At 42:10 the MRI scan is goosebumps. Fascinating Stuff Madam. The amount of thought that goes into an ASR system is mind boggling. It was a great communication of what you guys are working on. thank you.
Very good explanation :) It is very helpful to understand easily for this kind of topics, instead of going through vast research papers. Anyways thank you.
Isn't not only about using words as basic unit. But about comparison of input data and the stored one asa slight difference between background noise, speed,voices, amplitudes will cause problems in comparison
Sister you look like our state people (Tamil) I don't know exactly but I think u r from Tamil Nadu.. Anyway nice explanation... Preethi akka(sister)...
Dear Microsoft Research, At 1:09:00 you are at the section "What's next?" and talk about several problems. One of those problems is handling noisy real-life settings with many speakers (e.g., meetings, parties). Is it possible to implement the coktail party algorithm into your system? It may be possible to solve the problem partially. Yours sincerely, D.
Dynamic Belgian Networks ? I guess from context its "Bayesian". I can understand English, but I had to watch this with subtitles on, to test how good CC was.
thanks very much for this video. i wish to ask, which university is that? i wish to take part in your lectures. Multilingual speech recognition is my topic of research
how i can change live videos voice , like English voice to Hindi , how i can change your any video English voice to Hindi voice , i am not able to read caption
Funny when she says 'clearly speech recognition is not difficult for humans'. It takes at least 1-2 years for humans gain this skill and a decade to be fully developed, if you include speech in noise recognition. It is far from easy.
Luiz Felipe 😂😂 Btw, I want to ask a question, what one has to do if someone is interested in it ? I mean what educational background he/she has to have ?
I am currently learning about ASRs and I am trying to find good resources on modern ASRs so that I can make my own ASR system. This is largely to learn and have a new project on my resume. Do you have any recommendation for resources more relevant than this video?
While the talk is interesting, I am so sick of hearing "X is sexist" or any other "-ist" or "-ism". Can we just say that the application is too narrowly focused? For heaven's sake, it was a toy dog with a magnet, not a woman or elderly person hating machine. I
At 42:10 the MRI scan is goosebumps. Fascinating Stuff Madam. The amount of thought that goes into an ASR system is mind boggling.
It was a great communication of what you guys are working on. thank you.
Ldx
Lxllx
watching the video almost 2 years later with real-time translation and you can see improvements and still some misses, very nice one
A very intelligent researcher. I am passionate about ASR
Very good introduction and overview to ASR. Thanks.
nst as gx or not
Prof. Jyothi is presenting at Microsoft on an Apple laptop.
This was so insightful and clearly presented! Thank you so much!!! :)
Excellent talk about ASR by Preethi. Thank you very much
Very good explanation :) It is very helpful to understand easily for this kind of topics, instead of going through vast research papers. Anyways thank you.
Isn't not only about using words as basic unit.
But about comparison of input data and the stored one asa slight difference between background noise, speed,voices, amplitudes will cause problems in comparison
Got you! Prof Alan Black from CMU is among the audience!
noice
Impressive Presentation Skills
Very clearly overview for ASR, pretty good!
Wonderful explanation of ASR. Please share the PPT for this presentation with me.
Vice nice introduction about ASR
Very neat and concise overview on ASR. Thanks a lot!
Sister you look like our state people (Tamil) I don't know exactly but I think u r from Tamil Nadu..
Anyway nice explanation... Preethi akka(sister)...
Dear Microsoft Research,
At 1:09:00 you are at the section "What's next?" and talk about several problems. One of those problems is handling noisy real-life settings with many speakers (e.g., meetings, parties). Is it possible to implement the coktail party algorithm into your system? It may be possible to solve the problem partially.
Yours sincerely,
D.
Excellent overview of ASR. Thanks you.
Hello Ma'am Can you suggest some material or source related to speech recogniton.
I want to work on this in my M.Tech thesis.
Really nice introduction 👍
Absolutely amazing and very insightful session.
Thank you so much. ..Such an excellent presentation!
Dynamic Belgian Networks ? I guess from context its "Bayesian". I can understand English, but I had to watch this with subtitles on, to test how good CC was.
Thank u, and Thanks to the ASR of RUclips that i can understand clearly what u said as i m not english speaker.
Not all mistakes were caught. Palin for example said "that entrepreneurial spirit", not "their entrepreneurial spirit."
Very helpful introduction. Thanks!
That's a lot of talks, beautifully explained!
Good introduction of ASR system.
Great lecture! Thanks a lot! =)
pdf of this lecture?
good lecture. But expected something more on End-To-End System with RNN
ya same
this comment didn't age well, Attention is all you need to change the paradigm to more transformer-based models.
heyy....can I get the links of all videos use in this seminar
thanks very much for this video. i wish to ask, which university is that? i wish to take part in your lectures. Multilingual speech recognition is my topic of research
Indian institute of technology Bombay best college in india
Very good explanation. Thanks a lot
how i can change live videos voice , like English voice to Hindi , how i can change your any video English voice to Hindi voice , i am not able to read caption
16:20, subtitles have adapted to accent now
Excellent presentation, Stay Blessed.
1:03:19
Graves & Jaitley, "Towards End-to-End Speech Recognition with Recurrent
Neural Networks" ICML 2014
--:--:--
Maas et al, "Lexicon Free Conversional Speech Recognition with Neural Networks", NAACL 2015
1:09:36
Chan et al., "Listen, Attend and Spell: A NN for large Vocabulary Conversional Speech Recognition", ICASSP 2016
Wow. Well presented
Muito esclarecedor , obrigado!
Great clear presentation
Where can i get the graph at 1:13:10 title Languages with ASR
Can I have the slide used in this video, please...?
www.cse.iitb.ac.in/~pjyothi/cs753/slides/lecture1.pdf
@@maddai1764 thanks...
needed this for my research sub.
Funny when she says 'clearly speech recognition is not difficult for humans'. It takes at least 1-2 years for humans gain this skill and a decade to be fully developed, if you include speech in noise recognition. It is far from easy.
Your awesome, please post more lectures
The recordings aren't available, but you can get all other material here: www.cse.iitb.ac.in/~pjyothi/cs753/
Very nicely explained!
Thank you very much my sir
please my project for Speech Samples Recognition
and i need heelp please. Can i get on your emial please
Shall I get the ppt for reference?
Very nice lecture. Thank you.
Can I get the name of the speaker ?
Well explained.
this should be subtitled for sure :-D
Nice session....
IITB my favourite 💟💟💟
It’s Point Presentation.
27:42 subtitles are wrong and no way to fix them. I dislike who is not open for contribution.
51:04, lol, youtube CC has the joke in its model. let us pray vs lettuce spray.
Luiz Felipe 😂😂
Btw, I want to ask a question, what one has to do if someone is interested in it ? I mean what educational background he/she has to have ?
@@Aditya-te7oo my background is computing science. I guess that and specialization in statistics/data analysis how help.
@@monad_tcp
Okay, thanks for that.
Btw, it's 5:15 in the morning and I didn't sleep the whole night. 😂😅
so...at microsoft research they use macbook 🤔🤗
Slides link please
www.cse.iitb.ac.in/~pjyothi/cs753/slides/lecture1.pdf
why does she mispronounce phoneme for the first half hour
When was this talk recorded? It's way too much outdated and definitely not fashion even back in 2016...
I am currently learning about ASRs and I am trying to find good resources on modern ASRs so that I can make my own ASR system. This is largely to learn and have a new project on my resume. Do you have any recommendation for resources more relevant than this video?
Liked it.
Am I missing something - the moderator is using an APPLE computer - Isn't this a Microsoft event? Bill Gates must be going crazy.!!
This is class of iit Bombay
Cool laptop
18:02
Vocabulary ✌🏿
Punjabi so excellent.
14:18
31:05
I like you preethi!!!!
Speech white ❤️
She's wearing her formal pajamas.
I was attracted by this women
Texture Infact co so many likely presentation.
F(0^=) Prime index Equal Point Proof 0
While the talk is interesting, I am so sick of hearing "X is sexist" or any other "-ist" or "-ism". Can we just say that the application is too narrowly focused?
For heaven's sake, it was a toy dog with a magnet, not a woman or elderly person hating machine.
I