Djib2011
  • Member for 2 years, 7 months
  • Last seen more than a week ago
Why do we need explainable AI?
Accepted answer
73 votes

As argued by Selvaraju et al., there are three stages of AI evolution, in which interpretability is helpful. In the early stages of AI development, when AI is weaker than human performance, ...

View answer
How could artificial intelligence harm us?
Accepted answer
52 votes

tl;dr There are many valid reasons why people might fear (or better be concerned about) AI, not all involve robots and apocalyptic scenarios. To better illustrate these concerns, I'll try to split ...

View answer
Are neural networks prone to catastrophic forgetting?
21 votes

Yes, the problem of forgetting older training examples is a characteristic of Neural Networks. I wouldn't call it a "flaw" though because it helps them be more adaptive and allows for interesting ...

View answer
Is there a way to understand neural networks without using the concept of brain?
11 votes

tl;dr I always like to think of Neural Networks as a generalization of logistic regression. I too don't like that, traditionally, when introducing Neural Networks, books start with biological ...

View answer
What is the meaning of $V(D,G)$ in the GAN objective function?
Accepted answer
9 votes

To understand this equation first you need to understand the context in which it is first introduced. We have two neural networks (i.e. $D$ and $G$) that are playing a minimax game. This means that ...

View answer
What do the words "coarse" and "fine" mean in the context of computer vision?
6 votes

tl;dr What does that mean in the context of this paper? With "coarse segmentation" the author means a segmentation that doesn't have much detail. "Fine segmentation", on the other hand, refers to ...

View answer
What is the best measure for detecting overfitting?
Accepted answer
5 votes

tl;dr The safest method I've found is to use cross-validation for hyperparameter selection and a hold-out test set for a final evaluation. Why this isn't working for you... In your case, I suspect you'...

View answer
What is the “Hello World” problem of Unsupervised Learning?
Accepted answer
5 votes

I disagree with the context that MNIST is the "hello world" of supervised learning. It is definitely, though, the "hello world" of image classification, which is a very specific ...

View answer
What is the purpose of the batch size in neural networks?
5 votes

tl;dr: A batch size is the number of samples a network sees before updating its gradients. This number can range from a single sample to the whole training set. Empirically, there is a sweet spot in ...

View answer
Are deep learning models more prone to overfitting than machine learning ones?
Accepted answer
5 votes

Your reasoning isn't wrong. Deep Neural Networks (DNNs) have a much larger capacity than simpler ML algorithms (excluding NNs) and can easily memorize even a very complex dataset and overfit. DNNs, ...

View answer
Can we optimize an optimization algorithm?
Accepted answer
5 votes

First, you need to consider what are the "parameters" of this "optimization algorithm" that you want to "optimize". Let's take the most simple case, a SGD without momentum. The update rule for this ...

View answer
How should I deal with variable-length inputs for neural networks?
4 votes

The most common way people deal with inputs of varying length is padding. You first define the desired sequence length, i.e. the input length you want your model to have. Then any sequences with a ...

View answer
Are there any better visual models for transfer rather than ImageNet?
4 votes

Why is ImageNet so popular for transfer learning? Models pre-trained on the ImageNet datasets have been the de-facto choice for many years now. Many popular reasons as to why people think that ...

View answer
Can neurons in MLP and filters in CNN be compared?
Accepted answer
3 votes

tl;dr The equivalent to a neuron in a Fully-Connected (FC) layer is the kernel (or filter) of a Convolution layer Differences The neurons of these two types of layers have two key differences. These ...

View answer
How does the neural-network know how to tweak weights for a specific neuron?
3 votes

tl;dr The whole point of gradient descent is to assess the contribution of each parameter towards the loss. This information is uncovered through the gradient of the loss w.r.t each parameter. A ...

View answer
What is the use of the regular convolutional layer in expansion path of U-Net?
3 votes

The point is that in the expansive path you have two forms of information: the information from the contracting path, which includes all high-level features extracted from the original image. the ...

View answer
Can TensorFlow, PyTorch, and other mainstream ML frameworks be used for research-grade work in AI?
Accepted answer
3 votes

Your statement that researchers build their network from the ground-up using C++ or some other low level library couldn't be further from the truth. You could take a look at this analysis showing the ...

View answer
How does the generator in GAN's work?
Accepted answer
3 votes

What's the input to the Generator? In the basic implementation of GANs, the Generator only takes in a vector of random variables. This might seem strange, but after training, the generator can ...

View answer
What is the difference between model and data distributions?
3 votes

Yes. In Machine Learning we consider that the samples in your training set are sampled from an underlying distribution called the data generating distribution. Generative models classify the samples ...

View answer
How many layers exists in my neural network?
3 votes

tl;dr I'd say your model has 8 layers (5 conv, 3 dense), however a lot of people count layers in other ways. From what I've seen this is by far the most conventional way for counting layers. ...

View answer
Is unsupervised learning a branch of AI?
Accepted answer
3 votes

There is a problem with confining Artificial Intelligence to a single definition, because it has become an umbrella term encompassing many fields of science. It has come a long way from the "thinking ...

View answer
Are feature maps merged or are they passed on as they are?
Accepted answer
3 votes

tl;dr It helps to think that the channels dimension of a convolutional layer works like a fully connected layer (i.e. the layer computes the weighted sum over all channels). For a single pixel... Let'...

View answer
What is eager learning and lazy learning?
2 votes

What is eager learning or lazy learning? Eager learning is when a model does all its computation before needing to make a prediction for unseen data. For example, Neural Networks are eager models. ...

View answer
Is there any difference between the convolution operation applied to images and applied to other numerical 2D data?
2 votes

Short answer is no. You can't use a model trained for one task to predict on a totally different task. Even if the second task was another image classification task, the CNN would have to be fine ...

View answer
What does end-to-end training mean?
2 votes

This is relevant when you have two or more neural networks serving as components to a larger architecture. Training this architecture in an end-to-end manner means simultaneously training all ...

View answer
What are the features get from a feature extraction using a CNN?
Accepted answer
2 votes

You get what we call high-level features, which are basically abstract representations of the parts that carry information in the image you want to classify. Imagine you want to classify a car. The ...

View answer
Turn photos right-side up?
2 votes

I don't know if there is an existing pretrained NN that does this but it wouldn't be very hard to modify one to do this. First, I'd take a pretrained image classification NN (e.g. VGG, ResNet), drop ...

View answer
Are the training loss and validation loss plotted per sample or per batch?
Accepted answer
2 votes

You want to compute the mean loss over all batches. What you need to do is to divide the sum of batch losses with the number of batches! In your case: You have a training set of $21700$ samples and ...

View answer
Query regarding the minmax loss function formulation of the training of a Generative Adversarial Network (GAN)
Accepted answer
2 votes

I'll answer your questions one by one: In this equation are the $E_{z \sim p_z(z)}$ and $E_{x \sim p_{data}(x)}$ the means of the distributions of the mini batch samples? So let's take the first ...

View answer
How are exploding numbers in a forward pass of a CNN combated?
Accepted answer
2 votes

The most effective way to prevent both the forward and backward propagation of exploding is keeping the weights in a small range. The main way this is accomplished is through their initialization. ...

View answer