Skip to content
ml-recall

#reward-to-go

2 questions from Policy Gradient & Actor-Critic

Timed test

2 questions · 2 min · score and explanations at the end

Take a test
Q 1 / 2Score 0 / 0

Question 1: Short answer

easyAI-checkedPolicy Gradient & Actor-Critic

An episode has rewards r0,r1,r2,r3=1,0,2,3r_0, r_1, r_2, r_3 = 1, 0, 2, 3 (undiscounted). The simplest policy gradient weights every ∇log⁡π(at∣st)\nabla\log\pi(a_t\mid s_t) by the total return R(τ)=6R(\tau) = 6. With the reward-to-go form instead, what weight does the action taken at t=1t = 1 get?

a whole number