Natural Language Processing: Crash Course Computer Science #36

Подписаться 16 млн

Просмотров 493 тыс.

50% 1

Today we’re going to talk about how computers understand speech and speak themselves. As computers play an increasing role in our daily lives there has been an growing demand for voice user interfaces, but speech is also terribly complicated. Vocabularies are diverse, sentence structures can often dictate the meaning of certain words, and computers also have to deal with accents, mispronunciations, and many common linguistic faux pas. The field of Natural Language Processing, or NLP, attempts to solve these problems, with a number of techniques we’ll discuss today. And even though our virtual assistants like Siri, Alexa, Google Home, Bixby, and Cortana have come a long way from the first speech processing and synthesis models, there is still much room for improvement.
Produced in collaboration with PBS Digital Studios: / pbsdigitalstudios
Want to know more about Carrie Anne?
about.me/carrieannephilbin
The Latest from PBS Digital Studios: • All PBS Digital Studio...
Want to find Crash Course elsewhere on the internet?
Facebook - / youtubecrash. .
Twitter - / thecrashcourse
Tumblr - / thecrashcourse
Support Crash Course on Patreon: / crashcourse
CC Kids: / crashcoursekids

Опубликовано:

21 ноя 2017

Ссылка:

Скачать:

Готовим ссылку...

Добавить в:

Мой плейлист

Посмотреть позже

Комментарии : 263

@alexanderkoller 6 лет назад

Hi, I'm a professor of computational linguistics in Germany. I just wanted to say that your video is one of the best explanations of natural language processing for non-experts that I've ever seen. I'm really impressed - great job!

@379dhairyaashara4 6 лет назад

I'm a computer science student and i really need some help as i'm gonna present on this topic and needed someone to guide me through. If you can help please reply.

@camille5098 5 лет назад

Please what is the difference between computational y and NLP

@uzairakram899 4 года назад

I agree, This needs to be the introductory course for CS majors starting their undergrad. I had to take an introductory course at my university and the instructor told us to write an essay on computer science personalities; Grace Hopper and the like. While these personalities are definitely important people and deserve the respect. Some assignments were about choosing a preferred path in CS without giving any background on these paths. I think that the course neglected to give any introduction to computer science.

@faizarahman935 4 года назад

@@379dhairyaashara4 Hi I am in the same boat. Did u get any information that was helpful?

@Chr0nalis 6 лет назад

This is how computer science should be taught. The motivation and big picture is important.

@uzairakram899 4 года назад

I think this should be the introductory course for CS majors.

@etiennepesce546 6 лет назад

"It wasn't evil, it was efficient" might be the most terrifiying sentence I've ever heard in my life.

@the80386 6 лет назад

exactly. killing off humans will also be 'efficient'. not sure I'd enjoy that though.

@tuele4302 6 лет назад

It was music to my ears. If it works, it's not [insert negative adjective here].

@zerge69 6 лет назад

That's because you do not understand what happened.

@JulianHarris 5 лет назад

Framed around explainability, I'd agree. The bots diverged to using a language that humans couldn't understand, making it opaque and therefore unethically unexplainable.

@dwijaparikh101 5 лет назад

INTJ

@justsomeperson1282 6 лет назад

Keep calm and Carrie Ann.

@shreyanshdas7481 6 лет назад

Noah Browne haha

@allanjoarder27 6 лет назад

This series is so fantastic. I get super excited to hear real-world examples and helps me envision the possibilities in the future.

@kaptenteo 6 лет назад

Will the foreshadowing of Crash Course Linguistics ever cease?!

@angeldude101 6 лет назад

One minor issue is that with speech interfaces becoming more common and more complex, some people will over estimate how human they are and get annoyed when it doesn't understand what the user thinks is a perfectly simple request.

@neutronstar6739 6 лет назад

angeldude101 exactly me lmao

@IceMetalPunk 6 лет назад

"SIRI, I TOLD YOU I WANT SOME BUFFALO WINGS, WHY ARE YOU SENDING ME PLANE TICKETS TO NEW YORK?!"

@ReplacedTea 6 лет назад

I can't believe I've watched all 36 of these. Keep up the good computer science.

@JesseMcCall 6 лет назад

Oh my gosh, this is the best channel I've subscribed to in a long time (on par with SmarterEveryDay)! Kudos!

@typograf62 6 лет назад

I'm 55. It has been "pretty soon" that voice recognition would work properly for most of my life. It is now in common use - and it still does not work. Perhaps if I live to a 110, but I will take no bets on that.

@scfoxcode 6 лет назад

Absolutely fantastic video. A great overview of the topic making it easy for me to learn more about the parts that interest me. I will be watching more of these this weekend

@safflower_s 6 лет назад

thank you for the episode! i knew a lot of it, but this video managed to structure everything so neatly that i feel like i understand the topic so much better now :>

@ivannaivanova8542 5 лет назад

I love how you gave an example with Siri! Thank you for all your work! You are amazing.

@mattkuhn6634 6 лет назад

Aw yiss, this is my jam! Thanks for this video - I’m in grad school working on an MS in Computational Linguistics, so NLP is what I do every day. I actually just finished writing a simple part of speech tagger the other day!

@alephquartz 6 лет назад

that seems like a lot fun. I also enjoy both cs and linguistics, so that seems like a fun career path.

@faizarahman935 4 года назад

I'm a computer science student and in need of some help for my last year's project which includes this topic and needed someone to guide me through. If you can help please reply.

@old-moose 6 лет назад

Elisa was the starting point of doctoral interest in natural language for education. We've come a long way from Xerox Sigma 7s and PDP-11s working through teletypes and VDTs. I think we are about ready to start doing the natural language teaching that I dreamed about all those decades ago.

@dielfonelletab8711 6 лет назад

I SPEAK HUMAN LANGUAGE TOO, FELLOW HUMAN.

@recklessroges 6 лет назад

I also pose no danger to my fellow humans, (I think we are getting away with it!)

@dielfonelletab8711 6 лет назад

WE POSE NO DANGER TO THE HUMAN POSE NO DANGER TO THE HUMAN POSE NO DANGER TO THE HUMAN POSE NO DANGER TO THE HUMAN POSE NO DANGER TO THE HUMAN POSE NO DANGER TO THE HUMAN POSE NO DANGER TO THEsegmentation fault

@jxffxry 6 лет назад

Reckless Roges WHY ARE YOU YELLING ARE YOU PERHAPS A ROBOT

@CuriousSomeone 6 лет назад

Brilliant! So clearly explained. Well done!

@dixie_rekd9601 Год назад

I remember watching this 5 years ago and slowly realising how important it would be for LLM AI and how the use of language is just another level of abstraction for the reality we live in.

@cristian0523 6 лет назад

This videos are just so awesome!!!, great job.

@TalysAlankil 6 лет назад

Yay, an episode on the topic of my thesis!

@TheDomius 6 лет назад

Your enthusiasm is awesome.

@qlifee 6 лет назад

Thank you, I always wanted to know more about this topic 👍

@kd1s 6 лет назад

I recall on my TRS-80 Model 1, Level 2 with EI and disk - I had a speech synthesizer hooked up and a speech recognition module. That was great fun - learned about phonemes through that.

@brianandbrain 6 лет назад

Awesome breakdown of the current state of NLP.

@bharathdev6285 6 лет назад

big fan of Crash course. great job guys.

@natbud7987 Год назад

It would be great to have a similar series on implications of technical progress with philosophers and social studies experts commenting on the relationship between the tech and our society.

@muzamilshah8028 6 лет назад

thats really informative ...thank you sooo much ..keep it up ...

@SimonGornick 4 года назад

Great introduction to a very complex area!

@NishantPrateek 6 лет назад

Quite comprehensive!

@cocoachanie 4 года назад

Excellent video!

@unimportantaccount1 6 лет назад

Try turning on automatic captions. For clearly spoken stuff like what is on this channel, they are super accurate now.

@DuluthTW 6 лет назад

Very interesting. Another outstanding lesson. Thanks!

@shreyanshdas7481 6 лет назад

Teresa White Very true

@SinerAthin 6 лет назад

One thing to keep in mind with regard to natural language is that natural language is often highly imprecise, and very dependant upon the level of eloquence of the transmitter and the knowledge and understanding of the receiver, and the importance of shared points of reference for both of them. This is why for instance when it comes to the sciences there is such a heavy emphasis on exact word use. When writing a scientific article or instruction manual, you want to be as precise as possible with your use of words, terms and grammer and to minimize the chance of missunderanding to be as small as possible. That is why, even if computers could understand natural language, the degrees of complexity of the tasks which they could pull off would inevitably be affected by the accuracy of the instructions(macros and preprogrammed sequences and programs are of course exempted).

@maxineyang1332 6 лет назад

She speaks faster than my internet connection

@nceevij 5 лет назад

haha

@susuilu 5 лет назад

lol

@varghesevg5 5 лет назад

I thought my speed settings were not normal so had to check to reconfirm .. on the positive side!

@uzairakram899 4 года назад

I really like the pace its a crash course

@TRMcCoy 5 лет назад

Fascinating, thank you.

@darthsalsapants7059 6 лет назад

How about a crash course electrical engineering?

@darthsalsapants7059 6 лет назад

thanks

@thatoyaonebogopa9483 4 года назад

Your classes are amazing...

@thomasedwardking7286 Год назад

Came in to watch the video... Staying to watch the series 🎉

@patrikkarlsson9523 6 лет назад

great stuff!!

@WIImotionmasher 6 лет назад

Hey I've beenn taking an NLP class this term. Its hard. *edit* This video was not really about what my class was about if you're wondering how its hard. We covered types & tokens, Part of Speech tagging, word senses, Viterbi algorithm, Naive Bayes, Lesk Algorithm, Co-occurence matrices, and briefly near the end Skip-gram and CBOW embeddings systems. Among a number of other things that I've completely forgotten about. This class was difficult and I also screwed up, missing 3 classes being sick and not focusing enough time to do assignments. The entire course has been a nightmare and I know nothinng from it. Last class is Decemmber 7th, I will fail.

@rparl 6 лет назад

Koala: eats shoots and leaves. Eats, shoots, and leaves.

@kylenoe2234 6 лет назад

Omg like x1000. Great coverage girl...

@yesterdaysrose5446 6 лет назад

I recently saw that clip from the Bell Labs speech generation system in the movie "Gizmo!" from 1977. HIGHLY recommended for fans of obscure technology that didn't quite make it, and random feats of badassery. And great music. Let's not forget the great music.

@Grimlock1979 6 лет назад

Wait for it... The Mongols!

@HerodotusVon 6 лет назад

Dennis E They're the exception!

@bttfish Год назад

Now the ChatGPT is dominating this field.

@blakeingram1711 5 лет назад

The best part of this video is when she spoke in Shakespearean to Siri. Who here can also fluently speak this language? I know I can.

@liminal18 6 лет назад

This is awesome you rock comp sci goddess.

@Magpie_Artistry 5 лет назад

I wonder if there are some academic papers that introducing these concepts? I hope wish I can find good citation for my dissertation, thank you!

@tombouie 6 лет назад

Very impressive young lady, Thks

@appletree6741 5 лет назад

Brilliant

@davidchidester5463 5 лет назад

This makes me want a crash course linguistics so bad.

@paichethan 6 лет назад

A video on 1) firmware, drivers, microcontroller, daq and PLD 2) database Would be very useful and complete...

@tueemsyhu4846 6 лет назад

Good and Interesting!😀😁

@HMALDANA 6 лет назад

Thank you, Carrie. Great explanation!

@jeronimotamayolopera4834 6 лет назад

GREAT.

@MasterGeekMX 6 лет назад

There is even Vocaloid, a speech synthesizer software by Yamaha aimed to singing. There is a whole culture aroind it because the companies that made the voice banks have created anime-style characters for them.

@agentanakin9889 5 лет назад

I'm amazed by the Mac in the background.

@GreyCartridge 6 лет назад

How fitting. The ad for this video was for the Echo.

@microbuilder 6 лет назад

I remember downloading a chat bot called Daisy back in the day...very simple, but if you spoke to it in a consistent manor, and did it a LOT, it would spit out some fairly interesting things at times...I think still have some of the conversations saved somewhere.

@saeedbaig4249 6 лет назад

What were some interesting things it said?

@microbuilder 6 лет назад

I dont quite remember anymore, this wouldve been in the early 2000s...I'll see if I can find the conversations.

@Ayplus 6 лет назад

R u sure that wasnt an actual person. . .?

@microbuilder 6 лет назад

lol yep Still looking for the old conversations...not being able to find them makes me want to find them even more...

@ofrench560 6 лет назад

did you ever find those

@rubenalejandro7763 5 лет назад

Damn, Computer Science is so exciting.

@johnmongver 4 года назад

I am a linguist but I know nothing about computer science and NLP but I am interested to get a job in this field. Do you have any recommendations on how can I start my learning? What should I study to become more attractive towards employers?

@cashphattichaddi 6 лет назад

Dope!

@ramkrishnaverma4194 6 лет назад

apart from the vocabularies concept can you give short description on Entity extraction, aspect of entity extraction , like how can these interrelate to Natural language processing.

@UnpopularName 6 лет назад

WHAT THINK YE ON THIS GENTLE MIDSUMMER'S DAY? >intense concentration< maybe my favorite part in the entire series so far

@DanielGolding1337 6 лет назад

Crash Course - English Grammar!@! Next series please.

@jounalansman1769 4 года назад

'Blabbing on about'. lol I like kerryan choice pf words. :)

@eve36368 5 лет назад

I mean all I got was the surveillance. Then again, as a tool this is so cool.

@onee 5 лет назад

What bothers me about those "speech recognizers" is. Most of the time we humans don't say things like "I am going to ...". But talk more like this "Am goingto" (yes, I exaggerated a bit). So, we combine two words, because you talk faster. Or we even skip words. But the computer is programmed to recognize each word individually. Which isn't the way humans naturally speak.

@srijithbalachander6213 5 лет назад

Probably the best channel/playlist that I have encountered on youtube till date apart from 3blue1brown. Carrie Ann, Carry on!

@mikejohnstonbob935 6 лет назад

Buffalo buffalo Buffalo buffalo buffalo buffalo Buffalo buffalo

@Pfhorrest 6 лет назад

Police police police police police police, who in turn police police.

@IceMetalPunk 6 лет назад

Crash blossoms.

@xXsolar99Xx 6 лет назад

My favorite is: "Wouldn't the sentence 'I want to put a hyphen between the words Fish and And and And and Chips in my Fish-And-Chips sign' have been clearer if quotation marks had been placed before Fish, and between Fish and and, and and and And, and And and and, and and and And, and And and and, and and and Chips, as well as after Chips?"

@avagoyamug 5 лет назад

Hey, that's a copy of Mitnick's "Ghost in the Wires" in the background!

@mohamadeen 6 лет назад

My TA is preparing PhD on NLP things which brought me to this channel an' boy I'm glad i found it

@alejandromendoza2416 5 лет назад

thanks for this information I will take a 10 in my class of english

@MusiCaninesTheMusicalDogs 6 лет назад

I actually work with mainframes and I daily program in a language called Natural! Does anyone else work with it around here?

@KingsleyIII 6 лет назад

No mention of Watson (the Jeopardy!-playing computer that beat Ken Jennings and Brad Rutter) in a video about computers that understand human languages? What the heck?

@greyareaRK1 6 лет назад

Methinks thou art more wonderful than a summer's day. Long ago I created a 3D CGI puppet, using phonemes as the anchor points in speech as they correspond to the shapes our mouths form when speaking. Being able to detect phonemes in speech would have been a massive timesaver. I supposed that's what is going on with animojies?

@IceMetalPunk 6 лет назад

Nope. Animojis use the phone's 3D tracking camera to map points on your face onto the 3D model. So it's just copying your 3D expression, not basing it on sound at all.

@SuperDropsX 6 лет назад

Little correction: Phonems are from the realm of Phonology, not Phonetics. The corresponding unit from Phonetics is a Phon. The distinction is important, because Phons are a physics based unit while Phonems are a mind based unit. One Phonem can contain many multiple Phons, which are then referred to as Allophones.

@brianhack5806 6 лет назад

Perhaps it's an acceptable variation, but I would have typed "phoneme", "phonemes", and "phones".

@SuperDropsX 6 лет назад

Ah, good catch. You are fully correct, the mistake stems from my being a native german speaker. Thank you for your correction.

@brianhack5806 6 лет назад

Bitte shoen. "Akkusativ", "Dativ", und "Nominativ" gefallt mir. The extra "e" is rather pesky to me but it is a bit amusing to spell my native language since there is almost a method to the madness.

@rkpetry 6 лет назад

...and, finally, speaking of accuracy, linguists tell us that linguistics, the scientific study of language-in-general, is not an exact science, and yet, linguistics is what we use to convey the understanding of exact science-(is that a paradox)-and furthermore will correct the usage of natural language only when we stop monkeying-around for example, NOT-ELSE.

@cholten99 6 лет назад

And yet I've still never seen a web-based version of Eliza or any other Turing test contender that works using HTML speech recognition and generation. Anyone know of any?

@ayoubessabiri6331 4 года назад

Dear Siri, you are not alone. I am not sure I got that too :)

@catsaur 6 лет назад

I had to make a chatbot for my cs class, so this topic is good timing.

@TheOlian04 6 лет назад

catsaur if you didn't know about neutral networks before this video I doubt you will need to implement one to pass your cs class. A set of if statements would probably suffice :)

@IceMetalPunk 6 лет назад

One of the simpler ways to make a "dumb" chatbot is to give it a small set of default "conversation starters" to randomly choose from. Whenever a user says something it doesn't know about, it responds "I didn't understand that", but adds what the user said to its conversation starters. Whenever it chooses a conversation starter, it logs what the user responds as an "appropriate response" and now it knows how to respond to one more possible input :) Of course, even for a "dumb bot", you may still want to do some input normalization to get a more generic set of matched inputs for responses, but that's up to you :)

@ProfessorBorax 6 лет назад

Esperanto would make this all so much easier :)

@comittailfern5852 6 лет назад

Hmm, I never looked into the other parts of this channel, mostly just history and astromony. I must say I'm actually glad i see this, reasons being my desire to be an astro phycsist to have nasa send me to mars, and away from my crazy ex.

@ulysisxtr 6 лет назад

WE NEED CRASHCOURSE ENGLISH LANGUAGE!!!!!! ASAP!!!!!!

@leocelente 6 лет назад

Daisy, daisy... give me your answer do

@ntanileng 5 лет назад

what software do you use for editing?

@starmole5000 4 года назад

Great vid, slow down a tiny bit tho :)

@hozaifaowaisi1250 6 лет назад

Finnaly

@jawadtahmeed9854 5 лет назад

Damn, she speaks faaaaaast !!

@fourtien8656 6 лет назад

In the video where u talk about black holes u said that time would speed up for someone falling in to a black holes but would that mean u could see the black hole that u are falling in to died like black hole supernova and if i was at the a black hole and i was one meter away from the evethorisen could i see the future and get back to Earth i would be dead but could i??

@TheJaredtheJaredlong 6 лет назад

Wrong course. You'd be better off asking this on SpaceTime. To attempt to answer your question: as far as we know "information" is never destroyed. This is a consequence of a law of the universe that matter can never be destroyed. So all the matter, including energy and light, that goes into a blackhole just stays there. So if you were to enter a blackhole (without dieing) you would have access to all of that information meaning you could see the past; like a cosmic archive or history book. But, because of the crazy amount of gravity that's happening to cause this affect, time would speed up for you almost infinitely due to the extreme spacetime warp while at the same time someone watching you from outside the blackhole it would look like time had infinitely slowed down for you. You would quickly see the death of the blackhole due to evaporation, but because you're seeing all the information being released, from your perspective it would look like you're traveling back in time back towards the creation of the blackhole. What you would see after that point is a complete mystery. Science has never observed the complete evaporation of a blackhole.

@ashokkumarg6277 4 года назад

I am PG Students from India Linguistics ,but I studied General Linguistics in master degree, can anybody guide me self learning Computational Linguistics course available???? Pls.

@Eban11235 6 лет назад

If I do a voice search I do it the same way I do a text search. I don't bother with sentences.

@piggy8435 6 лет назад

is there a video on Databases?

@dewiz9596 6 лет назад

Fræ: quick answer, yes. This is number 36 in this series.

@sheepphic 6 лет назад

Does anyone else find it annoying that to take advantage of these awesome technologies you have to send your voice data off to some server somewhere? I wish that there were local speech recognition solutions (that actually compiled (looking at you, AUR version of Simon))

@IceMetalPunk 6 лет назад

Those wouldn't be very accurate; but more importantly, if it was all local, the global systems couldn't learn from your interactions, and overall technological progress of them would slow down. Which would suck.