Тёмный

14. Learning: Sparse Spaces, Phonology 

MIT OpenCourseWare
Подписаться 5 млн
Просмотров 80 тыс.
50% 1

MIT 6.034 Artificial Intelligence, Fall 2010
View the complete course: ocw.mit.edu/6-0...
Instructor: Patrick Winston
Why do "cats" and "dogs" end with different plural sounds, and how do we learn this? We can represent this problem in terms of distinctive features, and then generalize. We end this lecture with a brief discussion of how to approach AI problems.
License: Creative Commons BY-NC-SA
More information at ocw.mit.edu/terms
More courses at ocw.mit.edu

Опубликовано:

 

21 авг 2024

Поделиться:

Ссылка:

Скачать:

Готовим ссылку...

Добавить в:

Мой плейлист
Посмотреть позже
Комментарии : 37   
@xoARTxo1
@xoARTxo1 6 лет назад
This is almost better than Netflix bingewatching
@razzlfraz
@razzlfraz 5 лет назад
I'm not the only one! Yay!
@kyoungd
@kyoungd 10 лет назад
I was a bit lost in this one. I had trouble with slabic, voiced and few other phonetic terminology. Still, professor gave me a good overview of phonology. It is amazing that we use so few phonetics. Also, the last part about how to approach AI problems is valuable in any field.
@natfailsyoutube8163
@natfailsyoutube8163 9 лет назад
Not sure if this might help, but I think "syllabic" was what he meant when he wrote/said "slabic" in the lecture.
@kyoungd
@kyoungd 9 лет назад
nat4200 That does help. Thanks. :-)
@acidtears
@acidtears 3 года назад
44:01 - Mechanism envy in Artificial Intelligence research: trying to apply neural nets to every, possibly unsuitable problem.
@WepixGames
@WepixGames 4 года назад
R.I.P Patrick Winston
@FIBONACCIRICH-ql2pw
@FIBONACCIRICH-ql2pw 9 месяцев назад
:(( R.I.P.
@francescos7361
@francescos7361 Год назад
Grazie mille , interessante
@EranM
@EranM 6 лет назад
1:37 these ideas about neural nets.. Comon this is so not updated.. Why didnt you publish his recent courses from 2015 like classes 12a and 12b about neural networks..
@videotrash
@videotrash 8 лет назад
I wish it was true, that German cows would "müh", not moo, but I'm afraid it isn't so
@sprinkdesign7170
@sprinkdesign7170 8 лет назад
I think you meant to say "I wish it was true, that German cows would "müh", not moo, but I'm afraid it isn't soue."
@ekwmhroucg
@ekwmhroucg 4 года назад
Die Kuh macht "muh"!
@studywithjosh5109
@studywithjosh5109 3 года назад
great lecture Winston
@gumikebbap
@gumikebbap 7 лет назад
When Mr. Winston says that neural networks weren't a good approach to this problem, I guess it's because he actually tried them, and I'm very intrigued about the result. Or did he just know it wasn't a good method? In that case, how did he come to that conclusion?
@hailapathy2801
@hailapathy2801 7 лет назад
There have been a number of significant advances in neural networks since this was recorded. In fact, some of these advances prompted him to re-record the neural net portion of the course.
@AICodingAdventures
@AICodingAdventures 8 лет назад
It's very interesting how he said that the Neural Networks were history, maybe he meant that they're making history. Anyway, great lectures!
@sprinkdesign7170
@sprinkdesign7170 8 лет назад
Watch the lectures on Neural networks - videos 12a and 12b. You're right and I think Winston would also agree with your comment.
@NicolayGiraldo
@NicolayGiraldo 5 лет назад
'bi:tʃ I find the lack of IPA transcriptions a bit lacking. It's not "funny" letters, they help with many languages. Extremely good material though.
@ahgiynq
@ahgiynq 4 года назад
Thank you
@chrisoman87
@chrisoman87 10 лет назад
I noticed in this lecture he was talking about pronouncing a 'z' sound at the end of dogs, I think this a feature of 'American' English pronunciation, I wonder how an algorithm constructed around these arbitrary pronunciation quirks would deal with 'Australian' or 'British' English as they pronounce the 's' sounds in both cats and dogs.
@0MVR_0
@0MVR_0 7 лет назад
It is the same. The linguistic phenomenon this is referring to is called phonological assimilation. en.wikipedia.org/wiki/Assimilation_(phonology) British English experiences the same cross-phonemic bleeding, hence Brits pronounce /dogz/ with the voice plural morpheme.
@yuumayamamoto7970
@yuumayamamoto7970 Год назад
is this when genetic algorithms and neural nets only produced poor performance results?
@yuumayamamoto7970
@yuumayamamoto7970 Год назад
What does glimpse through a soda straw mean?
@EranM
@EranM 6 лет назад
45:30 Doesn't it?
@irinthirdwater6478
@irinthirdwater6478 6 лет назад
Does anyone have a resource to David Marr's catechism? I want to read more about it but couldn't find it anywhere.
@razzlfraz
@razzlfraz 5 лет назад
Page 279 of the Artificial Intelligence book ( www.amazon.com/exec/obidos/ASIN/0201533774/ref=nosim/mitopencourse-20 ).
@katateo328
@katateo328 2 года назад
yeah, localness is the most important.
@MikhailTrunovich
@MikhailTrunovich 4 года назад
41:24 Winston shows a diagram with all phonemes. Does anybody know if he published this somewhere?
@pmcate2
@pmcate2 3 года назад
He didn’t come up with those. You can simply look them up on Wikipedia
@ECOMMUSK
@ECOMMUSK 9 лет назад
the first lecture was better
@sirhendro
@sirhendro 8 лет назад
I miss some enthusiasm in his speech. This makes it less enjoyable to watch.
@sprinkdesign7170
@sprinkdesign7170 8 лет назад
I like his straightforward delivery. Makes me trust him more. I mean, more than I normally would trust some guy who happens to be a tenured member of the MIT faculty.,...,
@mathematics6199
@mathematics6199 3 года назад
who came after gaurav sen's video?
Далее
15. Learning: Near Misses, Felicity Conditions
46:54
Просмотров 69 тыс.
How to Speak
1:03:43
Просмотров 19 млн
1. Introduction and Scope
47:19
Просмотров 1,7 млн
13. Learning: Genetic Algorithms
47:16
Просмотров 520 тыс.
6. Search: Games, Minimax, and Alpha-Beta
48:17
Просмотров 450 тыс.
26. Chernobyl - How It Happened
54:24
Просмотров 2,8 млн
MIT Introduction to Deep Learning | 6.S191
1:09:58
Просмотров 490 тыс.
Richard Feynman: Can Machines Think?
18:27
Просмотров 1,5 млн
Calculus at a Fifth Grade Level
19:06
Просмотров 8 млн