Sir, you are a hero. I read a bunch of definitions, explanations and examples and only yours can make me really understand MCMC. Now I can continue my final assignment
I spent dozens of hours reading papers about MCMC. all that is sh... RUclips - the best source of any knowledge. Evidence of this - is the lecture above. Well done, author, well done... Thanks
Yeah I see you, League of Legends. hiding out there in the task bar-- thinking you'll just chill until Mr. Picton gets some free time. Well this great intellect has moved on. When given a choice between toxicity and flaming or creating helpful videos, I'll have you know, Jeff Picton chose the high road.
Thanks for the code. As a programmer, seeing how something would be coded makes a lot more sense than seeing a mathematical formula. :) The last example was also quite useful and a great way to tie it all together.
This is an awesome lecture that clears the mcmc concept. I am curious to know how can we apply it in partitioning of jobs on 2 parallel machines in order to minimize makespan?
Just in case it helps someone watching this very good video, here is some R code to demonstrate the Metropolis algorithm: # Metropolis algorithm -- Gaussian distribution library(ggplot2) mu
Some "Professors" teach students just to show how much they know about the topic, by using alien language (edit: but some are good prof). I spent hours in those language, but instead i can understand mcmc within 36 minutes. You're a superhero!!
Great video! Do you know how I would use the Metropolis algorithm to select random points from the tails of a Normal Distribution (or do we always have to sample from a Uniform distribution?) at a higher probability than selecting points close to the mean? i.e. I need the target distribution to be a Normal Distribution and the proposed Distribution to be the tails ((-4*sigma, -3*sigma) and (3*sigma 4*sigma)) of the Normal Distribution? Is this possible? Thanks a lot!
Thank you so much! this vid is really helpful Can you explain why the alrogithm(22:28) creates N(0,1) instead of N(0,10) or N(0,140), etc...? is it because that the normpdf is based on N(0,1)?
Hey Jeff, how does the software construct the normpdf of x(i) and x_c in the gaussian code example? Considering we start off with only a single x(i) value and then sample a single point x_c, how can one create an entire pdf to be used in the equation?
Hi, great tutorial, thanks. I have a couple of doubts 29'30" About the initial guess, what literature can I read to determine such a value of the initial guess? 30' About proposal distribution and the cost function, is there any other tutorial or literature to understand how to design such a proposed distribution or using exp(-cost) should suffice considering a wide range of phenomena and datasets? Thanks again
Thank for the video, I have some questions. Let's say that we didn't know that the distribution was gaussian, how do we decide what proposal distribution to use? Even if we knew that the distribution is gaussian, how did you know to use normpdf (which already centers at 0 with sigma of 1) ? If the actual distribution was N(2,1) instead, would you still use normpdf ?
everythig was brilliant!! great job.. m interested also in knowing your approach to the functions step_param and ebm_model while it could explain a more clearer picture.. Thanks in advance.
Typically all of the molecules would be altered at once, as the position of each molecule is a variable parameter and the collection of these constitutes a state of the system. I described moving them individually to simply convey the intuition of making small changes to the system. But, my intuition tells me that selecting single molecules with random reselection would be fine and preserve ergodicitiy.
You say the method will visit the nodes an amount proportional to "their probability" many times. But we don't give any probability to the nodes a-priori, so really the output of the method *defines* this "per node probability" no?
at 15:33 the first product between X0=[0.5 0.2 0.3] with T is not equals to [0.2 0.6 0.2]. actually it is [ 0.18 0.64 0.18], and converges to [0.2213 0.4098 0.3689]. am I missing missing some thing?
Interesting. About the climate example. Wouldn’t cloud formation be important since albedo was and perhaps that would be more important than the feedback, or just as?
The town of Monte Carlo is in the tiny principality of Monaco (that is, a territory originally ruled by a prince) on the Mediterranean coast of France. Monte Carlo was -- and still is -- famous for its iconic, palatial gambling casino.
Good job, but you missed the punch line at 7:10 that a histogram of the number of times you land in an interval matches the shape of the curve; i.e., the number of times is a maximum in an interval centered at 0 and falls off in both directions. Maybe it was obvious to others, but maybe I'm a little slow.
Nicely done. Would have been better if the Hastings correction to alpha was discussed. It was mentioned and even kept in the presentation, but then neglected. Seems either losing it, and justifying the loss would be good, or leaving it out would be better.
@@picjeffton When I find the product of the starting state X0 and the Markov transition matrix I do not get that the probabilities of the next state X1 are as shown [0.2, 0.6, 0.2] but rather [0.18, 0.64, 0.18]. Am I doing the multiplication wrong or is that part of the arithmatic error? Thanks for your help and the video.
I don't see the difference between irreducible and aperiodic. IMO the graph is aperiodic (in the sense that there is no subgraph where we will get stuck) iff it is irreducible (for every pair of states (x,y), x and y are mutually reachable with nonzero probability).
Paul Frischknecht irreducible is probability of reaching any state while starting at another state is positive. The periodicity, d, is the largest integer such that returning to a certain state i is always a multiple of d. ie if you can reach i after {2,4,6,8,10} steps then d=2 since {2,2(2),2(3),2(4)..} .. An aperoidic MC would be {2,3,4,6,7} here then is no d such that n*d will generate the periods.
The video is great, but why would you think that the name comes from a casino in Las Vegas and not from the original one in Monaco, that the american one was named after?? 😂😂
Jeff, you're fantastic for doing this. I've been struggling all semester trying to grasp this concept intuitively. I've finally seen the light
One of the best explanations of MCMC I have seen on the web. Wonderful job . Wonderful
This video is cool! I really like the two examples you give, illustrating the idea of MCMC concretely and clearly. Thanks!
Sir, you are a hero. I read a bunch of definitions, explanations and examples and only yours can make me really understand MCMC. Now I can continue my final assignment
Great lecture. Thanks for showing code. This is the most straight forward MCMC tutorial that I've seen on youtube thus far.
Excellent presentation. Very clear, with nice examples and simple codes. Thank you.
easily the best MCMC explanation I've seen, huge thanks
A great lecture indeed! Thanks very much :) The matlab code you shared really made it as clear as it gets. Keep them coming :)
Nice work!
You explained very clearly and the recording is also very nicely done...
This video might be the only thing saving my thesis. Thanks :D
I spent dozens of hours reading papers about MCMC. all that is sh...
RUclips - the best source of any knowledge. Evidence of this - is the lecture above.
Well done, author, well done...
Thanks
Yeah I see you, League of Legends. hiding out there in the task bar-- thinking you'll just chill until Mr. Picton gets some free time. Well this great intellect has moved on. When given a choice between toxicity and flaming or creating helpful videos, I'll have you know, Jeff Picton chose the high road.
Thanks for the code. As a programmer, seeing how something would be coded makes a lot more sense than seeing a mathematical formula. :) The last example was also quite useful and a great way to tie it all together.
Great video, Jeff Picton. It was really helpful. Thank you very much.
Great video. Thank you, Jeff!
Thank you. This really helped my understanding the model and the applications.
Great work, almost answered all my questions. Thanks
Thanks for this great explanation!
Thank you so much for a very clear explanation
Thanks so much for your inspiring explanation!
Something that presents MCMC in a concise and clear way. Like it a lot.
Thank you very much for your great lecture. It is really being of great help for me.
It is really being of great help for me, thank you very much!
Thank you for your clearness. Now I really have an idea.
That is a really good explanation! Thank you!!!
Very clear explaination!
Clear explaination. Thank you Jeff. A question existing in my brain is: How do we set the judgement criteria during the model iteration?
ths is so great, thank you!
So nicely explained!!!!! Thank you !!!!
Very very good explanation!! Thanks! :D
This is an awesome lecture that clears the mcmc concept. I am curious to know how can we apply it in partitioning of jobs on 2 parallel machines in order to minimize makespan?
Dude, You are awesome.
Just in case it helps someone watching this very good video, here is some R code to demonstrate the Metropolis algorithm:
# Metropolis algorithm -- Gaussian distribution
library(ggplot2)
mu
Great Lecture. Thankyou.
i couldnt agree more with you! Well done author!!
Some "Professors" teach students just to show how much they know about the topic, by using alien language (edit: but some are good prof). I spent hours in those language, but instead i can understand mcmc within 36 minutes. You're a superhero!!
Great video! Do you know how I would use the Metropolis algorithm to select random points from the tails of a Normal Distribution (or do we always have to sample from a Uniform distribution?) at a higher probability than selecting points close to the mean? i.e. I need the target distribution to be a Normal Distribution and the proposed Distribution to be the tails ((-4*sigma, -3*sigma) and (3*sigma 4*sigma)) of the Normal Distribution? Is this possible?
Thanks a lot!
Very well done!
Jeff thank you for these videos. I am self studying MCMC and is there any chance that you may share the simulation codes?
Brilliant tutorial !
Excellent explanation
Great presentation. Thnx
Well done! Thank you
Thank you very much, that helped my a lot!
I agree. I just didn't feel like opening latex to write out the equation and just took a screen cap of it from a paper I had.
nice simple examples. thank u
Thank you so much! this vid is really helpful
Can you explain why the alrogithm(22:28) creates N(0,1) instead of N(0,10) or N(0,140), etc...? is it because that the normpdf is based on N(0,1)?
@3:00 Wiki says it's from Monte Carlo in Monaco.
Hey Jeff, how does the software construct the normpdf of x(i) and x_c in the gaussian code example? Considering we start off with only a single x(i) value and then sample a single point x_c, how can one create an entire pdf to be used in the equation?
Thank you so much!
thank you. Really helped me
Hi, great tutorial, thanks.
I have a couple of doubts
29'30" About the initial guess, what literature can I read to determine such a value of the initial guess?
30' About proposal distribution and the cost function, is there any other tutorial or literature to understand how to design such a proposed distribution or using exp(-cost) should suffice considering a wide range of phenomena and datasets?
Thanks again
excellent tutorial
So at 19:30, the q distribution is equivalent to the transition matrix T from the markov chain formalism at 14:00, right?
very coherently explained, most videos go into unnecessary esoteric detail.
in 31:41, have you any evidence for choose the accept/reject test function like this? If yes, where can i find it?
You made it sound easy.
Thank for the video, I have some questions. Let's say that we didn't know that the distribution was gaussian, how do we decide what proposal distribution to use? Even if we knew that the distribution is gaussian, how did you know to use normpdf (which already centers at 0 with sigma of 1) ? If the actual distribution was N(2,1) instead, would you still use normpdf ?
The Markov condition is not "x4 depends only on x3", but "if we know x3, x4 becomes independent of x2 and x1"
At 9:58, when "another" molecule is chosen, was that through a process that had a chance of reselecting the first molecule again?
everythig was brilliant!! great job.. m interested also in knowing your approach to the functions step_param and ebm_model while it could explain a more clearer picture.. Thanks in advance.
one question: How do we choose the proposal q? Is there any requirements for this choice?
Great video - do you have any more from this class?
Typically all of the molecules would be altered at once, as the position of each molecule is a variable parameter and the collection of these constitutes a state of the system. I described moving them individually to simply convey the intuition of making small changes to the system. But, my intuition tells me that selecting single molecules with random reselection would be fine and preserve ergodicitiy.
Make more videos, Jeff!
Well there is a Monte Carlo in Vegas... but ya you're right.
At last a clear explanation of this
Gréât vidéo 👍🏻
I liked the song
nice video.
very good
very programm monte carlo
I'm also confused. A clarification about it would be welcome.
Have you got a written program for the disks?
I need the document presented in the video. Any help? Thanks
What is the explicit equation for the energy balance model?
You say the method will visit the nodes an amount proportional to "their probability" many times. But we don't give any probability to the nodes a-priori, so really the output of the method *defines* this "per node probability" no?
at 15:33 the first product between X0=[0.5 0.2 0.3] with T is not equals to [0.2 0.6 0.2]. actually it is [ 0.18 0.64 0.18], and converges to [0.2213 0.4098 0.3689]. am I missing missing some thing?
I guess you're right! I also realized that, that's why I decided to read the comments!
Interesting. About the climate example. Wouldn’t cloud formation be important since albedo was and perhaps that would be more important than the feedback, or just as?
I didn't understood the difference between Metropolis algorith and MCMC?
May I ask why?
Great explanation , but....MONTE CARLO IS IN MONTE CARLO , not in LAS VEGAS :-)
You Americans, so egocentric :-P
Very clear video BTY. Thanks!
Ah the legendary quality of American public education.
Yes! Monte Carlo is in Africa, and Africa is some place in the south of Europe. No?
The town of Monte Carlo is in the tiny principality of Monaco (that is, a territory originally ruled by a prince) on the Mediterranean coast of France. Monte Carlo was -- and still is -- famous for its iconic, palatial gambling casino.
Good job, but you missed the punch line at 7:10 that a histogram of the number of times you land in an interval matches the shape of the curve; i.e., the number of times is a maximum in an interval centered at 0 and falls off in both directions. Maybe it was obvious to others, but maybe I'm a little slow.
Nicely done. Would have been better if the Hastings correction to alpha was discussed. It was mentioned and even kept in the presentation, but then neglected. Seems either losing it, and justifying the loss would be good, or leaving it out would be better.
Nice lecture, whats your Elo? :p
Taleb brought me here; the Kali Yuga keeps me grinding.
snazzy intro music
I tried running that matrix program. Unfortunately it doesn' tconverge to (0.2, 0.4, 0.4) as you said. I don't know what I am doing wrong.
Jonathan Small I messed up the arithmetic in that example.
He he. Actually I tried again. This time using in-built matrix multiplication function in Python. It worked. Thanks :)
@@picjeffton When I find the product of the starting state X0 and the Markov transition matrix I do not get that the probabilities of the next state X1 are as shown [0.2, 0.6, 0.2] but rather [0.18, 0.64, 0.18]. Am I doing the multiplication wrong or is that part of the arithmatic error? Thanks for your help and the video.
Can someone explain this?
a good one! thank you!
thx alot ! also good old league of legends days XD
You're quite right. For the purposes of this video though, let's just pretend that is how arithmetic works.
monte carlo name came from famous casino in monaco not vegas i believe.
I get the same answer.
I don't see the difference between irreducible and aperiodic. IMO the graph is aperiodic (in the sense that there is no subgraph where we will get stuck) iff it is irreducible (for every pair of states (x,y), x and y are mutually reachable with nonzero probability).
Paul Frischknecht irreducible is probability of reaching any state while starting at another state is positive. The periodicity, d, is the largest integer such that returning to a certain state i is always a multiple of d. ie if you can reach i after {2,4,6,8,10} steps then d=2 since {2,2(2),2(3),2(4)..} .. An aperoidic MC would be {2,3,4,6,7} here then is no d such that n*d will generate the periods.
Jeff, I have an algorithm that I need help interpret.
Hey you play League of Legends too bro???
me too hahaaa
18:24 u are welcome
Learn more about logistic regression with R
drive.google.com/file/d/1qcq_186AMe2XK9aNiSLxLbvXlAmryWXX/view?usp=sharing
The video is great, but why would you think that the name comes from a casino in Las Vegas and not from the original one in Monaco, that the american one was named after?? 😂😂