hola
  • Member for 2 years
  • Last seen more than a month ago
  • France
What are the state-of-the-art meta-reinforcement learning methods?
Accepted answer
3 votes

One of the most recent papers on meta-RL is meta-Q-learning. This paper introduces Meta-Q-Learning (MQL), a new off-policy algorithm for meta-reinforcement learning (meta-RL). MQL builds upon three ...

View answer
Have GANs been used to solve regression problems?
3 votes

In reality GANs are not made for image classification, but for data generation, and they have gained popularity on image generation. They are also used for tabular data generation, see for example ...

View answer
How are newer weight initialization techniques better than zero or random initialization?
3 votes

There are several ways to answer this question. First of all, there are several mathematical arguments on why using some kind of initialization is better. Consider reading, for example, Xavier et al.. ...

View answer
How to make DNN learn multiplication/division?
Accepted answer
3 votes

In reallity any continous function on a compact can be approximated by a neural network having one hidden layer with a finite number of neurones (This is the Universal Approximation Theorem). Thus you ...

View answer
Should I apply image processing techniques to the inputs of convolution networks?
Accepted answer
2 votes

The whole interest of using deep learning-based solutions is that you don't have to do all those pre-processings, i.e. binarization, segmentation of background. CNNs, such as YOLO or FasterRCNN, can ...

View answer
Solving the supervised learning problem of learning $p(y \vert \mathbf{x})$ by using traditional unsupervised technologies to learn $p(\mathbf{x}, y)$
Accepted answer
2 votes

This is the definition of conditional probability + Total probability decomposition formula: $p(y|x) = \frac{p(y,x}{p(x)} = \frac{p(x,y)}{\sum_{y'}p(x,y')}$. The idea is to use some unsupervised ...

View answer
Why do we average gradients and not loss in distributed training?
2 votes

The whole idea behind those distributed optimization methods is that data should be local in every node/worker. Thus, if you only send the loss value to the central node, this node can't compute the ...

View answer
Do Seq2Seq models and the Bidirectional RNN do the same thing?
1 votes

Seq2Seq and Bidirectional RNNs are not doing the same thing, at least in their classic form. Seq2Seq models are used to generate a sequence from another sequence. Consider, for example, the ...

View answer
Using sigmoid in LSTM network for multi-step forecasting
1 votes

You have a problem in your code, you want to use "sigmoid" in the last layer. Fot the code you are showin you are using linear activation in the last layer.

View answer
Strategy of using intermediate layers of a neural network as features?
1 votes

This a typical transfer learning technique, a lot of people refer to it with fine-tunning. I would recommend that you have a look on PyTorch tutorial: it explains well how to use.

View answer
What kind of optimizer is suggested to use for binary classification of similar images?
0 votes

The fact that images are similar to each other or the fact that you are using binray classification, don't give you a particular choice of Optimizer, when an optimization algorithm is developped, ...

View answer
Are embeddings in multi-lingual language models comparable across languages?
-1 votes

There is a general idea in the field of NLP that there is a mapping between embeddings in different langauges. Figure 1 explains this. In Figure 1. we have the embedding of English words and Spanish ...

View answer