I'm learning about how helpful it is to have your variables in a gaussian distribution. It's not very clear to me how to apply this to a supervised learning problem, say, using a neural network. The following is the target variable for my dataset, the sqrt and log of the target variable respectively.
Question
Is there any benefit/point is transforming a continuous target variable in a regression problem?
Sometimes. Some algorithms, for ex. ANN, work with gradients, and in this case having a better behaving variable works in favor of the optimization. However, other times, for ex. with linear models, scaling won't affect the model fit.
A lot of people unnecessarily transform their variables to be normal because of a false belief that this will improve the model. Unless the method requires normal data, or the optimization algorithm will work better because of it, this belief is unfounded.
Related
I read this line today :
Every regression gets better with the addition of more features or variables... But adding more features increases complexity and reduces interpretability of the model as well.
I am unable to understand what is interpretability? (searched it on google but still did not get it)
Please help thank you
I would say that interpretability in a regression problems is when you can explain the result of your model to non statistician / domain experts.
For example: you try to predict the size of people depending on many variable, including sex. If you use linear regression, you will be able to say that the model will add 20cm (again, for example) to the predicted size if the person is a man (compared to a woman). The domain expert will understand the relationship between explanatory variable and the predicted result, without understanding statistics or how a linear regression works.
In addition, I disagree with the fact that the addition of more features or variables always improve regression result.
What is a better regression ? Improvement in choosen metrics ? For training or test set ? A "better regression" doesn't mean anything...
If we assume that a better regression is a regression which is better to predict the target for a new dataset, more variable doesn't always improve prediction power, especially when there is no regularization, if the added feature contains futures variables or many others cases.
Many of the papers I have read so far have this mentioned "pre-training network could improve computational efficiency in terms of back-propagating errors", and could be achieved using RBMs or Autoencoders.
If I have understood correctly, AutoEncoders work by learning the
identity function, and if it has hidden units less than the size of
input data, then it also does compression, BUT what does this even have
anything to do with improving computational efficiency in propagating
error signal backwards? Is it because the weights of the pre
trained hidden units does not diverge much from its initial values?
Assuming data scientists who are reading this would by theirselves
know already that AutoEncoders take inputs as target values since
they are learning identity function, which is regarded as
unsupervised learning, but can such method be applied to
Convolutional Neural Networks for which the first hidden layer is
feature map? Each feature map is created by convolving a learned
kernel with a receptive field in the image. This learned kernel, how
could this be obtained by pre-training (unsupervised fashion)?
One thing to note is that autoencoders try to learn the non-trivial identify function, not the identify function itself. Otherwise they wouldn't have been useful at all. Well the pre-training helps moving the weight vectors towards a good starting point on the error surface. Then the backpropagation algorithm, which is basically doing gradient descent, is used improve upon those weights. Note that gradient descent gets stuck in the closes local minima.
[Ignore the term Global Minima in the image posted and think of it as another, better, local minima]
Intuitively speaking, suppose you are looking for an optimal path to get from origin A to destination B. Having a map with no routes shown on it (the errors you obtain at the last layer of the neural network model) kind of tells you where to to go. But you may put yourself in a route which has a lot of obstacles, up hills and down hills. Then suppose someone tells you about a route a a direction he has gone through before (the pre-training) and hands you a new map (the pre=training phase's starting point).
This could be an intuitive reason on why starting with random weights and immediately start to optimize the model with backpropagation may not necessarily help you achieve the performance you obtain with a pre-trained model. However, note that many models achieving state-of-the-art results do not use pre-training necessarily and they may use the backpropagation in combination with other optimization methods (e.g. adagrad, RMSProp, Momentum and ...) to hopefully avoid getting stuck in a bad local minima.
Here's the source for the second image.
I don't know a lot about autoencoder theory, but I've done a bit of work with RBMs. What RBMs do is they predict what the probability is of seeing the specific type of data in order to get the weights initialized to the right ball park- it is considered an (unsupervised) probabilistic model, so you don't correct using the known labels. Basically, the idea here is that having a learning rate that is too big will never lead to convergence but having one that is too small will take forever to train. Thus, by "pretraining" in this way you find out the ball park of the weights and then can set the learning rate to be small in order to get them down to the optimal values.
As for the second question, no, you don't generally prelearn kernels, at least not in an unsupervised fashion. I suspect that what is meant by pretraining here is a bit different than in your first question- this is to say, that what is happening is that they are taking a pretrained model (say from model zoo) and fine tuning it with a new set of data.
Which model you use generally depends on the type of data you have and the task at hand. Convnets I've found to train faster and efficiently, but not all data has meaning when convolved, in which case dbns may be the way to go. Unless say, you have a small amount of data then I'd use something other than neural networks entirely.
Anyways, I hope this helps clear some of your questions.
The motivating idea behind neural nets seems to be that they learn the "right" features to apply logistic regression to. Is there a similar approach for linear regression? (or just regression problems in general?)
Would doing the obvious thing of removing the application of a sigmoid function for all neurons (ie, including the hidden layers) make sense/work? (ie, each neuron is performing linear regression instead of logistic regression).
Alternatively, would doing the (maybe even more obvious) thing of just scaling output values to [0,1] work? (intuitively I would think not, as the sigmoid function seems like it would cause the net to arbitrarily favor extreme values) (edit: though I was just searching around some more, and saw that one technique is to scale based on mean and variance, which seems like it might deal with this issue -- so maybe this is more viable than I thought).
Or is there some other technique for doing "feature learning" for regression problems?
Check out this applet. Try to learn different functions. When you dictate linear activation functions at both hidden and output layers, it even fails to learn the quadratic function. At least one layer needs to be set to sigmoid function, see figures below.
There are different kinds of scaling. Standard scaling, as you mentioned, eliminates the impact of mean and standard deviation of the training sample, is most often used in machine learning. Just make sure you are using the same mean and std value from training sample in the test sample.
The reason why scaling is required is because the output of sigmoid function ranges at (0,1). I didn't try, but I think it is better to scale the output even if you select linear function at output layer. Otherwise large input at hidden layer (with sigmoid) won't lead to drastic output (the sigmoid function is approximately linear when the input is at a small range, out of such range will make the output changes much slowly). You can try this by yourself in your own data.
Besides, if you have various features, the feature normalization that makes different features in the same scale is also recommended. The scaling speeds up gradient descent by avoiding many extra iterations that are required when one or more features take on much larger values than the rest.
As #Ray mentioned, deep learning that many levels of features are involved can help you with the feature learning, it's not all linear combinations though.
I was learning about different techniques for classification, like probablistic classifiers etc , and stubled upon the question Why cant we implement a binary classifier as a Regression function of all the attributes and classify on the basis of the output of the function , say if the output is less than a certain value it belongs to class A , else in class B . Is there any limitation to this method compared to probablistic approach ?
You can do this and it is often done in practice, for example in Logistic Regression. It is not even limited to binary classes. There is no inherent limitation compared to a probabilistic approach, although you should keep in mind that both are fundamentally different approaches and hard to compare.
I think you have some misunderstanding in classification. No matter what kind of classifier you are using (svm, or logistic regression), you can always view the output model as
f(x)>b ===> positive
f(x) negative
This applies to both probabilistic model and non-probabilistic model. In fact, this is something related to risk minimization which results the cut-off branch naturally.
Yes, this is possible. For example, a perceptron does exactly that.
However, it is limited in its use to linearly separable problems. But multiple of them can be combined to solve arbitrarily complex problems in general neural networks.
Another machine learning technique, SVM, works in a similar way. It first transforms the input data into some high dimensional space and then separates it via a linear function.
I'm having trouble with some of the concepts in machine learning through neural networks. One of them is backpropagation. In the weight updating equation,
delta_w = a*(t - y)*g'(h)*x
t is the "target output", which would be your class label, or something, in the case of supervised learning. But what would the "target output" be for unsupervised learning?
Can someone kindly provide an example of how you'd use BP in unsupervised learning, specifically for clustering of classification?
Thanks in advance.
The most common thing to do is train an autoencoder, where the desired outputs are equal to the inputs. This makes the network try to learn a representation that best "compresses" the input distribution.
Here's a patent describing a different approach, where the output labels are assigned randomly and then sometimes flipped based on convergence rates. It seems weird to me, but okay.
I'm not familiar with other methods that use backpropogation for clustering or other unsupervised tasks. Clustering approaches with ANNs seem to use other algorithms (example 1, example 2).
I'm not sure which unsupervised machine learning algorithm uses backpropagation specifically; if there is one I haven't heard of it. Can you point to an example?
Backpropagation is used to compute the derivatives of the error function for training an artificial neural network with respect to the weights in the network. It's named as such because the "errors" are "propagating" through the network "backwards". You need it in this case because the final error with respect to the target depends on a function of functions (of functions ... depending on how many layers in your ANN.) The derivatives allow you to then adjust the values to improve the error function, tempered by the learning rate (this is gradient descent).
In unsupervised algorithms, you don't need to do this. For example, in k-Means, where you are trying to minimize the mean squared error (MSE), you can minimize the error directly at each step given the assignments; no gradients needed. In other clustering models, such as a mixture of Gaussians, the expectation-maximization (EM) algorithm is much more powerful and accurate than any gradient-descent based method.
What you might be asking is about unsupervised feature learning and deep learning.
Feature learning is the only unsupervised method I can think of with respect of NN or its recent variant.(a variant called mixture of RBM's is there analogous to mixture of gaussians but you can build a lot of models based on the two). But basically Two models I am familiar with are RBM's(restricted boltzman machines) and Autoencoders.
Autoencoders(optionally sparse activations can be encoded in optimization function) are just feedforward neural networks which tune its weights in such a way that the output is a reconstructed input. Multiple hidden layers can be used but the weight initialization uses a greedy layer wise training for better starting point. So to answer the question the target function will be input itself.
RBM's are stochastic networks usually interpreted as graphical model which has restrictions on connections. In this setting there is no output layer and the connection between input and latent layer is bidirectional like an undirected graphical model. What it tries to learn is a distribution on inputs(observed and unobserved variables). Here also your answer would be input is the target.
Mixture of RBM's(analogous to mixture of gaussians) can be used for soft clustering or KRBM(analogous to K-means) can be used for hard clustering. Which in effect feels like learning multiple non-linear subspaces.
http://deeplearning.net/tutorial/rbm.html
http://ufldl.stanford.edu/wiki/index.php/UFLDL_Tutorial
An alternative approach is to use something like generative backpropagation. In this scenario, you train a neural network updating the weights AND the input values. The given values are used as the output values since you can compute an error value directly. This approach has been used in dimensionality reduction, matrix completion (missing value imputation) among other applications. For more information, see non-linear principal component analysis (NLPCA) and unsupervised backpropagation (UBP) which uses the idea of generative backpropagation. UBP extends NLPCA by introducing a pre-training stage. An implementation of UBP and NLPCA and unsupervised backpropagation can be found in the waffles machine learning toolkit. The documentation for UBP and NLPCA can be found using the nlpca command.
To use back-propagation for unsupervised learning it is merely necessary to set t, the target output, at each stage of the algorithm to the class for which the average distance to each element of the class before updating is least. In short we always try to train the ANN to place its input into the class whose members are most similar in terms of our input. Because this process is sensitive to input scale it is necessary to first normalize the input data in each dimension by subtracting the average and dividing by the standard deviation for each component in order to calculate the distance in a scale-invariant manner.
The advantage to using a back-prop neural network rather than a simple distance from a center definition of the clusters is that neural networks can allow for more complex and irregular boundaries between clusters.