answers
6 views

Can we do autoregressive using pad_packed_sequence?

I’m curious about if we want to do the autoregressive manner. Is it possible to do with implemented using pad_packed_sequence and pack_padded_sequence input to some recurrent network? Because we need ...
1
vote
0answers
13 views

How do I create a dok matrix with split files correctly?

I let my model run normally and have defined an early stopping as a callback. The model breaks off, I let it run through without early stopping, loss and val_loss go further and further apart (see ...
0
votes
0answers
7 views

torch.save(the_model, PATH) vs torch.save(the_model.state_dict(), PATH) - model loading incorrectly for one method

I just now noticed that the model does not get loaded correctly if I use the the_model.state_dict() method to save it. On the other hand, using ...
0
votes
0answers
6 views

How to generate fixed number of superpixels?

A lot of work regarding Graph Neural Networks require fixed number of nodes. In the case of image processing using graph, the image representation is often super pixels (like in this work https://...
-1
votes
0answers
14 views

Is Transformer better then GRU for human acitivity recognition

I am working on human activity recognition and I was wondering If there are documented studies( Papers) about GRU vs Transformers in this context?
0
votes
0answers
10 views

Usage of Doc2Vec as feature extractor for text classification of websites with political articles

I have gathered political articles from polish websites for my engineering thesis. The main goal is to try to predict the website that input text belongs to. So for this few websites I want to create ...
4
votes
4answers
628 views

What is the difference between AI, ML, NN and DL? [closed]

What is the difference between the following four categories: Artificial Intelligence (AI) Machine Learning (ML) Neural Network (NN) Deep Learning (DL) Data Science My current understanding is that ...
0
votes
0answers
8 views

Help understanding input to biaxial network for generating music

I am reading Composing Music With Recurrent Neural Networks by Daniel D. Johnson. But I am really confused about the input passed to this network. If we pass notes of music along the time axis, then ...
0
votes
0answers
19 views

Is padding the right way to allow your model to make prediction with test sequences of shorter lengths?

Say I have a RNN-lstm encoder-decoder model trained on fixed timesteps (no padding when training, all sequences are treated as if having the same lengths). My testing criteria requires me to provide ...
1
vote
1answer
17 views

Working Behavior of BERT vs Transformers vs Self-Attention+LSTM vs Attention+LSTM on the scientific STEM data classification task?

So I just used BERT pre-trained with Focal Loss to classify Physics, Chemistry, Biology and Mathematics and got a good f-1 macro of 0.91. It is good given it only had to look for the tokens like ...
0
votes
0answers
31 views

Stateful LSTM in Deployment

Knowing the nature of my time series problem, I am using a stateful LSTM to forecast one step ahead. My question is quite straightforward. Do I need to explicitly save and pass the hidden cell in ...

1
2 3 4 5
70