#concept-pamphlet
Key question of CS238: how do we automate the process of good decision making?
Based on this graph, what are the 4 sources of uncertainty in decision making?
?
Our focus is on agents that interact intelligently to achieve their objectives over time. Given the past sequence of observations, o1, . . . , o_t and knowledge of the environment, the agent must choose an action at that best achieves its objectives in the presence of various sources of uncertainty, including the following:
- outcome uncertainty, where the effects of our actions are uncertain,
- model uncertainty, where our model of the problem is uncertain,
- state uncertainty, where the true state of the environment is uncertain, and
- interaction uncertainty, where the behavior of the other agents interacting in the environment is uncertain.
intention of reading this book: for anything that can be described as input -> output, how can I and others make informed decisions in uncertain (chaotic?) environments that maximize towards an intention?
methods for designing decision making agents (5)
Fields (8): economics, psychology, neuroscience, computer science, engineering, mathematics, operations research, societal impact Another one unlisted…self-fulfillment
q: how are linear programming, dynamic programming, and queuing theory related to operations research? (p12)
Each of the 5 methods has varying level of how much the designer and automata interact with each other:
___. decision tree paths- supervised learning.
- optimization.
- planning.
_____. ? explicit programming, reinforcement learning
What are the downsides of automating away decisions?
- it’s also automating away
____, and taste for____. - it overweights to
_____, rather than what might happen. ____are not accounted for ? nuance, what’s already happened, Black swans
