upshard

Save your progress

Everything on Upshard is free and readable without an account. Signing in saves what you have finished, your bookmarks and your notes — and nothing else.

or

We store your email, your progress, your bookmarks and your notes. Nothing else, ever. What we do with it.

Your account

Delete your account?

This removes your progress, bookmarks and notes immediately and permanently. It cannot be undone. Everything on Upshard stays free to read without an account.

Advantage function

Reinforcement learning · Formula · Not yet linked from a lesson

Aπ(s,a)=Qπ(s,a)−Vπ(s)A^{\pi}(s, a) = Q^{\pi}(s, a) - V^{\pi}(s)

How much better one action is than the state's average. Subtracting the state value removes everything the choice had no say in, which cuts gradient variance without biasing the estimate.

Symbols

AπA^{\pi}
The advantage of action a in state s.
QπQ^{\pi}
The value of that specific action.
VπV^{\pi}
The average value of the state.
Esc
↑↓ navigate↵ open? shortcuts

Keyboard shortcuts

Press Esc to close

Anywhere

⌘K
Search and commands
/
Search
?
This sheet
Esc
Close

Reading a lesson

J
Next lesson
K
Previous lesson
B
Bookmark