Politics as statistical learner
January 31, 2022 — February 17, 2022
In mind-as-ml I wondered whether statistical learning models can give us insight into the individual human mind. Here I wonder if we can learn anything about humans en masse from statistical analogies. Probably, since this spans Herbert Simon’s interests, I should be ransacking his work more thoroughly.
1 Greedy algorithms as pragmatism
Contrast with utopian ideals as analytic solutions.
2 Game theory
3 Learning agents
4 Federated learning
5 Machines of loving grace
The cybernetics folks. Politics as operations research. Insert Adam Curtis documentary here, and a brief history of Operations Research.
TODO: Mention Pigs for the Ancestors, and the Water Temples (Lansing 2000; Rappaport 1967; Sage 2013).
6 Loss functions and public risk aversion
Which experiments do we permit the state to do?
7 Legibility and labelled data sets
Which census category do you fit in and which benefits do you reap thereby? Affirmative action, Trans-exclusionary Radical Feminists etc.
8 Evolutionary learning in institutions
9 Dimension reduction of politics
See political axes