# Tag Info

3

Try Rectification Improve the features available to your model, Remove some of the NOISE present in the data. In audio data, a common way to do this is to smooth the data and then rectify it so that the total amount of sound energy over time is more distinguishable. # Rectify the audio signal audio_rectified = audio.apply(np.abs) You can also calculate ...

2

Nope! Our number of coefficients will be driven by the vocabulary, and we'll use each of those 10K samples to estimate values for those coefficients - so, 'just' 100K samples. However, word frequency in human languages follows a Zipf distribution => most of those words will be rare, seen in only a few samples (=> won't even be able to determine whether ...

2

In my experience, GANs work really well for the scenario of semi-supervised learning, where you don't necessarily have labels for all your class $B$ data, but you do have a balanced dataset. In my (limited) experience, you do have to have a balanced (in numbers) set of $A$ and $B$ objects, even though you are not sure of the labels. And yes, GANs can overfit ...

2

The short answer is no, you shouldn't do that. There is a "distribution shift" thing when you have different x-y relation on the validation set then on the train set. The distribution shift would deteriorate your model performance and you should try to avoid that. The reason it's bad - ok, you find the way to fix the model for validation data, but ...

2

If I understood things correctly: You have a task which you need to estimate two values, gender and age. Your question revolves about the difference between networks which share layers for both inputs, whether the shared layers should be followed independent linear layers. Firstly, using shared layers in the networks of two related tasks may be useful to ...

2

If you want to determine if something is either a cat/dog or neither you need 2 classes: one for dog or cat, and one for anything else. However, if you assign all cats and dogs to the same class $A$, if an input is classified as $A$, then you won't be able to know whether it is a dog or a cat, you will just know that it is either a dog or a cat. In case ...

2

The hinge loss/error function is the typical loss function used for binary classification (but it can also be extended to multi-class classification) in the context of support vector machines, although it can also be used in the context of neural networks, as described here. The hinge loss function is defined as follows $$\ell(y) = \max(0, 1-t \cdot y) \tag{... 2 It is important to note that the exact statement is the eqation given below can never be 0 for misclassified points in  S^+$$ E(X) = (y - \text{sign}\{\overline{W} \cdot \overline{X}\}) $$And S+ is defined as the set of all misclassified training points X \in S that satisfy the condition y(\overline{W} \cdot \overline{X})<0  which means that y ... 1 If you find the Hessian matrix (the matrix of second order derivatives) for the binary cross entropy loss function, you'll see that it is positive semidefinite for any possible value of the parameters. This concludes that it is a convex function. A side effect of it being convex is that it will have a single minimum as mentioned in the textbook you cited. 1 So actually I managed to get hold of my lecturer to explain the argmax to argmin conversion. Generally speaking maximising \frac{1}{||w||} is identical to minimising ||w||. As ||w|| in \frac{1}{||w||} decreases, the overall value increases, i.e. we maximise it. The reason for choosing \frac{1}{2}||w||^2 turns out to be a less mathematic and more ... 1 Let's first recap the definition of the binary cross-entropy (BCE) and the categorical cross-entropy (CCE). Here's the BCE (equation 4.90 from this book)$$-\sum_{n=1}^{N}\left( t_{n} \ln y_{n}+\left(1-t_{n}\right) \ln \left(1-y_{n}\right)\right) \label{1}\tag{1}, where $t_{n} \in\{0,1\}$ is the target $y_n \in [0, 1]$ is the prediction (as produced by ...

1

You can try using Fourier basis functions to transform your observable variables and then apply a general linear regression model. To clarify, if you have pairs of observables $(y_i, x_i)$ where $y_i$ is $i$-th output, and $x_i$ is $i$-th input then you can transform your input variable into vector \phi = [1, \sin(x), \cos(x), \sin(2x), \cos(...

1

Yes, due to this issue, you should use temperature scaling after training your model. It will calibrate your probability and you will start to get the same kind of distributions. Here are a good article along with implementation on it.

1

So you have a network pretrained on 80 classes. I also assume that one of these classes are human (or else this is just not the way to go*) I suspect that the final layer contains 80 labels, correct? Then you then 'rescale' this layer to 1 label and then train on some data you possess? Then you're basically trying to teach the network that it shouldn't care ...

1

As far as generalization error is concerned, you are better off by learning the data distribution of (A and B) classes using unsupervised criterion. If you capture the underlying factors that explain most of the variations belong to A and B classes, after that, fine-tune it using a supervised criterion. in this way if you used two classes one for (A or B)...

1

The best approach may be to have a cat, dog, and neither class (3 classes total) and go with a regression approach — specifically, outputting the probabilities of each class for any given input. From there, you can always take the probabilities of each output and derive the probability of a cat and dog class or neither class. Also, make sure you use the ...

Only top voted, non community-wiki answers of a minimum length are eligible