I think that when most people picture a Bayesian agent, they imagine a system which:

• Enumerates every possible state/trajectory of “the world”, and assigns a probability to each.
• When new observations come in, loops over every state/trajectory, checks the probability of the observations conditional on each, and then updates via Bayes rule.
• To select actions, computes the utility which each action will yield under each state/trajectory, then averages over state/trajectory weighted by probability, and picks the action with the largest weighted-average utility.

Typically, we define Bayesian agents as agents which behaviorally match that picture.

But that's not really the picture David and I typically have in mind, when we picture Bayesian agents. Yes, behaviorally they act that way. But I think people get overly-anchored imagining the internals of the agent that way, and then mistakenly imagine that a Bayesian model of agency is incompatible with various features of [...]

---

Outline:

(01:20) Causal Models and Submodels

(03:15) Updates

(04:04) Latents

(06:19) Aside: Map-Territory Correspondence

(07:50) Utility Over Latents

(08:33) Lazy Utility Maximization

(09:52) Caching and Inconsistency

(11:06) Putting It All Together

The original text contained 1 footnote which was omitted from this narration.

---

First published:

Source:

Narrated by TYPE III AUDIO.