(ML 14.12) Viterbi algorithm (part 2)

[@kjpr5641]

Just wanted to acknowledge how great these series of videos on the HMM have been. Really enjoyed how you linked up the formula derivation with how it translates at a more intuitive level.

[@rishabhdev4666]

Whole ML.14 series just G.O.A.T

[@caseyli5580]

Little path image was so helpful. Explained in 14 minutes what a 1 hour lecture could not. Thank you!

[@keli5670]

Thank you for the videos! They are the most clear explanations on HMM I have ever seen

[@phananh101010]

Thanks so much for your video, make me understand the materials really faster than reading Theodoridis's book.

[@matifnawaz]

Great lecture. Thank you very much. I like that you don't just deliver the lecture, you convince your students on what you teach. Best wishes.

[@biturboism]

You are a true educator, not just a teacher! Thank you!

[@julianserban]

These videos are great. They've helped me understand the basic theory of HMMs very quickly. Thanks a lot mathematicsmonk!

[@ብሌናይጻዕዳ]

you saved my day...thank you after 9 yrs

[@subhomoyghosh9074]

your thought process is extremely helpful..... thank you!!!

[@q0x]

I am missing Baum-Welch but still a very good lecture.

[@NazerkeSafina]

Thank you. Very clear explanation

[@VanTeeeee]

What a great HMM series! Thank you

[@orchisamadas2222]

Good video!Please can you upload an explanation of Baum Welch algorithm?

[@zloop]

have just pushed through all the hmm vids,,,awesome job!! thx!

[@ruiwang862]

very helpful for the first time.

[@med0ize]

cheers bro. great videos.

[@farzind25]

Thanks for the great job! A point that I did not get is how to calculate the transition, emission and initial values from the training data after all?

[@mikel5264]

Impresive!

[@saurabhdubey6656]

That's a really good explanation but the visual representation of the most probable path is incorrect. Since mu_n is recursively calculated, the path upto x_n-1 is fixed and we can't have multiple most probable paths that could lead to the red point in the plot.

[@artomeri7266]

The only thing missing from HMM set of lectures is Baum-Welch or MCMC to estimate transition/emission distribution parameters :)

[@juaneugeniodebenedettti6187]

yes true, is a shame to not have the Baum-Welch algorithm, do you know a good video to wath it?

[@ruebena9640]

good explanation

[@haoguoxuan411]

Could somebody explain the transformation at 4:35?

[@Rijndhadu]

can u please explain the backtracking thing in viterbi algorithm :)