New answers tagged

0

There is sadly very little theory that is useful for predicting the behavior of neural networks. Also, their experiment tends to be dependent on the particular workload/task you are trying to solve it. While I know it's not what you were hoping to hear, I suggest that you try it in empirical experiments.


0

As far as I understand you are trying to infer 3D position of the various points (i.e. height) on the beach using single monocular RGB camera. I will try to give the best answer that with the CV knowledge that I have. This is a difficult problem because you do not have the depth (distance from the camera) information, so you cannot infer full 3D structure ...


0

A general rule of thumb would be: look at the most successful architectures for your given task, take it as a prior and try to vary (by retraining) the number of layers if you need a specific point in the accuracy-speed tradeoff curve. Previously people believed that one can get more accuracy just by stacking more layers. Even though this is true in most ...


1

I would expect most of the works use Generative Adversarial Networks (GANs) for this because they are powerful generative models capable of learning the complex underlying probability distribution. In this amazing work the authors used a Conditional GAN, in which they can generate an image conditioned on semantic segmentation map. In your case, you might ...


0

Try changing your input in this case if you wanted to take only the difference of two states into consideration. Think of each rounds in a discrete manner, they will be of their own distinctive state if not encountered the exactly same configuration of the board beforehand. What do you get? You get another set of states that takes differentiated states into ...


1

The value of $C = \sqrt 2$ was shown to ensure the asymptotic optimality when rewards are in the $[0,1]$ range (Kocsis, Szepesvári, 2006). In many games, that reward range is straightforward: maximum and minimum possible scores can be translated to 0 and 1 (0 could mean a loss and 1 a win). The accuracy of this squashing seems to have a minimal impact (...


-1

Don't train by first training all the negatives or first training all the positives. Instead, randomize the order in which examples are used during training. Moreover, you should use a different random order in each epoch. When you use a non-random order, catastrophic forgetting can cause performance to suffer, as you have observed. This is not specific ...


-1

RNNs are pattern recognition tools. It is't entirely clear to me what it is exactly that you are trying to do, but if you simply intend for it to classify positive and negative messages a regular Neural Network might be better suited. What your RNN does (if implemented correctly) is learn classifications in the context of the sequence. i. e. It is learning: ...


Top 50 recent answers are included