Bayesian Reasoning and Rule of Conditionalization

[Thought Art]

Hey, can anyone explain this to me in a simple straightforward way as I am not sure I understand it. 
 

Rule of a Conditionalization: 

‘Upon learning evidence e,P*new*(H) should equal P(H/e)’ 

I understand it explained out that upon learning new evidence the possibility that our hypothesis is true improves or gets worse?. So, when our theory makes correct predictions then… we increase the confidence in our models, hypothesis  and world view? 

This the Bayesian view of scientific inference right? 
 

I find the algebra a bit mind confusing and will need to sit with it for awhile I think. I understand the logic when it’s explained out but the algebra confuses me a bit. 

 

 

Edited June 12, 2022 by Thought Art

[Thought Art]

Is this why Leo says everyone uses rationality and why rationality is so limited ?

 

I don’t really grasp the final sentence though. Basically deduction through probability as an escape from induction? Doesn’t work and points to something higher… because Bayesian reasoning assumes our initial creedences are subjective (this seems tangental to inductive)? 
 

not knowing?

Edited June 12, 2022 by Thought Art

[Nilsi]

19 hours ago, Thought Art said:

Hey, can anyone explain this to me in a simple straightforward way as I am not sure I understand it. 
 

Rule of a Conditionalization: 

‘Upon learning evidence e,P*new*(H) should equal P(H/e)’ 

I understand it explained out that upon learning new evidence the possibility that our hypothesis is true improves or gets worse?. So, when our theory makes correct predictions then… we increase the confidence in our models, hypothesis  and world view? 

This the Bayesian view of scientific inference right? 

It just means that when you gather new evidence you still take all the previous evidence (= Base-Rate) into consideration. The common human bias is neglecting the Base-Rate and just taking everything we hear and see as being absolute evidence.

Bayesian reasoning is basically completely antagonistic to mysticism but it's definitely useful to not make foolish decisions in the world. 

The point is not that it is right or wrong. It's useful for approximating the confidence you should have in certain events being true/happening versus not.

Edited June 13, 2022 by Nilsi

[Leo Gura]

Just throw Bayesian reasoning in the trash. People who use it are idiots.

[flyingguitarist]

@Leo Gura

33 minutes ago, Leo Gura said:

Just throw Bayesian reasoning in the trash. People who use it are idiots.

 why?

[Nilsi]

1 hour ago, Leo Gura said:

Just throw Bayesian reasoning in the trash. People who use it are idiots.

So the next night you go out approaching girls and don't take one home with you, you are going to tell yourself that you are incapable of picking up girls?

Point is, you are using bayesian logic intuitively anyways.

The average person would greatly benefit from understanding bayesian logic and making it more conscious. It would prevent a lot of foolishness.

[Thought Art]

Yeah, whether idiots use it or not I wanna understand it because I think it’s actually a facet of my thinking anyway 

[Thought Art]

4 hours ago, Nilsi said:

It just means that when you gather new evidence you still take all the previous evidence (= Base-Rate) into consideration. The common human bias is neglecting the Base-Rate and just taking everything we hear and see as being absolute evidence.

Bayesian reasoning is basically completely antagonistic to mysticism but it's definitely useful to not make foolish decisions in the world. 

The point is not that it is right or wrong. It's useful for approximating the confidence you should have in certain events being true/happening versus not.

I don’t think it is antagonistic the mystic perspective that I have because my mystic perspective includes science within its scope. 
 

Doesn’t it really mean as we gain new evidence it affects the Creedence on our initial hypothesis?

[Nilsi]

1 hour ago, Thought Art said:

 

Doesn’t it really mean as we gain new evidence it affects the Creedence on our initial hypothesis?

I think the theorem is best explained with an example. Say I want to know how confident I should be that I am enlightened if i have had a mind-blowing psychedelic trip (P A/B). I now take the probability of being enlightened (P A) times the probability of having a mind-blowing psychedelic trip given that I am enlightened (P B/A) divided by the probability of having a mind-blowing psychedelic trip. 

This puts things into perspective since being enlightened is probably much less probable than having a mind-blowing psychedelic trip. The evidence of having had a mind-blowing psychedelic trip is now put into context of the probability of being enlightened thus weakening it considerably. 

This also answers why bayesian reason is not the appropriate tool for spiritual work as logically you would always have to doubt yourself and talk yourself out of actually realizing the Absolute.

[AtheisticNonduality]

3 hours ago, Nilsi said:

I think the theorem is best explained with an example. Say I want to know how confident I should be that I am enlightened if i have had a mind-blowing psychedelic trip (P A/B). I now take the probability of being enlightened (P A) times the probability of having a mind-blowing psychedelic trip given that I am enlightened (P B/A) divided by the probability of having a mind-blowing psychedelic trip. 

This puts things into perspective since being enlightened is probably much less probable than having a mind-blowing psychedelic trip. The evidence of having had a mind-blowing psychedelic trip is now put into context of the probability of being enlightened thus weakening it considerably. 

This also answers why bayesian reason is not the appropriate tool for spiritual work as logically you would always have to doubt yourself and talk yourself out of actually realizing the Absolute.

Probabilities and confidences in them work when you have objects in relation, or subjects relativized by an outside world. The certainty and Absolution of enlightenment negates that. @Thought Art That's what makes it antithetical to mystical states, but mystical states may also include physical and mental science constructions / equations / events; ie. "Brahman is the world." This probabilistic content is transcendable but also are part of the complete reality, as just a part.

7 hours ago, Leo Gura said:

Just throw Bayesian reasoning in the trash. People who use it are idiots.

That isn't very nice.

6 hours ago, Nilsi said:

Point is, you are using bayesian logic intuitively anyways.

They aren't good at it.

[something_else]

9 hours ago, Nilsi said:

Point is, you are using bayesian logic intuitively anyways

Humans are completely and utterly terrible at intuiting probability theory

I don’t know if I’d say Bayesian logic is useless. I remember from uni that it has some excellent applications for creating ‘AI’ that behaves perfectly rationally based on the info it has available

And scientifically it’s pretty useful too

[Thought Art]

6 hours ago, Nilsi said:

I think the theorem is best explained with an example. Say I want to know how confident I should be that I am enlightened if i have had a mind-blowing psychedelic trip (P A/B). I now take the probability of being enlightened (P A) times the probability of having a mind-blowing psychedelic trip given that I am enlightened (P B/A) divided by the probability of having a mind-blowing psychedelic trip. 

 

Is this coherent?

Isn’t it more when new evidence comes into play given a hypothesis you already have? or is that the law of Conditionalization. 
 

I’ll take a look when I have free time tomorrow to look at this is in more detail

Edited June 14, 2022 by Thought Art

[Thought Art]

2 hours ago, AtheisticNonduality said:

Probabilities and confidences in them work when you have objects in relation, or subjects relativized by an outside world. The certainty and Absolution of enlightenment negates that. @Thought Art

 

I know.