Keyboard Shortcuts?f

×
  • →Next step
  • ←Previous step
  • ↓Skip this slide
  • ↑Previous slide
  • mShow slide thumbnails
  • tShow transcript (+SHIFT = all transcript)
  • nShow notes (+SHIFT = all notes)

Please be cautious in using the transcripts.

They were created mechanically and have mostly not been checked or revised.

Here is how they were created:

  1. live lecture recorded;
  2. machine transcription of live recording;
  3. ask LLM to clean up transcript, and link to individual slides.

This is an error-prone process.

Click here and press the right key for the next slide.

(This may not work on mobile or ipad. You can try using chrome or firefox, but even that may fail. Sorry.)

also ...

Press the left key to go backwards (or swipe right)

Press n to toggle whether notes are shown (or add '?notes' to the url before the #)

Press m or double tap to slide thumbnails (menu)

Press ? at any time to show the keyboard shortcuts

 

Expected Utility

in this lecture we're going to explore how some formal methods borrowed from decision theory might enhance our understanding of the discoveries about action
we're also going to explore, conversely, how those discoveries might bear on our use of these formal models.
Although this is going to seem quite dry at the beginning, when i taught this course in the past this was one of the areas that most interested people. So do bear with me.
insert-transcript#f0fb7d60-3bec-4d36-90d0-47b0718cad59-here

decision theory

How do rational agents decide which of several available actions to perform?

simple question: How do rational agents decide which of several available actions to perform ?
insert-transcript#be4f04e2-51de-4d7f-897b-c19d88d7ad2a-here
insert-transcript#36954e84-7fba-40cd-9191-e97ec9d37531-here
One is a lottery ticket, the other a coin toss
According to Lanchester’s novel Mr Phillips, I’d need to buy a lottery ticket at most 11 minutes before the draw in order for my chance of winning to be greater than my chance of dying before the draw.
Most people here (16-24) could buy a ticket as much as 68 minutes before the draw to achieve the same result.
insert-transcript#ec04d8c9-74c7-4772-ac15-9de9368c26e1-here
insert-transcript#c1621be8-233e-4b43-b3eb-972be08405b3-here
Expected utilities can be derived from desirabilities and probabilities. One procedure for choosing an action is to compute expected utilities (in the way illustrated) and then perform the action with the highest expected utility.
insert-transcript#0e9f2a9d-e637-4391-a783-c65b992f701f-here

Terminology

actions have outcomes

(Jeffrey: outcome = ‘consequence’)

which outcome an action causes depends on **conditions**

subjective probabilities attach to conditions.

preferences rank outcomes

expected utilities attach to actions.

(Jeffrey: expected utility = ‘estimated desirabilities’)

Jeffrey (1983)

Make sure you understand the terminology and can relate it to the scenario.
Be sure to use the terminology consistently, and with precision, in your writing.
insert-transcript#a6ff8b87-c3d9-4740-81a6-8f932becf499-here
It is important to understand that we haven’t done anything more than introduce a way of representing things.

so far: the representation

next step: the theory

This book has ‘a philosophical end: elucidation of the notions of subjective probability and subjective desirability or utility’

(Jeffrey, 1983, p. xi)

‘we [...] view
subjective values and probabilities
as interrelated constructs of decision theory’

(Davidson, 1974, p. 146)

Why is what we have done so far not sufficient?
(Because we have not anchored preference or subjective probability to subjects.)
(We have not said anything at all about what it is for someone to have a particular preference, nor what it is for them to assign a particular subjective probability to a given condition.)
So what we've done so far is just talked about representation. We just talked about purely formal representation. We haven't done what Jeffrey and Davidson say they can do, which is to anchor a shared understanding of belief and desire. The reason we haven't done that, I think, is relatively simple. We have assumed that we already know what preferences and subjective expectations are, and how they connect to individual people.
So what we've got is a purely formal way of representing things. We haven't got any way of connecting that to any particular person. And therefore we can't say that we have found an articulation of anything.
But just here I want to stop for 90s.
There's two things that you might discuss with the person next to you or behind you. It might be, why is it that we've only got the representation? What's the thing that we need?
I've told you what we need is the connection between the representation and the person. But it might be that that doesn't make a lot of sense at this stage.
On the other hand, if that's completely clear to you, then the question is how do you take the next step? What's the next step that I'm going to offer you in connecting up the formal representation to the individual person and their actions?