I use torch.nn.Embedding to embed my model’s categorical input features, however, I face problems when I set the max_norm parameter to not None.
There is a note on the pytorch docs page that explains how to use max_norm parameter through the following example:
n, d, m = 3, 5, 7
embedding = nn.Embedding(n, d, max_norm=True)
W = torch.randn((m, d), requires_grad=True)
idx = torch.tensor(\[1, 2\])
a = embedding.weight.clone() # W.t() # weight must be cloned for this to be differentiable
b = embedding(idx) # W.t() # modifies weight in-place
out = (a.unsqueeze(0) + b.unsqueeze(1))
loss = out.sigmoid().prod()
loss.backward()
I can’t easily understand this example from the docs. What is the purpose of having both ‘a’ and ‘b’ and why ‘out’ is defined as, out = (a.unsqueeze(0) + b.unsqueeze(1))?
Do we need to first clone the entire embedding tensor as in ‘a’, and then finding the embeddings for our desired indices as in ‘b’? Then how do ‘a’ and ‘b’ need to be added?
In my code, I don’t have W explicitly, I am assuming that W is representative of the weights applied by the torch.nn.Linear layers. So, I just need to prepare the input (which includes the embeddings for categorical features) that goes into my network.
I greatly appreciate any instructions on this, as understanding this example would help me adapt my code accordingly.
Because W in the line computing a requires gradients, we must save embedding.weight to compute those gradients in the backward pass. However, in the line computing b, executing embedding(idx) will scale embedding.weight by max_norm - in place. So, without cloning it in line a, embedding.weight will be modified when line b is executed - changing what was saved for the backward pass to update W. Hence the requirement to clone embedding.weight - to save it before it gets scaled in line b.
If you don't use embedding.weight outside of the normal forward pass, you don't need to worry about all this.
If you get an error, post it (and your code).
My bold claim is that the Octave implementation of the cubic spline, as implemented in interp1(..., "spline") differs from the "natural cubic spline" algorithm outlined in, e.g., Wolfram's Mathworld. I have written my own implementation of the latter and compared it to the output of the interp1(..., "spline") function, with the following results:
I discovered that when I try the same comparison with 4 points, the solutions also differ, and, moreover, the Octave solution is identical to fitting a single cubic polynomial to all four points (and not actually producing a piecewise spline for the three intervals).
I also tried to look under the hood at Octave's implementation of splines, and found it was too obtuse to read and understand in 5 minutes.
I know that there are a few options for boundary conditions one can choose ("natural" vs "clamped") when implementing a cubic spline. My implementation uses "natural" boundary conditions (in which the second derivative of the two exterior points is set to zero).
If Octave's cubic spline is indeed different to the standard cubic spline, then what actually is it?
EDIT:
The second order differences of the two solutions shown in the Comparison plot above are plotted here:
Firstly, there appear to be only two cubic polynomials in Octave's case: one that is fit over the first two intervals, and one that is fit over the last two intervals. Secondly, they are clearly not using "natural" splines, since the second derivatives at the extremes do not tend to zero.
Also, I think the fact that the second order difference for my implementation at the middle (i.e. 3rd) point is zero is just a coincidence, and not demanded by the algorithm. Repeating this test for a different set of points will confirm/refute this.
Different end conditions explains the difference between your implementation and Octave's. Octave uses the not-a-knot condition (depending on input)
See help spline
To explain your observations: the third derivative is continuous at the 2nd and (n-1)th break due to the not-a-knot condition, so that's why Octave's second derivative looks like it has less 'breaks', because it is a continuous straight line over the first two and last two segments. If you look at the third derivative, you can see the effect more clearly - the 3rd derivative is discontinuous only at the 3rd break (the middle)
x = 1:5;
y = rand(1,5);
xx = linspace(1,5);
pp = interp1(x, y, 'spline', 'pp');
yy = ppval(pp, xx);
dyy = ppval(ppder(pp, 3), xx);
plot(xx, yy, xx, dyy);
Also the pp data structure looks like this
pp =
scalar structure containing the fields:
form = pp
breaks =
1 2 3 4 5
coefs =
0.427823 -1.767499 1.994444 0.240388
0.427823 -0.484030 -0.257085 0.895156
-0.442232 0.799439 0.058324 0.581864
-0.442232 -0.527258 0.330506 0.997395
pieces = 4
order = 4
dim = 1
orient = first
So, I have a vector that corresponds to a given feature (same dimensionality). Is there a package in Julia that would provide a mathematical function that fits these data points, in relation to the original feature? In other words, I have x and y (both vectors) and need to find a decent mapping between the two, even if it's a highly complex one. The output of this process should be a symbolic formula that connects x and y, e.g. (:x)^3 + log(:x) - 4.2454. It's fine if it's just a polynomial approximation.
I imagine this is a walk in the park if you employ Genetic Programming, but I'd rather opt for a simpler (and faster) approach, if it's available. Thanks
Turns out the Polynomials.jl package includes the function polyfit which does Lagrange interpolation. A usage example would go:
using Polynomials # install with Pkg.add("Polynomials")
x = [1,2,3] # demo x
y = [10,12,4] # demo y
polyfit(x,y)
The last line returns:
Poly(-2.0 + 17.0x - 5.0x^2)`
which evaluates to the correct values.
The polyfit function accepts a maximal degree for the output polynomial, but defaults to using the length of the input vectors x and y minus 1. This is the same degree as the polynomial from the Lagrange formula, and since polynomials of such degree agree on the inputs only if they are identical (this is a basic theorem) - it can be certain this is the same Lagrange polynomial and in fact the only one of such a degree to have this property.
Thanks to the developers of Polynomial.jl for leaving me just to google my way to an Answer.
Take a look to MARS regression. Multi adaptive regression splines.
I need an algorithm to perform a 2D bisection method for solving a 2x2 non-linear problem. Example: two equations f(x,y)=0 and g(x,y)=0 which I want to solve simultaneously. I am very familiar with the 1D bisection ( as well as other numerical methods ). Assume I already know the solution lies between the bounds x1 < x < x2 and y1 < y < y2.
In a grid the starting bounds are:
^
| C D
y2 -+ o-------o
| | |
| | |
| | |
y1 -+ o-------o
| A B
o--+------+---->
x1 x2
and I know the values f(A), f(B), f(C) and f(D) as well as g(A), g(B), g(C) and g(D). To start the bisection I guess we need to divide the points out along the edges as well as the middle.
^
| C F D
y2 -+ o---o---o
| | |
|G o o M o H
| | |
y1 -+ o---o---o
| A E B
o--+------+---->
x1 x2
Now considering the possibilities of combinations such as checking if f(G)*f(M)<0 AND g(G)*g(M)<0 seems overwhelming. Maybe I am making this a little too complicated, but I think there should be a multidimensional version of the Bisection, just as Newton-Raphson can be easily be multidimed using gradient operators.
Any clues, comments, or links are welcomed.
Sorry, while bisection works in 1-d, it fails in higher dimensions. You simply cannot break a 2-d region into subregions using only information about the function at the corners of the region and a point in the interior. In the words of Mick Jagger, "You can't always get what you want".
I just stumbled upon the answer to this from geometrictools.com and C++ code.
edit: the code is now on github.
I would split the area along a single dimension only, alternating dimensions. The condition you have for existence of zero of a single function would be "you have two points of different sign on the boundary of the region", so I'd just check that fro the two functions. However, I don't think it would work well, since zeros of both functions in a particular region don't guarantee a common zero (this might even exist in a different region that doesn't meet the criterion).
For example, look at this image:
There is no way you can distinguish the squares ABED and EFIH given only f() and g()'s behaviour on their boundary. However, ABED doesn't contain a common zero and EFIH does.
This would be similar to region queries using eg. kD-trees, if you could positively identify that a region doesn't contain zero of eg. f. Still, this can be slow under some circumstances.
If you can assume (per your comment to woodchips) that f(x,y)=0 defines a continuous monotone function y=f2(x), i.e. for each x1<=x<=x2 there is a unique solution for y (you just can't express it analytically due to the messy form of f), and similarly y=g2(x) is a continuous monotone function, then there is a way to find the joint solution.
If you could calculate f2 and g2, then you could use a 1-d bisection method on [x1,x2] to solve f2(x)-g2(x)=0. And you can do that by using 1-d bisection on [y1,y2] again for solving f(x,y)=0 for y for any given fixed x that you need to consider (x1, x2, (x1+x2)/2, etc) - that's where the continuous monotonicity is helpful -and similarly for g. You have to make sure to update x1-x2 and y1-y2 after each step.
This approach might not be efficient, but should work. Of course, lots of two-variable functions don't intersect the z-plane as continuous monotone functions.
I'm not much experient on optimization, but I built a solution to this problem with a bisection algorithm like the question describes. I think is necessary to fix a bug in my solution because it compute tow times a root in some cases, but i think it's simple and will try it later.
EDIT: I seem the comment of jpalecek, and now I anderstand that some premises I assumed are wrong, but the methods still works on most cases. More especificaly, the zero is garanteed only if the two functions variate the signals at oposite direction, but is need to handle the cases of zero at the vertices. I think is possible to build a justificated and satisfatory heuristic to that, but it is a little complicated and now I consider more promising get the function given by f_abs = abs(f, g) and build a heuristic to find the local minimuns, looking to the gradient direction on the points of the middle of edges.
Introduction
Consider the configuration in the question:
^
| C D
y2 -+ o-------o
| | |
| | |
| | |
y1 -+ o-------o
| A B
o--+------+---->
x1 x2
There are many ways to do that, but I chose to use only the corner points (A, B, C, D) and not middle or center points liky the question sugests. Assume I have tow function f(x,y) and g(x,y) as you describe. In truth it's generaly a function (x,y) -> (f(x,y), g(x,y)).
The steps are the following, and there is a resume (with a Python code) at the end.
Step by step explanation
Calculate the product each scalar function (f and g) by them self at adjacent points. Compute the minimum product for each one for each direction of variation (axis, x and y).
Fx = min(f(C)*f(B), f(D)*f(A))
Fy = min(f(A)*f(B), f(D)*f(C))
Gx = min(g(C)*g(B), g(D)*g(A))
Gy = min(g(A)*g(B), g(D)*g(C))
It looks to the product through tow oposite sides of the rectangle and computes the minimum of them, whats represents the existence of a changing of signal if its negative. It's a bit of redundance but work's well. Alternativaly you can try other configuration like use the points (E, F, G and H show in the question), but I think make sense to use the corner points because it consider better the whole area of the rectangle, but it is only a impression.
Compute the minimum of the tow axis for each function.
F = min(Fx, Fy)
G = min(Gx, Gy)
It of this values represents the existence of a zero for each function, f and g, within the rectangle.
Compute the maximum of them:
max(F, G)
If max(F, G) < 0, then there is a root inside the rectangle. Additionaly, if f(C) = 0 and g(C) = 0, there is a root too and we do the same, but if the root is in other corner we ignore him, because other rectangle will compute it (I want to avoid double computation of roots). The statement bellow resumes:
guaranteed_contain_zeros = max(F, G) < 0 or (f(C) == 0 and g(C) == 0)
In this case we have to proceed breaking the region recursively ultil the rectangles are as small as we want.
Else, may still exist a root inside the rectangle. Because of that, we have to use some criterion to break this regions ultil the we have a minimum granularity. The criterion I used is to assert the largest dimension of the current rectangle is smaller than the smallest dimension of the original rectangle (delta in the code sample bellow).
Resume
This Python code resume:
def balance_points(x_min, x_max, y_min, y_max, delta, eps=2e-32):
width = x_max - x_min
height = y_max - y_min
x_middle = (x_min + x_max)/2
y_middle = (y_min + y_max)/2
Fx = min(f(C)*f(B), f(D)*f(A))
Fy = min(f(A)*f(B), f(D)*f(C))
Gx = min(g(C)*g(B), g(D)*g(A))
Gy = min(g(A)*g(B), g(D)*g(C))
F = min(Fx, Fy)
G = min(Gx, Gy)
largest_dim = max(width, height)
guaranteed_contain_zeros = max(F, G) < 0 or (f(C) == 0 and g(C) == 0)
if guaranteed_contain_zeros and largest_dim <= eps:
return [(x_middle, y_middle)]
elif guaranteed_contain_zeros or largest_dim > delta:
if width >= height:
return balance_points(x_min, x_middle, y_min, y_max, delta) + balance_points(x_middle, x_max, y_min, y_max, delta)
else:
return balance_points(x_min, x_max, y_min, y_middle, delta) + balance_points(x_min, x_max, y_middle, y_max, delta)
else:
return []
Results
I have used a similar code similar in a personal project (GitHub here) and it draw the rectangles of the algorithm and the root (the system have a balance point at the origin):
Rectangles
It works well.
Improvements
In some cases the algorithm compute tow times the same zero. I thinh it can have tow reasons:
I the case the functions gives exatly zero at neighbour rectangles (because of an numerical truncation). In this case the remedy is to incrise eps (increase the rectangles). I chose eps=2e-32, because 32 bits is a half of the precision (on 64 bits archtecture), then is problable that the function don't gives a zero... but it was more like a guess, I don't now if is the better. But, if we decrease much the eps, it extrapolates the recursion limit of Python interpreter.
The case in witch the f(A), f(B), etc, are near to zero and the product is truncated to zero. I think it can be reduced if we use the product of the signals of f and g in place of the product of the functions.
I think is possible improve the criterion to discard a rectangle. It can be made considering how much the functions are variating in the region of the rectangle and how distante the function is of zero. Perhaps a simple relation between the average and variance of the function values on the corners. In another way (and more complicated) we can use a stack to store the values on each recursion instance and garantee that this values are convergent to stop recursion.
This is a similar problem to finding critical points in vector fields (see http://alglobus.net/NASAwork/topology/Papers/alsVugraphs93.ps).
If you have the values of f(x,y) and g(x,y) at the vertexes of your quadrilateral and you are in a discrete problem (such that you don't have an analytical expression for f(x,y) and g(x,y) nor the values at other locations inside the quadrilateral), then you can use bilinear interpolation to get two equations (for f and g). For the 2D case the analytical solution will be a quadratic equation which, according to the solution (1 root, 2 real roots, 2 imaginary roots) you may have 1 solution, 2 solutions, no solutions, solutions inside or outside your quadrilateral.
If instead you have analytic functions of f(x,y) and g(x,y) and want to use them, this is not useful. Instead you could divide your quadrilateral recursively, however as it was already pointed out by jpalecek (2nd post), you would need a way to stop your divisions by figuring out a test that would assure you would have no zeros inside a quadrilateral.
Let f_1(x,y), f_2(x,y) be two functions which are continuous and monotonic with respect to x and y. The problem is to solve the system f_1(x,y) = 0, f_2(x,y) = 0.
The alternating-direction algorithm is illustrated below. Here, the lines depict sets {f_1 = 0} and {f_2 = 0}. It is easy to see that the direction of movement of the algorithm (right-down or left-up) depends on the order of solving the equations f_i(x,y) = 0 (e.g., solve f_1(x,y) = 0 w.r.t. x then solve f_2(x,y) = 0 w.r.t. y OR first solve f_1(x,y) = 0 w.r.t. y and then solve f_2(x,y) = 0 w.r.t. x).
Given the initial guess, we don't know where the root is. So, in order to find all roots of the system, we have to move in both directions.