After this sequence of transitions, if we use a learning rate of = 0.5, what would temporal difference learning learn for the values of the states A, B, C?

2 Temporal Difference Learning Consider the following mini-grid, with the state names shown. L and R are terminal states. Re- wards are zero except moving to L (+4) or to R (+16), = 1 and actions (left or right) are deterministic. We do not know the MDP, so we use reinforcement learning to compute various values. We observe the following transition sequence. B Right C 0 C Right R 16 B Left A 0 A Left L 4 B Right C 0 C Right R 16 B Left A 0 A Left L 4 After this sequence of transitions, if we use a learning rate of difference learning learn for the values of the states A, B, C? 0.5, what would temporal s a s'r V (A) V (B) V(C) 0 0 0 B Right C 0 C Right R 16 B Left A 0 A Left L 4 B Right C0 C Right R 16 B Left A 0 A Left L 4 2 Temporal Difference Learning Consider the following mini-grid, with the state names shown. L and R are terminal states. Re- wards are zero except moving to L (+4) or to R (+16), = 1 and actions (left or right) are deterministic. We do not know the MDP, so we use reinforcement learning to compute various values. We observe the following transition sequence. B Right C 0 C Right R 16 B Left A 0 A Left L 4 B Right C 0 C Right R 16 B Left A 0 A Left L 4 After this sequence of transitions, if we use a learning rate of difference learning learn for the values of the states A, B, C? 0.5, what would temporal s a s'r V (A) V (B) V(C) 0 0 0 B Right C 0 C Right R 16 B Left A 0 A Left L 4 B Right C0 C Right R 16 B Left A 0 A Left L 4