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
7 views

GradientTape not computing gradient

I understand that so long as i am defining a computation in tf.GradientTape() context, the gradient tape would compute the gradient w.r.t all the variables that the ...
0
votes
0answers
4 views

Better understanding of Integrated Gradients

I've been trying to use Integrated Gradients to get a better understanding of the attribution of different features in my NN. I've read the original publication(https://arxiv.org/pdf/1703.01365.pdf) ...
1
vote
0answers
8 views

TensorFlow1.15, multi-GPU-1-machine, how to set batch_size?

The input function code: ...
1
vote
2answers
10 views

Is GNU Octave a perfect place to code neural networks

GNU Octave is used for its simplicity and compiling speed to write numerical algorithms (such as eg machine learning problems), but I wanted to know if I can also use it for faster coding of neural ...
0
votes
1answer
10 views

How to verify a CNN encoder works as expected?

I am using CNN as a part of kernel warping. The purpose here is to reduce input dimension (from N*M to K *1). The input data is not image data. I suspected that the CNN network might not work as I ...
1
vote
1answer
7 views

Neural network regression problem, integer output neuron constraint

The problem I'm solving is a regression problem using neural networks, and the "y" value covers a very large range (let's say y represents the number of people, ranging from 0 people to 10000 people), ...
0
votes
0answers
24 views

I need a dialogue dataset [migrated]

Where can I find a dataset of 2 people chit-chat dialogues? I need the data to be of full conversations, from "Hi" to "bye" so to speak.
1
vote
0answers
8 views

Why N-pair Loss (NIPS 2016) stops minimizing in Image retrieval task?

Currently, I am using a Deep Learning model to build a search engine for retrieving images. With a dataset of pairs of (image, description), I am using a ...
1
vote
0answers
10 views

how do deep Q network deal with varying input size?

I am conducting research with multiply agents in an environment. The main concept of my methodology is a centralized control system, which means we take the positions, as well as other information, of ...
2
votes
0answers
22 views

What's the best way to validate a rare event detection model during training?

When training a deep model for rare event detection (e.g. sound of an alarm in a home device audio stream), is it best to use a balanced validation set (50% alarm, 50% normal) to determine early ...
0
votes
1answer
21 views

Training accuracy is ~97% but validation accuracy is stuck at ~40%

I am trying to classify images into 27 classes using a Conv2D network. The training accuracy rises through epochs as expected but the val_accuracy and val_loss values fluctuate severely and are not ...
0
votes
1answer
13 views

How to preprocess data for image classification from a .txt file?

Basically my issue is that im building an image classification model using AlexNet. I have this pre-split dataset thats already split into training, test, validation. However the issue is that these ...
0
votes
1answer
17 views

Don't know how to preprocess my dataset for image classification

I'm trying to do image classification using CNN. The exact model isn't important but I decided to try use AlexNet and I'm getting abysmal accuracy. I believe the issue might be with my data ...
0
votes
0answers
9 views

How to decide whether to use categorical embeddings in a neural network?

I have a binary classification task with a whole slew of binary categorical features, one multiclass categorical and a few continuous features. I initially treated the categorical features using one-...
-1
votes
0answers
11 views

How Reconstructing the Original Image and Cycle Consistency Constraints help the GAN in image to Image Transalation

Currently, I am reading few papers related to image to image translation using Generative Adversarial Network. What I have found is that many papers involves Reconstructing the same original image ...
0
votes
0answers
10 views

What's the best way to store and then call 3 Million+ records for training?

Good afternoon all, I am working on a Deep Learning project where I'm proceeding in batches (rather than a continuous pipeline). Essentially I've built a function that can convert raw data into ...
0
votes
0answers
5 views

Need for doing grad.zero_() after setting torch.no_grad() in pytorch

In the following lines of code, ...
1
vote
1answer
21 views

How to tune learning rate with HParams Dashboard on Tensorflow?

In Tensorflow documentation, it is shown how to tune several hyperparameters but not the learning rate.I have searched how to tune learning rate using HParams dashboard but could not find much. The ...
0
votes
0answers
4 views

Data points are highly overlapped and do not follow smoothness rule assumption

I am working on a very high dimensional categorical features based data set. There are two output classes and 2-dimensional PCA plot suggests that the data points belonging to both +ve and -ve classes ...
-1
votes
0answers
11 views

Training dice coefficient is increasing but validation dice coefficient remains constant at zero? [closed]

I am training an U-Net model for image segmentation. The model takes RGB (3-channel) images as input and outputs a segmented binary image (2-channel). The model is working well and delivering the dice ...
0
votes
0answers
13 views

Is this Tensorflow 1.x network get trained?

I've just started to learn Tensorflow 2.1.0 with Keras 2.3.1 and Python 3.7.7. I have found the following code from this "Omniglot Character Set Classification Using Prototypical Network": ...
0
votes
0answers
17 views

Can I used a pre-trained auto encoder as an embedding layer in the model?

The dataset has an ordinal target variable [0,1,2]. Each observation in the dataset has multiple time series. In addition, each observation have various tabular features. The purpose is to build a ...
0
votes
0answers
11 views

Using Keras fit_generator for functional keras models and custom dataset

I have to fit a model that takes three discrete inputs and produces two discrete outputs using a generator made as follows: ...
0
votes
0answers
16 views

Splitting a 10 year long time-series into multiple year time-series on Deep Learning Models

I'm using recent Deep Learning models for time series analysis such as DeepAR[1] and DeepFactors[2] for my masters. My target time series was given to me by a cement factory, 10 years of compositions ...
0
votes
0answers
8 views

Batch Normalization as input layer to learn an optimal scaling?

We all know that Batch Normalization reduces "Internal Covariance Shift" and therefore helps Neural Networks to train faster (Batch Normalization: Accelerating Deep Network Training by Reducing ...
0
votes
0answers
11 views

How to deal with images with textual noise?

I have a dataset of images collected from google and bing images (scraped). basically I want to classify these images into binary classes (positive, negative). Images that contain a text originally ...
0
votes
0answers
9 views

How good is batch normalization in avoiding vanishing/exploding gradients?

Incase of deep neural network with many layers, how good is batch normalization in avoiding vanishing/exploding gradients. Consider the problem is only due to too many layers(too many multiplications ...
0
votes
0answers
33 views

Using word embeddings for kaggle?

Not sure, if this is the right forum so redirect me if it wrong. I have started on an NLP problem in kaggle. There i have word embeddings from google news, wiki, glove in a zipped folder. I want to ...
2
votes
0answers
12 views

Xavier initialisation vs He initialisation

After reading the famous paper, Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classification, I understand two things:- He initilization borrows on the benefits of ...
0
votes
0answers
18 views

How to estimate the OCR accuracy

I am building a system which uses ocr to extract text but i have no way to flag that information on how correct it can be and if the information needs to be discarded by just looking at the image and ...
0
votes
1answer
20 views

How to use fine tuning of BERT when i have unlabelled dataset of text documents?

I have gained a basic understanding of using BERT for various NLP/text mining tasks. When it comes to fine-tuning of BERT, I always see that fine-tuning is performed using some classification tasks. ...
0
votes
1answer
41 views

Data augmentation for multiple output heads in Keras

I have a transfer learning based two output classification problem. So, accordingly, I have formatted my data to have X_train as a ...
0
votes
0answers
6 views

Approaches For Recommender System Using Complicated Novel Dataset

I have a question about the best approach(s) I should take in building a recommender system for a project I'm working on. I have created a dataset. The dataset has the following: 400,000 users For ...
0
votes
0answers
11 views

Skip gram model on multiple sequence

From different examples, I have seen that getting an embedding from skip-gram model on a single sequence or a single corpus. However, if I have multiple sequences of same word or phases. How can I use ...
0
votes
0answers
6 views

How exactly the hidden state works in an RNN ? How to decide on how many past instances to consider?

I am unable to grasp the working of RNN because in different tutorials, it is explained differently. Please correct me as I have considered that: In a ...
1
vote
2answers
18 views

Implementing training in PyTorch

I wish to accomplish the following task in PyTorch- I have the COCO dataset, wherein each data sample is used in training YOLO v3. After being processed by the model, the sample is to be deleted if ...
-1
votes
0answers
4 views

which parameter should i feed to the next RBM as it's visible layer in a Deep Belief Network?

i'm currently building a Deep Belief Network by stacking several RBMs on top of each other. The 1st RBM takes standardized image pixel values (float between 0 and 1) as it's visible layer. But ...
0
votes
0answers
8 views

Low GPU utilisation and High GPU memory

I am training a conv net for classifying 3 classes of images of size 512,512 using Pytorch framework. I have 3 Tesla V100s(16 Gb). My GPUs utilization is really low - <10% and GPU memory is really ...
0
votes
0answers
24 views
+50

BPR TripletLoss Recommender System

I am trying to modify the code of this repo to build a recommender system based on BPR triplet loss. In particular I modified the TripletLoss layer class like this ...
0
votes
1answer
26 views

I need some guidance how to develop myself in deep learning [closed]

Few months ago I've taken an Andrew Ng course at coursera and started my data science/machine learning journey. During those months, I've take another courses from the internet - about Deep Learning, ...
1
vote
0answers
15 views

What does anneal mean in the context of machine learning?

An article released by Open AI gives an overview of how Open AI Five works. There is a paragraph in the article stating: Our agent is trained to maximize the exponentially decayed sum of future ...
0
votes
2answers
37 views

Tensorflow keras fit - accuracy and loss both increasing drastically

ubuntu - 20.04 tensorflow 2.2 dataset used = MNIST I am testing tensorflow and i notice that validation sparse_categorical_accuracy (accuracy) and validation <...
0
votes
0answers
7 views

Untrained Network Giving 80% Accuracy

I have a two class classification problem and my neural network prior to training predicts with an accuracy of 80%. After training i have an accuracy of 75%. Can you tell me how this is possible? ...
0
votes
0answers
6 views

What is difference in Neural Netowork if we train any object detector like Yolo or Mobilenet, for 1 class and 80 classes? [closed]

I want to know If I train a model (Lets say Yolo) for 1(Person) class and other with 80(including Person) classes, does the accuracy of model with 1 class would be greater than that of model with 80 ...
1
vote
2answers
51 views

Image Classification low accuracy

I have a dataset that has two folders for training and testing. I am trying to determine whether a patient has an eye disease or not. However, the images I have are hard to work with. I've ran this ...
1
vote
0answers
12 views

Generate text using user-supplied keywords

I've got a use case where I need to generate sentences based on a set of user supplied keywords. Here is an example of what I need: User input: End-User: Data Scientists Region: Middle East ...
4
votes
0answers
34 views

Robustness of hyperparameter tuning

I use a Bayesian hyperparameter (HP) optimization approach (BOHB) to tune a deep learning model. However, the resulting model is not robust when repeatedly applied to the same data. I know, I could ...
0
votes
0answers
16 views

attention mechanism [closed]

has anybody here any knowledge about attention mechanism? did anybody probably write a master thesis about this or so? I'm looking for some insights regarding theme in this area for a master thesis. ...
1
vote
0answers
15 views

How to use multiple text features for NLP classifier?

I am trying to build text classifier, Usually, we have one text column and ground truth. But I am working on a problem where dataset contains many text features. I am exploring different ways how to ...
1
vote
1answer
14 views

How do I build a DQN which selects the correct objects in an environment based on the environment state?

I have an environment with 4 objects in it. All of these objects can either be selected or not selected. So the actions taken by my DQN should look like - ...

1
2 3 4 5
64