A Basic Introduction to Speech Recognition (Hidden Markov Model & Neural Networks)

Рет қаралды 61,344

Күн бұрын

This video provides a very basic introduction to speech recognition, explaining linguistics (phonemes), the Hidden Markov Model and Neural Networks. In short: Speech Recognition for Dummies!
00:51 - From an analog to a digital environment
04:22 - Linguistics
05:47 - Hidden Markov Model
10:54 - Artificial Neural Networks
In August 2018, we organised a workshop at our job about speech recognition. Being in charge to explain the technology behind speech recognition, we created a dynamic session where we explained the concepts both in person and via video for a good visualization. To facilitate this interaction, we created digital versions of ourselves: e-Rodi & e-Hanni. Satisfied with the feedback during the workshop, we decided to also film the parts in between to make it a full story to share online. We hope this video can help you understand speech recognition better!
Enjoy!
Rodolphe & Hannes
NB: the probabilities used in the part of the HMM (e.g. 0,6; 0,9; ...) were just chosen to explain the concept and are not a true representation of the actual probabilities.
#speechrecognition #hiddenmarkovmodel #neuralnetworks #easilyexplained #speechprocessing

Пікірлер: 91

@dwyercbp6311 4 жыл бұрын

far and away THE best explanation of HMM or any speech recognition related thing i have ever seen. love your technique. simple, visual, full explained. why can't the other youtube vids be this instructive?

@user-ef2pv2du3j 8 ай бұрын

As a visual learner this is an amazing demonstration, thank you!!

@digabulful Жыл бұрын

This was BY FAR the best explanation I've seen on this topic. Thank you guys.

@WushuAII 3 жыл бұрын

Let's complete the homework, for my midterm assignment !!!!

@jtlunsford780 Жыл бұрын

You guys broke that down so even I could see how the big picture works. Thank you so much....JT

@kalagaarun9638 4 жыл бұрын

It was really cool way of expanation of HMM ... good one

@zeynepkurt2801 5 жыл бұрын

You guys have done a great job!

@karimbenyezza1029 4 жыл бұрын

concise, simple, awesome! Thank you!!!

@spp626 Жыл бұрын

I think this could be my first KZfaq video where I got a clearest explanation of this speech recognition..hats off to you both. Thanks a ton!!!

@gabriellaswan1917 4 жыл бұрын

I wish you guys made more of these videos, this is great

@madisonworley931 Жыл бұрын

Thank you so much for this!! I'm doing a project on speech recognition, and I was starting to wonder why I chose this topic because it's so overly complicated. But you guys made it easy to grasp and now I don't have to stress as much.

@ItsDanHere Жыл бұрын

Great video! Super easy to understand intro to HMM and NN. Exactly what I was looking for.

@jonasbayer 3 жыл бұрын

Wow, the explanation is very easy to understand and covers the topic in detail. Great job!!

@rajaanss3519 4 жыл бұрын

amazing presentation, and the info too, just love it.

@rapoliit 4 жыл бұрын

a great combo for sure...thanks a million.

@noveltyrose 3 жыл бұрын

thank you guys for doing this video you have saved me by helping me to understand HMM and ANN for my presentation, thank you SO MUCH

@takethewalk2128 2 жыл бұрын

Best video so far for a non technical to understand speech recognition

@016_areebashabir8 Жыл бұрын

this is the best explanation one can find on online platform. Much appreciated. Thankyou. May you rise and shine.

@niyongaboeric Жыл бұрын

I enjoyed it. You really simplified for me and I 'm thinking to go ahead and code my first speech recognition. Thanks fellas.

@nova4536 Жыл бұрын

Finally a much needed explanation.. As I've just started to get into a voice assistant from scratch in c++ lol 😂

@njerurichard3581 Жыл бұрын

Good luck in your journey. I'm trying the same thing.

@vertikasahu8468 2 жыл бұрын

Excellent explanation!! Thankyou for simplifying such complicated topics

@Sunny-qe5el Жыл бұрын

Thank a ton gentlemen for making this informative video. Concepts like Hidden Markhov Model is quite hard to understand but after watching this video I am quite comfortable with the topic of HMM Kudos to you and your team.

@muhammadsaimhashmi2425 4 жыл бұрын

Thanks alot man, you saved my day.

@richajaiswal3882 3 жыл бұрын

great explanation... I was working on a project of speech recognition and couldnt understand much from reading...but this helped a lot :)

@user-yk2zu5px3u Жыл бұрын

Thank you!!

@sidrode 2 жыл бұрын

Thanks a lot for not just the initiative but the clarity of the explanation too

@rohanshetgaonkar255 3 жыл бұрын

To good. This is what I was searching for past 2 days. Thanks ❤️

@DummBistDu 3 жыл бұрын

This is so well made, thank you guys!! :)

@amruthag8235 3 жыл бұрын

dude now I truly understand the concept. Liked it, keep going.

@manohar3626 4 жыл бұрын

thanks for the simple and short explanation

@arshiarahimi3779 4 жыл бұрын

Thank you! This was awesome

@estelleyu9445 2 жыл бұрын

Thank you so much for the detailed explanation!!! Great help for my research topic :)

@anubhavsingh2555 Ай бұрын

well done guys the way you explained this makes totaly worth to watch

@lancelotdsouza4705 2 жыл бұрын

Beautiful loved it,made it so simple

@jiayiwu4101 4 жыл бұрын

Thanks for your explanation! Very helpful! What does the input look like for NN? How does NN and HMM work together?

@ganeshkharad 4 жыл бұрын

thank you for this video....you have really done hard work to make this....i can clearly see that

@hayatt143 4 жыл бұрын

Great Video. Can you please give an example of sequential nature of speech and in what terms HMM is not flexible?

@ashishmehta1744 7 ай бұрын

This is the best explanation of the underlining principle.

@abdulkadiryapici5341 Жыл бұрын

Wonderful explanation with great level of understanding. Thanks for this beautiful content.

@asutosh123 2 жыл бұрын

excellent explanation! thank you.

@roshcoben8820 4 жыл бұрын

Thanks guys! Next for me is to check how deep learning is used for speech recognition

@nihatguliyev2837 2 жыл бұрын

Nice job, thank you both, guys.

@aryanyekrangi7093 3 жыл бұрын

Great video, helped me out!

@fahnub 2 жыл бұрын

that was a great explanation. so helpful.

@tobiaszpietrala8643 3 жыл бұрын

Thanks a lot guys!!!

@maikhellmeister54 4 жыл бұрын

Very good! Thank you guys. Da you have a video about the GMM+HMM?

@huyangquang1711 4 жыл бұрын

you should do more of this !

@JL-pg4pj Жыл бұрын

Best one, thanks a lot.

@ihsibustim5930 2 жыл бұрын

Best video about Speech Recognition by far

@chinmayakishore6219 3 жыл бұрын

brilliant mate.cheers.can u make a video on tacotron ,concatenative..and all the major methods too.

@micaismyname Жыл бұрын

Thank u so much for the video!! It was really helpfull!!

@aneesanaushad5654 2 жыл бұрын

so far the best explanation. KUDOS!!!

@roxy_2000 8 ай бұрын

very nice and interesting! the person pretending the model is so cute!

@amirmasoud_iravani Жыл бұрын

best video on asr ever!😍

@laibaaaqil7678 2 жыл бұрын

Exceptional Brilliant 👏 👏 👏 👏 👏 Thank you so much 💓 💗 💛 💖

@samrithkohli2272 2 жыл бұрын

Superb explanation

@williamgomez6226 2 жыл бұрын

Excelente more than excellent!!!

@mdbadiuzzamanbodi8250 2 ай бұрын

1:25 The video is supe helpful, just a tinay thought. I think calling analog to digital converter hardware would be appropriate .

@parth1450 4 жыл бұрын

Thanks people!!!

@dina4913 Жыл бұрын

Amazing, deserves more likes/views

@lotiumdupeuple3661 5 жыл бұрын

Great explanation thanks How is an hmm trained to define the good probabilities ?

@irfananizar5376 Жыл бұрын

Good explanation .. Need more topics..plss

@rabiadamlaozyer9622 Ай бұрын

That's a great informative video of such a complex system. Thank you so much! I wonder where you got the likeliest probabilities of phonemes and vowels with the values, though. Is there such a thing?

@rabiadamlaozyer9622 Ай бұрын

NB: the probabilities used in the part of the HMM (e.g. 0,6; 0,9; ...) were just chosen to explain the concept and are not a true representation of the actual probabilities. You guys already wrote it! you're the best!

@matth8355 2 жыл бұрын

So good 👌🏻

@niwdenus8256 4 жыл бұрын

Great stuff! Do you have any papers or resources to go from here?

@YouTubist666 Жыл бұрын

Nice job.

@renew44 4 жыл бұрын

1:40 I see what you did there! ;)

@shubhamg9495 Жыл бұрын

this video was so good

@noemimp4910 8 ай бұрын

Wow!! Very good communicatiors, great video, really helpful

@muhibullah4201 5 жыл бұрын

good job

@saimadhavp 5 жыл бұрын

Appreciate your nature of sharing your learning(s). Just curious to know if there are any good pointers to understand how HMM is trained for words with Phonemes. I mean how arrive at TranstionProbability matrix and Emission probability matrix. Is it through dictionary of words./Phonemes we train HMM ?

@joshuanandrekar 4 жыл бұрын

Hey, did you get any info about it? Was looking for the same thing

@florain5040 4 жыл бұрын

Many thanks for this introduction. Especially the HMM Model was perfect for me since I read about it and the formulas always got me confused.

@ritaheller921 3 жыл бұрын

I love you guys

@sarthakpatwari7988 6 ай бұрын

Thank you for such a wonderful explaination of such a tricky concept can you also please provide the code for the same ?

@melinagensler9582 4 жыл бұрын

First of all, thanks for the interesting and good explanation. Is there any reference for this? Any papers or books where you get that information from? I need it for a scientific research. Would be nice if any one could help me out :)

@two697 3 жыл бұрын

Did you find any?

@a2sirmotivationdoses782 3 жыл бұрын

Sir what is the Data set for Speech To Text Machine learning Model?

@rajeevanand5731 8 ай бұрын

could you share a video that compares the performance of a markov chain on speech recognition and then a HMM doing speech recognition on the same speech

@Hannesvl 8 ай бұрын

Thanks for watching and sharing your question! This video and our related knowledge was the result of a one-off workshop, so unfortunately our expertise in this matter doesn’t reach further than what we’ve put in the video. Good luck with your further research!

@gdevification 3 жыл бұрын

Can you share your workshop?

@ediwidodo499 Жыл бұрын

how to calculate the probabilities used in real application ? is it stored in a database or how ?

@Hannesvl Жыл бұрын

I'm sorry, I could make a guess, but I don't have an actual answer to that. This video is the result of a crash course we made as a peer-to-peer learning exercise, although this is not my field of expertise (at all). Since it has been some years now, I don't know/remember anything more than captured in this video. I do hope this video helped you in getting the basics, but for more advanced questions & information you'll have to look elsewhere. Good luck!

@Democracy_Manifest Жыл бұрын

You could use a language model to calculate the probabilities.