Sign inSign in
ConceptAI & ML

Value function

The expected sum of future rewards starting from state st and following policy pi thereafter.
as Sergey Levine defines itUC BerkeleyThe wording closest to all 11 definitions
Watch it explained

Know first

Named as a prerequisite by the lectures that define this. The number is how many say so. A grey one has no page yet.

Opens up

Concepts that name this one as a prerequisite.

Lecturers

8

Each in their own terms, from the easiest level.

  • David Silver

    Google DeepMind · Introductory

    A function that tells us the expected total reward given policies by all players.
    • 55:44Introduction to Reinforcement Learning
    • 23:22Markov Processes, Reward Processes & Bellman Equations
    • 21:52Classic Games
  • Ian Ball

    MIT · Undergraduate

    The average discounted payoff from today onward if we're in a given state, assuming players follow the strategy profile.
  • Robert Shiller

    Yale University · Undergraduate

    A function representing how people subjectively value gains and losses, characterized by a concave curve for gains, a convex curve for losses, and a kink at the reference point.
    • 21:46Behavioral Finance and the Role of Psychology
  • Fei-Fei Li, Justin Johnson & Serena Yeung

    Stanford University · Undergraduate

    The expected cumulative reward from following a policy starting from a given state.
    • 11:53Deep Reinforcement Learning
  • Sergey Levine

    UC Berkeley · Graduate

    The expected value over all actions in a given state under your current policy of the Q-value.
    • 8:35Natural Policy Gradients, Covariant Updates & The Fisher Information Matrix
  • Stephen Boyd

    Stanford University · Graduate

    The optimal value of the control problem expressed as a function of the initial state z.
    • 10:33Convex Optimization II: Model Predictive Control
  • Parham Panahi

    Alberta Machine Intelligence Institute (Amii) · Graduate

    The expected sum of future rewards when taking a certain action in a certain state.
    • 7:27Endpoint Replay: Compressing the Recency Buffer in Deep Reinforcement Learning
  • Ashia Wilson

    UC Berkeley · Research

    A function that assigns a value to each score given the true label and the classifier's decision.

Lectures that assume it

32

The first 12 are shown.

Connections

53

The lectures that define it or assume it, and the concepts next to it.

Concepts10

Show 4 more

Lectures43

Show 37 more

8 lecturers. Definitions are written by AI from each lecture’s transcript. A time with a play mark starts the lecture at the moment the lecturer says it.