Ciro Santilli @cirosantilli 37

CNN convolution kernels are not hardcoded. They are learnt and optimized via backpropagation. You just specify their size! Example in PyTorch you'd do just:as used for example at: activatedgeek/LeNet-5.

This can also be inferred from: stackoverflow.com/questions/55594969/how-to-visualise-filters-in-a-cnn-with-pytorch where we see that the kernels are not perfectly regular as you'd expected from something hand coded.

The CNOT gate is a controlled quantum gate that operates on two qubits, flipping the second (operand) qubit if the first (control) qubit is set.

This gate is the first example of a controlled quantum gate that you should study.

To understand why the gate is called a CNOT gate, you should think as follows.

First let's produce a generic quantum state vector where the control qubit is certain to be 0.

On the standard basis:we see that this means that only $|00⟩$ and $|01⟩$ should be possible. Therefore, the state must be of the form:where $x$ and $y$ are two complex numbers such that $|x|+|y|=1.0$

If we operate the CNOT gate on that state, we obtain:and so the input is unchanged as desired, because the control qubit is 0.

If however we take only states where the control qubit is for sure 1:

Therefore, in that case, what happened is that the probabilities of $|10⟩$ and $|11⟩$ were swapped from $x$ and $y$ to $y$ and $x$ respectively, which is exactly what the quantum NOT gate does.

So from this we understand more concretely what "the gate only operates if the first qubit is set to one" means.

Now go and study the Bell state and understand intuitively how this gate is used to produce it.

Finding a complete basis such that each vector solves a given differential equation is the basic method of solving partial differential equation through separation of variables.

The first example of this you must see is solving partial differential equations with the Fourier series.

Notable examples:

This section is about the definition of the dot product over ${\mathbb{C}}^n$, which extends the definition of the dot product over ${\mathbb{R}}^n$.

Some motivation is discussed at: math.stackexchange.com/questions/2459814/what-is-the-dot-product-of-complex-vectors/4300169#4300169

The complex dot product is defined as:

E.g. in ${\mathbb{C}}^1$:

We can see therefore that this is a form, and a positive definite because:

Just like the usual dot product, this will be a positive definite symmetric bilinear form by definition.

A:

math.stackexchange.com/questions/2382011/computational-complexity-of-modular-exponentiation-from-rosens-discrete-mathem mentions:can be calculated in:Remember that $log(m)$ and $log(n)$ are the lengths in bits of $m$ and $n$, so in terms of the length in bits $M$ and $N$ we'd get:

The intersection of two beautiful arts: coding and physics!

Computational physics is a good way to get valuable intuition about the key equations of physics, and train your numerical analysis skills:

Other child sections:

Programming is hard. To Ciro Santilli, it's almost masochistic.

What makes Ciro especially mad when programming is not the hard things.

It is the things that should be easy, but aren't, and which take up a lot of your programming time.

Especially when you are already a few levels of "simple problems" down from your original goal, and another one of them shows up.

This is basically the cause of Hofstadter's law.

But of course, it is because it is hard that it feels amazing when you achieve your goal.

Putting a complex and useful program together is like composing a symphony, or reaching the summit of a hard rock climbing proble.

Programming can be an art form. There can be great beauty in code and what it does. It is a shame that this is hard to see from within the walls of most companies, where you are stuck doing a small specific task as fast as possible.