Тёмный

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

Hannes van Lier
Подписаться 500
Просмотров 62 тыс.
50% 1

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

Наука

Опубликовано:

 

1 авг 2024

Поделиться:

Ссылка:

Скачать:

Готовим ссылку...

Добавить в:

Мой плейлист
Посмотреть позже
Комментарии : 91   
@dwyercbp6311
@dwyercbp6311 4 года назад
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
@user-ef2pv2du3j 8 месяцев назад
As a visual learner this is an amazing demonstration, thank you!!
@spp626
@spp626 Год назад
I think this could be my first RU-vid video where I got a clearest explanation of this speech recognition..hats off to you both. Thanks a ton!!!
@digabulful
@digabulful 2 года назад
This was BY FAR the best explanation I've seen on this topic. Thank you guys.
@016_areebashabir8
@016_areebashabir8 Год назад
this is the best explanation one can find on online platform. Much appreciated. Thankyou. May you rise and shine.
@takethewalk2128
@takethewalk2128 2 года назад
Best video so far for a non technical to understand speech recognition
@karimbenyezza1029
@karimbenyezza1029 4 года назад
concise, simple, awesome! Thank you!!!
@sidrode
@sidrode 2 года назад
Thanks a lot for not just the initiative but the clarity of the explanation too
@ItsDanHere
@ItsDanHere Год назад
Great video! Super easy to understand intro to HMM and NN. Exactly what I was looking for.
@jonasbayer
@jonasbayer 3 года назад
Wow, the explanation is very easy to understand and covers the topic in detail. Great job!!
@rohanshetgaonkar255
@rohanshetgaonkar255 3 года назад
To good. This is what I was searching for past 2 days. Thanks ❤️
@WushuAII
@WushuAII 3 года назад
Let's complete the homework, for my midterm assignment !!!!
@zeynepkurt2801
@zeynepkurt2801 5 лет назад
You guys have done a great job!
@madisonworley931
@madisonworley931 Год назад
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.
@rapoliit
@rapoliit 4 года назад
a great combo for sure...thanks a million.
@noveltyrose
@noveltyrose 3 года назад
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
@gabriellaswan1917
@gabriellaswan1917 4 года назад
I wish you guys made more of these videos, this is great
@vertikasahu8468
@vertikasahu8468 3 года назад
Excellent explanation!! Thankyou for simplifying such complicated topics
@rajaanss3519
@rajaanss3519 4 года назад
amazing presentation, and the info too, just love it.
@jtlunsford780
@jtlunsford780 Год назад
You guys broke that down so even I could see how the big picture works. Thank you so much....JT
@ashishmehta1744
@ashishmehta1744 7 месяцев назад
This is the best explanation of the underlining principle.
@niyongaboeric
@niyongaboeric Год назад
I enjoyed it. You really simplified for me and I 'm thinking to go ahead and code my first speech recognition. Thanks fellas.
@manohar3626
@manohar3626 4 года назад
thanks for the simple and short explanation
@anubhavsingh2555
@anubhavsingh2555 Месяц назад
well done guys the way you explained this makes totaly worth to watch
@amruthag8235
@amruthag8235 3 года назад
dude now I truly understand the concept. Liked it, keep going.
@ihsibustim5930
@ihsibustim5930 2 года назад
Best video about Speech Recognition by far
@aneesanaushad5654
@aneesanaushad5654 2 года назад
so far the best explanation. KUDOS!!!
@lancelotdsouza4705
@lancelotdsouza4705 2 года назад
Beautiful loved it,made it so simple
@DummBistDu
@DummBistDu 3 года назад
This is so well made, thank you guys!! :)
@abdulkadiryapici5341
@abdulkadiryapici5341 Год назад
Wonderful explanation with great level of understanding. Thanks for this beautiful content.
@estelleyu9445
@estelleyu9445 2 года назад
Thank you so much for the detailed explanation!!! Great help for my research topic :)
@Sunny-qe5el
@Sunny-qe5el Год назад
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.
@kalagaarun9638
@kalagaarun9638 4 года назад
It was really cool way of expanation of HMM ... good one
@muhammadsaimhashmi2425
@muhammadsaimhashmi2425 4 года назад
Thanks alot man, you saved my day.
@ganeshkharad
@ganeshkharad 4 года назад
thank you for this video....you have really done hard work to make this....i can clearly see that
@richajaiswal3882
@richajaiswal3882 3 года назад
great explanation... I was working on a project of speech recognition and couldnt understand much from reading...but this helped a lot :)
@nihatguliyev2837
@nihatguliyev2837 2 года назад
Nice job, thank you both, guys.
@asutosh123
@asutosh123 2 года назад
excellent explanation! thank you.
@user-yk2zu5px3u
@user-yk2zu5px3u Год назад
Thank you!!
@arshiarahimi3779
@arshiarahimi3779 4 года назад
Thank you! This was awesome
@JL-pg4pj
@JL-pg4pj Год назад
Best one, thanks a lot.
@fahnub
@fahnub 2 года назад
that was a great explanation. so helpful.
@mdbadiuzzamanbodi8250
@mdbadiuzzamanbodi8250 2 месяца назад
1:25 The video is supe helpful, just a tinay thought. I think calling analog to digital converter hardware would be appropriate .
@amirmasoud_iravani
@amirmasoud_iravani Год назад
best video on asr ever!😍
@nova4536
@nova4536 Год назад
Finally a much needed explanation.. As I've just started to get into a voice assistant from scratch in c++ lol 😂
@njerurichard3581
@njerurichard3581 Год назад
Good luck in your journey. I'm trying the same thing.
@tobiaszpietrala8643
@tobiaszpietrala8643 3 года назад
Thanks a lot guys!!!
@micaismyname
@micaismyname Год назад
Thank u so much for the video!! It was really helpfull!!
@aryanyekrangi7093
@aryanyekrangi7093 3 года назад
Great video, helped me out!
@samrithkohli2272
@samrithkohli2272 2 года назад
Superb explanation
@noemimp4910
@noemimp4910 8 месяцев назад
Wow!! Very good communicatiors, great video, really helpful
@laibaaaqil7678
@laibaaaqil7678 2 года назад
Exceptional Brilliant 👏 👏 👏 👏 👏 Thank you so much 💓 💗 💛 💖
@roshcoben8820
@roshcoben8820 4 года назад
Thanks guys! Next for me is to check how deep learning is used for speech recognition
@dina4913
@dina4913 Год назад
Amazing, deserves more likes/views
@huyangquang1711
@huyangquang1711 4 года назад
you should do more of this !
@roxy_2000
@roxy_2000 8 месяцев назад
very nice and interesting! the person pretending the model is so cute!
@williamgomez6226
@williamgomez6226 2 года назад
Excelente more than excellent!!!
@shubhamg9495
@shubhamg9495 Год назад
this video was so good
@hayatt143
@hayatt143 4 года назад
Great Video. Can you please give an example of sequential nature of speech and in what terms HMM is not flexible?
@matth8355
@matth8355 2 года назад
So good 👌🏻
@irfananizar5376
@irfananizar5376 Год назад
Good explanation .. Need more topics..plss
@YouTubist666
@YouTubist666 2 года назад
Nice job.
@maikhellmeister54
@maikhellmeister54 4 года назад
Very good! Thank you guys. Da you have a video about the GMM+HMM?
@jiayiwu4101
@jiayiwu4101 4 года назад
Thanks for your explanation! Very helpful! What does the input look like for NN? How does NN and HMM work together?
@chinmayakishore6219
@chinmayakishore6219 3 года назад
brilliant mate.cheers.can u make a video on tacotron ,concatenative..and all the major methods too.
@parth1450
@parth1450 4 года назад
Thanks people!!!
@sarthakpatwari7988
@sarthakpatwari7988 6 месяцев назад
Thank you for such a wonderful explaination of such a tricky concept can you also please provide the code for the same ?
@ritaheller921
@ritaheller921 3 года назад
I love you guys
@florain5040
@florain5040 4 года назад
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.
@muhibullah4201
@muhibullah4201 5 лет назад
good job
@renew44
@renew44 4 года назад
1:40 I see what you did there! ;)
@lotiumdupeuple3661
@lotiumdupeuple3661 5 лет назад
Great explanation thanks How is an hmm trained to define the good probabilities ?
@niwdenus8256
@niwdenus8256 4 года назад
Great stuff! Do you have any papers or resources to go from here?
@rabiadamlaozyer9622
@rabiadamlaozyer9622 Месяц назад
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
@rabiadamlaozyer9622 Месяц назад
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!
@saimadhavp
@saimadhavp 5 лет назад
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
@joshuanandrekar 4 года назад
Hey, did you get any info about it? Was looking for the same thing
@a2sirmotivationdoses782
@a2sirmotivationdoses782 3 года назад
Sir what is the Data set for Speech To Text Machine learning Model?
@rajeevanand5731
@rajeevanand5731 8 месяцев назад
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
@Hannesvl 8 месяцев назад
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!
@melinagensler9582
@melinagensler9582 4 года назад
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
@two697 3 года назад
Did you find any?
@gdevification
@gdevification 3 года назад
Can you share your workshop?
@ediwidodo499
@ediwidodo499 Год назад
how to calculate the probabilities used in real application ? is it stored in a database or how ?
@Hannesvl
@Hannesvl Год назад
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
@Democracy_Manifest Год назад
You could use a language model to calculate the probabilities.
@laenetmoloto9716
@laenetmoloto9716 2 года назад
06:43
@isvelalonso1376
@isvelalonso1376 7 месяцев назад
Hola,hablo espàñol
@arnavdas3139
@arnavdas3139 4 года назад
For anyone wanting to know about Fourier transform can refer to these videos : ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-spUNpyF58BY.html
@anubhavsingh2555
@anubhavsingh2555 Месяц назад
well done guys the way you explained this makes totaly worth to watch
Далее
Hidden Markov Model : Data Science Concepts
13:52
Просмотров 115 тыс.
Automatic Speech Recognition: Chapter 1
5:53
Просмотров 8 тыс.
How AIs, like ChatGPT, Learn
8:55
Просмотров 10 млн
Why Neural Networks can learn (almost) anything
10:30
Markov Decision Processes - Computerphile
17:42
Просмотров 163 тыс.
How Voice Recognition Works
5:00
Просмотров 134 тыс.
Hidden Markov Models 12: the Baum-Welch algorithm
27:02
How to train simple AIs to balance a double pendulum
24:59