Closed. This question does not meet Stack Overflow guidelines. It is not currently accepting answers.
This question does not appear to be about programming within the scope defined in the help center.
Closed last year.
Improve this question
I asked a question earlier that might have been too specific so I'll ask again in more general terms. How does error get propagated backwards through a pooling layer when there are no weights to train? In the tensorflow video at 6:36 https://www.youtube.com/watch?v=Y_hzMnRXjhI there's a GlobalAveragePooling1D after Embedding, How does the error go backwards?
A layer doesn't need to have weights in order to back-prop.
You can compute the gradients of a global avg pool w.r.t the inputs - it's simply dividing by the number of elements pooled.
It is a bit more tricky when it comes to max pooling: in that case, you propagate gradients through the pooled indices. That is, during back-prop, the gradients are "routed" to the input elements that contributed the maximal elements, no gradient is propagated to the other elements.
Related
Closed. This question needs to be more focused. It is not currently accepting answers.
Want to improve this question? Update the question so it focuses on one problem only by editing this post.
Closed 6 years ago.
Improve this question
Hy, I am working on project that based on sentiment analysis for product rating.
I have data set for good words and Negative words. When any user comment on website for product it will rate automatically out of 10
So i am confused with clustering technique and ago that solve my problem Plzzx Help
Thanks in Advance.
You are basically asking us what would be best for you to use as a classifier for your program while we have to idea how is your data stored.
However, it seems you only have two classes, positive and negative. And you want to classify new data based on word analysis of the data.
I have worked earlier in such problem, I used Rocchio's TF-IDF algorithm for such classification. You give it a set of training data (negative and positive words) and it classifies what later comes to the system.
It is based on vector classification and cosine similarity distance measure.
For more information you can read this paper.
You can find an example of how the method works (on very small data) here.
Note: the provided example is a section of a project I worked on.
Closed. This question does not meet Stack Overflow guidelines. It is not currently accepting answers.
This question does not appear to be about programming within the scope defined in the help center.
Closed 7 years ago.
Improve this question
While solving a puzzle, I ended up having a complexity of T(n)=T(n-3)+T(n-5).
I was trying subtraction method. But I am unable to solve this. Please explain what should be the procedure.
This is a linear homogeneous difference equation with constant coeffs.. It is usually solved by transforming it to the complex plane and solving a polynomial.
Without a CS background (as you state), I'm afraid the details wouldn't fit in here. Start with the Wikipedia entry, if you're interested.
If you want to skip to the final solution, here is the Wolfram Alpha for it.
Closed. This question does not meet Stack Overflow guidelines. It is not currently accepting answers.
Questions asking for code must demonstrate a minimal understanding of the problem being solved. Include attempted solutions, why they didn't work, and the expected results. See also: Stack Overflow question checklist
Closed 9 years ago.
Improve this question
I am working on large scale data like 300000 x 300000 matrix currently may interest. It is really hard to process in matlab due to "Out of memory" error so I decide to use EIGEN. Is there any restriciton for eigen in the matrix size?
The dense matrices in EIGEN are stored in a contiguous block of memory, which cannot exceed 2 GB in 32-bit application, so if you're running a 32-bit application, the allocation will start to crash for matrices half this size, that is to say around 10,000x10,000. See for example this duplicated question.
The same issue will happen with other libraries, since you're bounded by your RAM, not by the library.
Yet, if your big matrix has a vast majority of 0 in it, you can resort to SparseMatrix.
If your matrix isn't sparse, then you may store it on disk, but you'll get terrible performance when manipulating it.
Closed. This question does not meet Stack Overflow guidelines. It is not currently accepting answers.
This question does not appear to be about programming within the scope defined in the help center.
Closed 9 years ago.
Improve this question
Does the zero-value or sensitivity of an accelerometer changes over time? I.e is there something similar to the bias drift effect for gyroscopes, for accelerometers?
Yes. They also depend on temperature. And the temperature coefficient of offset (TCO) and the temperature coefficient of sensitivity (TCS) are different for the three axes. That's why you have the x,y,z calibration commands.
Closed. This question is off-topic. It is not currently accepting answers.
Want to improve this question? Update the question so it's on-topic for Stack Overflow.
Closed 10 years ago.
Improve this question
I'm trying to minimize this function (by A):
argmin_A (||L(A)||^2 + a||A-B||^2*)
where:
A is a MxN image L is the Laplacian Operator
||.|| is the usual norm (Frobrenius)
a is a weight parameter
B is a matrix of size (M+2*k)xN
where k is an integer parameter.
(*) indicates that we just consider the pixels in the boundary (we want to preserve in A the pixels in the boundary of B).
Maybe the problem has a trivial solution, but I'm absolutely blocked.
If you need more details, it's (4) equation in this paper.
I will be very grateful for any help provided.
Without looking carefully at that paper, gridfit does essentially that, although it does not employ the boundary conditions you ask for. You could certainly write what you want however, or you could simply apply a weight to the boundary points. This is not a difficult problem, but in order to solve it efficiently, you would want to employ the sparse matrix capabilities of MATLAB.