Q3. Temporal Difference Learning (10 points) Consider the gridworld shown below. The left panel shows the name of each state A through E. The middle panel shows the current estimate of the value function V" for each state. A transi- tion is observed, that takes the agent from state B through taking action east into state C, and the agent receives a reward of -2. Assuming 7 = 1,2 = 0.5, what are the value estimates of U*(A), "(B), *(C), *(D), and (E) after the TD learning update? (note: the value will change for one of the states only) States Observed Transition: B, east, C, -2 A 1 BCD 1.2 | 8 10 E 10 Assume: y = 1, a = 1/2 V*(s) + (1 - a)V"(s) +a (R(s, (8), s') +yV*(s')]