What applications are there for Bayesian Statistics in ML? Does it apply to Deep Learning?
How is Bayesian Statistics related to Machine learning?
Eli Hill
Matthew Ward
Many, e.g. shitpost detection
Yes, keeping track of probability is very useful but complicates things; it would be annoying as fuck to implement while keeping track of state
Tyler Gonzalez
> it would be annoying to implement while keeping track of state
What do you mean by that?
Nathan Turner
you have to keep the probability of being in a state
real bayesian deep learning would be fucking difficult to implement with current algorithms and hardware. You approximate it with estimators.
Look into POMDPs - pretty much the 'bayesian' approach to deep learning
Naive Bayes for general ML is pretty well known
Jack Rivera
so for example you use a normal distribution as the likelihood of being right, and with bayesian inference you reduce the standard deviation to be more and more certain of your beliefs
there are other tools, bayesian inference isn't the only one used, but it has the advantage of being very fast on computers
Noah Cruz
Bayes and fisher would btfo of you
James Hall
>Look into POMDPs - pretty much the 'bayesian' approach to deep learning
Dude what? No... no at all
Isaiah Bailey
Bayesian deep learning = deep learning with a prior
Josiah Hill
>Look into POMDPs - pretty much the 'bayesian' approach to deep learning
>Naive Bayes for general ML is pretty well known
Ethan Howard
good way to put it
Jace Kelly
i fucking hate you ML brainlets, especially bayesian ones
probably couldn't even write bayes theorem on a chalkboard
Isaac Rivera
>Dude what? No... no at all
Yes, yes at all.
Bayesian RL can be (and usually is) modeled as a POMDP
Or we can ignore the fact that the belief estimator in a POMDP is maintaining a probability distribution using Bayes' rule.
Asher Adams
>chimes into thread without any constructive input
>implies someone is a brainlet
kek
A model using posterior probability distribution explicitly isn't bayesian?
Naive Bayes is a pretty common usage of bayesian stats for ML applications, dunno what your getting at
How would you answer OP's question user?
John Ross
also papers that you may be interested in:
this one is pretty comprehensive
people.eecs.berkeley.edu
jmlr.org
people.eecs.berkeley.edu
cs.cmu.edu
also to clarify:
the belief/value/policy estimators in the MDP/POMDP are typically NNs in practice (e.g. AlphaGo, Deep Blue, other deepmind stuff)
Also note that normal MDPs also hold a probability distribution as well, just not explicitly
Connor Wood
Bayesian RL =/= Bayesian Deep Learning
There is nothing inherently bayesian about POMDPs
You are just confused friendo
Asher James
In practice it does. Deep Learning being a buzzword for using multi-layered NNs.
>There is nothing inherently bayesian about POMDPs
the 'M' in MDP, the markov property, is what makes it inherently bayesian.
en.wikipedia.org
The state transition probability term is what gives MDP formulations the markov property.
This term is a representation of the *prior* of different states.
A prior is a form of bayesian statistical inference.
There is a reason that POMDP literature refers to POMDP states as prior belief states(because the transition term is explicit) and updates the states using the Bayes Rule.
How is this wrong?
Jonathan Hill
Why is she leeching off the labor of other teammates? Typical attention whore.
Jacob Green
Inverse problems boiiiiii
Samuel Jones
MDPs are essentially like a "data structure" used for modeling. They can be used with non-Bayesian methods.
You seem like you have a general knowledge of the field, but you are making statements that just don't make sense.
Carson Hill
Thanks!
Ian Fisher
What about free energy minimising.