(a) (b) (c) The CITS3001 project this semester featured the game Hanabi, and it is assumed that you are familiar with the game. Suppose that we have observed an agent playing several games, and built a table showing what actions they played, depending on what state the game was in (how many cards had been discarded, whether thay had a playable card, whether someone else had a playable card, how many hints were remaining and how many fuse tokens were left). A small section of the table is below: Discards Can play Other can play card card Hints remaining 1 3 6 Fuse remaining 2 1 3 3 1 2 3 1 4 5 3 20 Yes Yes 15 No Yes 23 Yes Yes 30 No No 15 Yes Yes 12 Yes No 23 No No 1 27 No No 0 3 No Yes 8 1 Hint Describe the process of inducing a decision tree from this data. (You do not have to build the full tree, but you should describe the required steps). 4 marks Action Describe the process of temporal-difference learning. Play Hint Play Discard Hint Play Play Discard 3 marks Describe the process of Q-learning and give its advantages and disadvantages relative to temporal difference learning.