See more
…he social realm. But the question of knowledge brings us to a pertinent question: Is this success — the ability to go from a predetermined point A to an imagined point B — a product of luck or agency?
…about minimising the error cost function (the loss function) over the whole dataset. This is called batch learning, and might be very slow for big data. What we can do instead, is to to update the weights every bat…
…ese functions. This can be done by de-stacking through the function calls. This technique is called auto-differentiation, and requires only that each function is provided with the implementation of its derivative. In a f…