After digging a bit into machine learning i feel like there is a tendency to be fancy with terminology. For example, do we really need to use the obscure term ’stochastic’ when ’random’ exists?
For example "a random process" could mean "any process in existence" whereas "a stochastic process" clearly states stochastic as being a property of the process.
I however agree that things can sometimes get a bit wrapped up in sounding cool just to sound cool.
1. Stochastic.
2. Unbiased.
3. Typical and average.
4. Unprompted or unjustifiable.
5. Habitually using non sequiturs.
So, I might use a random sort algorithm like, I don't know, heapsort. Or I might be reading through some code and find a random sort algorithm --- why is there a mergesort in the middle of this calendar routine? Or, if you want a random sort algorithm, you should be sure to use the Fisher–Yates shuffle. Or, to avoid the quadratic worst case of treesort and quicksort with high probability, you can convert them into random sort algorithms by picking partition elements randomly.
Of these four random sort algorithms, only the fourth is stochastic. (I haven't figured out how a sort algorithm could habitually use non sequiturs.)
Random is a property of the world itself; stochastic means "we allow ourselves to think of this process as random because we can't/don't want to deal with it exactly".
Following this is the difference between Random Differential Equations and Stochastic Differential Equations.