I am trying to implement the equation for asymmetrically quantized matrix multiplcation from the paper https://arxiv.org/abs/2106.08295 (equation 13)
In this system, real-number matrices are represented as an affine map with a scale factor s and a zero point z. The terms in blue are possible to pre-compute.
Computing each of final terms is simple enough, but I do not understand how they are supposed to sum together. The first term s_w * s_x * W_int * X_int is a matrix multiplication of integer values, so if the input matrices W,X had dimensions m by n and n by p respectively, the result would be a m by p matrix. The second term s_w * z_w * s_x * X_int would have dimensions n by p, so I do not understand how it could be subtracted from the first term.
I understand this algorithm is implemented in many tinyML applications, eg TensorFlow lite, has anyone any experience here? Or know how it was done?
I have a set of signals S1, S2, ....,SN, for which I am numerically computing the Fourier transforms F1, F2, ,,,,FN. Where Si's and Fi's are C++ vectors (my computations are in C++).
My Computation objective is as follows:
compute product: F = F1*F2*...*FN
inverse Fourier transform F to get a S.
What I numerically observe is when I am trying to compute the product either I am running into situation where numbers become too small. Or numbers become too big.
I thought of scaling F1 by say a1 so that the overflow or underflow is avoided.
With the scaling my step 1 above will become
F' = (F1/a1)*(F2/a2)*...*(FN/aN) = F'1*F'2*...*F'N
And when I inverse transform my final S' will differ from S by a scale factor. The structural form of the S will not change. By this I mean only the normalization of S is different.
My question is :
Is my rationale correct.
If my rationale is correct then given a C++ vector "Fi" how can I choose a good scale "ai" to scale up or down the Fi's.
Many thanks in advance.
You can change the range of the Fi vector in a new domain, let's say [a, b]. So from your vector Fi, the minimum number will become a and the maximum number will become b. You can scale the vector using this equation:
Fnew[i] = (b-a)/(max-min) * (F[i] - min) + a,
where: min = the minimum value from F
max= the maximum value from F
The only problem now is choosing a and b. I would choose [1,2], because the values are small (so you won't get overflow easily), but not smaller than 0.
I started to learn Machine Learning. Now i tried to play around with tensorflow.
Often i see examples like this:
pred = tf.add(tf.mul(X, W), b)
I also saw such a line in a plain numpy implementation. Why is always x*W+b used instead of W*x+b? Is there an advantage if matrices are multiplied in this way? I see that it is possible (if X, W and b are transposed), but i do not see an advantage. In school in the math class we always only used Wx+b.
Thank you very much
This is the reason:
By default w is a vector of weights and in maths a vector is considered a column, not a row.
X is a collection of data. And it is a matrix nxd (where n is the number of data and d the number of features) (upper case X is a matrix n x d and lower case only 1 data 1 x d matrix).
To correctly multiply both and use the correct weight in the correct feature you must use X*w+b:
With X*w you mutliply every feature by its corresponding weight and by adding b you add the bias term on every prediction.
If you multiply w * X you multipy a (1 x d)*(n x d) and it has no sense.
I'm also confused with this. I guess this may be a dimension matter. For a n*m-dimension matrix W and a n-dimension vector x, using xW+b can be easily viewed as that maping a n-dimension feature to a m-dimension feature, i.e., you can easily think W as a n-dimension -> m-dimension operation, where as Wx+b (x must be m-dimension vector now) becomes a m-dimension -> n-dimension operation, which looks less comfortable in my opinion. :D
I am newbie in convolutional neural networks and just have idea about feature maps and how convolution is done on images to extract features. I would be glad to know some details on applying batch normalisation in CNN.
I read this paper https://arxiv.org/pdf/1502.03167v3.pdf and could understand the BN algorithm applied on a data but in the end they mentioned that a slight modification is required when applied to CNN:
For convolutional layers, we additionally want the normalization to obey the convolutional property – so that different elements of the same feature map, at different locations, are normalized in the same way. To achieve this, we jointly normalize all the activations in a mini- batch, over all locations. In Alg. 1, we let B be the set of all values in a feature map across both the elements of a mini-batch and spatial locations – so for a mini-batch of size m and feature maps of size p × q, we use the effec- tive mini-batch of size m′ = |B| = m · pq. We learn a pair of parameters γ(k) and β(k) per feature map, rather than per activation. Alg. 2 is modified similarly, so that during inference the BN transform applies the same linear transformation to each activation in a given feature map.
I am total confused when they say
"so that different elements of the same feature map, at different locations, are normalized in the same way"
I know what feature maps mean and different elements are the weights in every feature map. But I could not understand what location or spatial location means.
I could not understand the below sentence at all
"In Alg. 1, we let B be the set of all values in a feature map across both the elements of a mini-batch and spatial locations"
I would be glad if someone cold elaborate and explain me in much simpler terms
Let's start with the terms. Remember that the output of the convolutional layer is a 4-rank tensor [B, H, W, C], where B is the batch size, (H, W) is the feature map size, C is the number of channels. An index (x, y) where 0 <= x < H and 0 <= y < W is a spatial location.
Usual batchnorm
Now, here's how the batchnorm is applied in a usual way (in pseudo-code):
# t is the incoming tensor of shape [B, H, W, C]
# mean and stddev are computed along 0 axis and have shape [H, W, C]
mean = mean(t, axis=0)
stddev = stddev(t, axis=0)
for i in 0..B-1:
out[i,:,:,:] = norm(t[i,:,:,:], mean, stddev)
Basically, it computes H*W*C means and H*W*C standard deviations across B elements. You may notice that different elements at different spatial locations have their own mean and variance and gather only B values.
Batchnorm in conv layer
This way is totally possible. But the convolutional layer has a special property: filter weights are shared across the input image (you can read it in detail in this post). That's why it's reasonable to normalize the output in the same way, so that each output value takes the mean and variance of B*H*W values, at different locations.
Here's how the code looks like in this case (again pseudo-code):
# t is still the incoming tensor of shape [B, H, W, C]
# but mean and stddev are computed along (0, 1, 2) axes and have just [C] shape
mean = mean(t, axis=(0, 1, 2))
stddev = stddev(t, axis=(0, 1, 2))
for i in 0..B-1, x in 0..H-1, y in 0..W-1:
out[i,x,y,:] = norm(t[i,x,y,:], mean, stddev)
In total, there are only C means and standard deviations and each one of them is computed over B*H*W values. That's what they mean when they say "effective mini-batch": the difference between the two is only in axis selection (or equivalently "mini-batch selection").
Some clarification on Maxim's answer.
I was puzzled by seeing in Keras that the axis you specify is the channels axis, as it doesn't make sense to normalize over the channels - as every channel in a conv-net is considered a different "feature". I.e. normalizing over all channels is equivalent to normalizing number of bedrooms with size in square feet (multivariate regression example from Andrew's ML course). This is usually not what you want - what you do is normalize every feature by itself. I.e. you normalize the number of bedrooms across all examples to be with mu=0 and std=1, and you normalize the the square feet across all examples to be with mu=0 and std=1.
This is why you want C means and stds, because you want a mean and std per channel/feature.
After checking and testing it myself I realized the issue: there's a bit of a confusion/misconception here. The axis you specify in Keras is actually the axis which is not in the calculations. i.e. you get average over every axis except the one specified by this argument. This is confusing, as it is exactly the opposite behavior of how NumPy works, where the specified axis is the one you do the operation on (e.g. np.mean, np.std, etc.).
I actually built a toy model with only BN, and then calculated the BN manually - took the mean, std across all the 3 first dimensions [m, n_W, n_H] and got n_C results, calculated (X-mu)/std (using broadcasting) and got identical results to the Keras results.
Hope this helps anyone who was confused as I was.
I'm only 70% sure of what I say, so if it does not make sense, please edit or mention it before downvoting.
About location or spatial location: they mean the position of pixels in an image or feature map. A feature map is comparable to a sparse modified version of image where concepts are represented.
About so that different elements of the same feature map, at different locations, are normalized in the same way:
some normalisation algorithms are local, so they are dependent of their close surrounding (location) and not the things far apart in the image. They probably mean that every pixel, regardless of their location, is treated just like the element of a set, independently of it's direct special surrounding.
About In Alg. 1, we let B be the set of all values in a feature map across both the elements of a mini-batch and spatial locations: They get a flat list of every values of every training example in the minibatch, and this list combines things whatever their location is on the feature map.
Firstly we need to make it clear that the depth of a kernel is determined by previous feature map's channel num, and the number of kernel in this layer determins the channel num of next feature map (the next layer).
then we should make it clear that each kernel(three dimentional usually) will generate just one channel of feature map in the next layer.
thirdly we should try to accept the idea of each points in the generated feature map (regardless of their position) are generated by the same kernel, by sliding on previous layer. So they could be seen as a distribution generated by this kernel, and they could be seen as samples of a stochastic variable. Then they should be averaged to obtain the mean and then the variance. (it not rigid, only helps to understand)
This is what they say "so that different elements of the same feature map, at different locations, are normalized in the same way"
I am currently reading the Machine Learning book by Tom Mitchell. When talking about neural networks, Mitchell states:
"Although the perceptron rule finds a successful weight vector when
the training examples are linearly separable, it can fail to converge
if the examples are not linearly separable. "
I am having problems understanding what he means with "linearly separable"? Wikipedia tells me that "two sets of points in a two-dimensional space are linearly separable if they can be completely separated by a single line."
But how does this apply to the training set for neural networks? How can inputs (or action units) be linearly separable or not?
I'm not the best at geometry and maths - could anybody explain it to me as though I were 5? ;) Thanks!
Suppose you want to write an algorithm that decides, based on two parameters, size and price, if an house will sell in the same year it was put on sale or not. So you have 2 inputs, size and price, and one output, will sell or will not sell. Now, when you receive your training sets, it could happen that the output is not accumulated to make our prediction easy (Can you tell me, based on the first graph if X will be an N or S? How about the second graph):
^
| N S N
s| S X N
i| N N S
z| S N S N
e| N S S N
+----------->
price
^
| S S N
s| X S N
i| S N N
z| S N N N
e| N N N
+----------->
price
Where:
S-sold,
N-not sold
As you can see in the first graph, you can't really separate the two possible outputs (sold/not sold) by a straight line, no matter how you try there will always be both S and N on the both sides of the line, which means that your algorithm will have a lot of possible lines but no ultimate, correct line to split the 2 outputs (and of course to predict new ones, which is the goal from the very beginning). That's why linearly separable (the second graph) data sets are much easier to predict.
This means that there is a hyperplane (which splits your input space into two half-spaces) such that all points of the first class are in one half-space and those of the second class are in the other half-space.
In two dimensions, that means that there is a line which separates points of one class from points of the other class.
EDIT: for example, in this image, if blue circles represent points from one class and red circles represent points from the other class, then these points are linearly separable.
In three dimensions, it means that there is a plane which separates points of one class from points of the other class.
In higher dimensions, it's similar: there must exist a hyperplane which separates the two sets of points.
You mention that you're not good at math, so I'm not writing the formal definition, but let me know (in the comments) if that would help.
Look at the following two data sets:
^ ^
| X O | AA /
| | A /
| | / B
| O X | A / BB
| | / B
+-----------> +----------->
The left data set is not linearly separable (without using a kernel). The right one is separable into two parts for A' andB` by the indicated line.
I.e. You cannot draw a straight line into the left image, so that all the X are on one side, and all the O are on the other. That is why it is called "not linearly separable" == there exist no linear manifold separating the two classes.
Now the famous kernel trick (which will certainly be discussed in the book next) actually allows many linear methods to be used for non-linear problems by virtually adding additional dimensions to make a non-linear problem linearly separable.