March 2010 April 2010 May 2010 June 2010 July 2010
August 2010
September 2010 October 2010
November 2010
December 2010
January 2011
February 2011 March 2011 April 2011 May 2011 June 2011 July 2011 August 2011 September 2011 October 2011 November 2011 December 2011 January 2012 February 2012 March 2012 April 2012 May 2012 June 2012 July 2012 August 2012 September 2012 October 2012 November 2012 December 2012 January 2013 February 2013 March 2013 April 2013 May 2013 June 2013 July 2013 August 2013 September 2013 October 2013 November 2013 December 2013 January 2014 February 2014 March 2014 April 2014 May 2014 June 2014 July 2014 August 2014 September 2014 October 2014 November 2014 December 2014 January 2015 February 2015 March 2015 April 2015 May 2015 June 2015 July 2015 August 2015 September 2015 October 2015 November 2015 December 2015 January 2016 February 2016 March 2016 April 2016 May 2016 June 2016 July 2016 August 2016 September 2016 October 2016 November 2016 December 2016 January 2017 February 2017 March 2017 April 2017 May 2017 June 2017 July 2017 August 2017 September 2017 October 2017 November 2017 December 2017 January 2018 February 2018 March 2018 April 2018 May 2018 June 2018 July 2018 August 2018 September 2018 October 2018 November 2018 December 2018 January 2019 February 2019 March 2019 April 2019 May 2019 June 2019 July 2019 August 2019 September 2019 October 2019 November 2019 December 2019 January 2020 February 2020 March 2020 April 2020 May 2020 June 2020 July 2020 August 2020 September 2020 October 2020 November 2020 December 2020 January 2021 February 2021 March 2021 April 2021 May 2021 June 2021 July 2021 August 2021 September 2021 October 2021 November 2021 December 2021 January 2022 February 2022 March 2022 April 2022 May 2022 June 2022 July 2022 August 2022 September 2022 October 2022 November 2022 December 2022 January 2023 February 2023 March 2023 April 2023 May 2023 June 2023 July 2023 August 2023 September 2023 October 2023 November 2023 December 2023 January 2024 February 2024 March 2024
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29
30
31
News Every Day |

Meta’s new learning algorithm can teach AI to multi-task

If you can recognize a dog by sight, then you can probably recognize a dog when it is described to you in words. Not so for today’s artificial intelligence. Deep neural networks have become very good at identifying objects in photos and conversing in natural language, but not at the same time: there are AI models that excel at one or the other, but not both. 

Part of the problem is that these models learn different skills using different techniques. This is a major obstacle for the development of more general-purpose AI, machines that can multi-task and adapt. It also means that advances in deep learning for one skill often do not transfer to others.

A team at Meta AI (previously Facebook AI Research) wants to change that. The researchers have developed a single algorithm that can be used to train a neural network to recognize images, text, or speech. The algorithm, called Data2vec, not only unifies the learning process but performs at least as well as existing techniques in all three skills. “We hope it will change the way people think about doing this type of work,” says Michael Auli, a researcher at Meta AI.

The research builds on an approach known as self-supervised learning, in which neural networks learn to spot patterns in data sets by themselves, without being guided by labeled examples. This is how large language models like GPT-3 learn from vast bodies of unlabeled text scraped from the internet, and it has driven many of the recent advances in deep learning.

Auli and his colleagues at Meta AI had been working on self-supervised learning for speech recognition. But when they looked at what other researchers were doing with self-supervised learning for images and text, they realized that they were all using different techniques to chase the same goals.

Data2vec uses two neural networks, a student and a teacher. First, the teacher network is trained on images, text, or speech in the usual way, learning an internal representation of this data that allows it to predict what it is seeing when shown new examples. When it is shown a photo of a dog, it recognizes it as a dog.

The twist is that the student network is then trained to predict the internal representations of the teacher. In other words, it is trained not to guess that it is looking at a photo of a dog when shown a dog, but to guess what the teacher sees when shown that image.

Because the student does not try to guess the actual image or sentence but, rather, the teacher’s representation of that image or sentence, the algorithm does not need to be tailored to a particular type of input.

Data2vec is part of a big trend in AI toward models that can learn to understand the world in more than one way. “It’s a clever idea,” says Ani Kembhavi at the Allen Institute for AI in Seattle, who works on vision and language. “It’s a promising advance when it comes to generalized systems for learning.”

An important caveat is that although the same learning algorithm can be used for different skills, it can only learn one skill at a time. Once it has learned to recognize images, it must start from scratch to learn to recognize speech. Giving an AI multiple skills at once is hard, but that’s something the Meta AI team wants to look at next.  

The researchers were surprised to find that their approach actually performed better than existing techniques at recognizing images and speech, and performed as well as leading language models on text understanding.

Mark Zuckerberg is already dreaming up potential metaverse applications. “This will all eventually get built into AR glasses with an AI assistant,” he posted to Facebook today. “It could help you cook dinner, noticing if you miss an ingredient, prompting you to turn down the heat, or more complex tasks.”

For Auli, the main takeaway is that researchers should step out of their silos. “Hey, you don’t need to focus on one thing,” he says. “If you have a good idea, it might actually help across the board.”

Происшествия

На Урале будут судить виновника ДТП с микроавтобусом, в котором ехала группа учителей

Top 5 Websites to Watch FREE Movies - TV Shows (No Sign up!)

The 10 Intense New Action Movies on Netflix That Left Me on the Edge of My Seat!

I was diagnosed with cancer aged 39… you are never too rich, too famous or too young, says Dr Philippa Kaye

Top 10 Emmanuelle Seigner Movies

Ria.city






Read also

EXCLUSIVE: Congress raids presidential campaign fund in surprise reversal

'That's not who I am': Igor Severino apologizes for bite, hopes one mistake doesn't define career

BBC schedules in Easter shake-up as big shows cancelled or moved to make way for blockbuster movie

News, articles, comments, with a minute-by-minute update, now on Today24.pro

News Every Day

The 10 Intense New Action Movies on Netflix That Left Me on the Edge of My Seat!

Today24.pro — latest news 24/7. You can add your news instantly now — here


News Every Day

I was diagnosed with cancer aged 39… you are never too rich, too famous or too young, says Dr Philippa Kaye



Sports today


Новости тенниса
Янник Синнер

Янник Синнер поделился впечатлениями от общения с игроками сборной Италии по футболу



Спорт в России и мире
Москва

«Радио Зенит» – информационный партнер форума «Мы вместе. Спорт»



All sports news today





Sports in Russia today

Москва

Федерация бокса России на ВДНХ


Новости России

Game News

Шапки женские вязаные на Wildberries, 2024 — новый цвет от 392 руб. (модель 466)


Russian.city


Москва

В Вешкайме загорелся склад с готовой продукцией местной птицефабрики


Губернаторы России
#123ru.net

Архитектурные доминанты округов Москвы


Нелли Закирова, вдова музыканта Георгия Гараняна: «Он разок-другой дунул в саксофон и моментально понял, что это — его инструмент»

Пассажир не успел войти в поезд и разбил стекло вагона на станции «Киевская»

Шапки женские вязаные на Wildberries, 2024 — новый цвет от 392 руб. (модель 466)

Заказать недорогой ремонт кухонной мебели в районе в Москве и Московской области


Как Розенбаум судился с мусорщиками, Шнур прогорел на «Пушкине», а Нагиев заработал 4 тысячи рублей за год

Атомная взятка потребовала ареста // Топ-менеджер «Росатома» Геннадий Сахаров отправлен в СИЗО на два месяца

Певица Mia Boyka призналась, что в детстве ходила в церковь почти каждый день

Ольга Бузова стала человеком-невидимкой на ТВ-3 в новом сезоне легендарного шоу о звёздах


Битые корты: Медведев и Александрова вышли в четвертьфинал Miami Open

Хачанов победил Черундоло и пробился в 1/8 финала турнира ATP в Майами

Хачанов: хочу вернуться в топ-10 ATP

Александрова обыграла первую ракетку мира на турнире WTA



Шапки женские вязаные на Wildberries, 2024 — новый цвет от 392 руб. (модель 466)

Пассажир рейса Москва — Пермь попал в реанимацию

«Радио Зенит» – информационный партнер форума «Мы вместе. Спорт»

Перенос дат II этапа культурного проекта «Классика: история и современность» в Дмитрове


Всем по местам: в России назвали топ благополучных регионов

Водителя водовоза, врезавшегося в лайнер в Домодедово, парализовало за рулем

Чемпионат по зимнему плаванию пройдет в Пскове

В России назвали топ самых благополучных регионов по итогам 2023 года


В Самарской области будут судить наркобанду из 14 человек

Захарова: главы СЕ опозорились, оставив без реакции теракт в «Крокусе»

В Подмосковье сотрудники Росгвардии задержали подозреваемых в кражах из сетевого супермаркета

В Вешкайме загорелся склад с готовой продукцией местной птицефабрики



Путин в России и мире






Персональные новости Russian.city
Певица

Певица Ирина Понаровская перенесла концерт в Москве



News Every Day

Top 10 Emmanuelle Seigner Movies




Friends of Today24

Музыкальные новости

Персональные новости