Connect with us

Published

on

inflection point — The generative AI revolution has begunhow did we get here? A new class of incredibly powerful AI models has made recent breakthroughs possible.

HaomiaoHuang – Jan 30, 2023 12:00 pm UTC Enlarge / This image was partially AI-generated with the prompt “a pair of robot hands holding pencils drawing a pair of human hands, oil painting, colorful,” inspired by the classic M.C. Escher drawing. Watching AI mangle drawing hands helps us feel superior to the machines… for now. AurichAurich Lawson | Stable Diffusion reader comments 38 with 0 posters participating Share this story Share on Facebook Share on Twitter Share on Reddit

Progress in AI systems often feels cyclical. Every few years, computers can suddenly do something theyve never been able to do before. Behold! the AI true believers proclaim, the age of artificial general intelligence is at hand! Nonsense! the skeptics say. Remember self-driving cars?

The truth usually lies somewhere in between.

Were in another cycle, this time with generative AI. Media headlines are dominated by news about AI art, but theres also unprecedented progress in many widely disparate fields. Everything from videos to biology, programming, writing, translation, and more is seeing AI progress at the same incredible pace. Why is all this happening now?

Further ReadingThe basics of modern AIhow does it work and will it destroy society this year?You may be familiar with the latest happenings in the world of AI. Youve seen the prize-winning artwork, heard the interviews between dead people, and read about the protein-folding breakthroughs. But these new AI systems arent just producing cool demos in research labs. Theyre quickly being turned into practical tools and real commercial products that anyone can use.

Theres a reason all of this has come at once. The breakthroughs are all underpinned by a new class of AI models that are more flexible and powerful than anything that has come before. Because they were first used for language tasks like answering questions and writing essays, theyre often known as large language models (LLMs). OpenAIs GPT3, Googles BERT, and so on are all LLMs. Advertisement

But these models are extremely flexible and adaptable. The same mathematical structures have been so useful in computer vision, biology, and more that some researchers have taken to calling them “foundation models” to better articulate their role in modern AI.

Where did these foundation models came from, and how have they broken out beyond language to drive so much of what we see in AI today? The foundation of foundation models

Theres a holy trinity in machine learning: models, data, and compute. Models are algorithms that take inputs and produce outputs. Data refers to the examples the algorithms are trained on. To learn something, there must be enough data with enough richness that the algorithms can produce useful output. Models must be flexible enough to capture the complexity in the data. And finally, there has to be enough computing power to run the algorithms.

The first modern AI revolution took place with deep learning in 2012, when solving computer vision problems with convolutional neural networks (CNNs) took off. CNNs are similar in structure to the brain’s visual cortex. Theyve been around since the 1990s but werent yet practical due to their intense computing power requirements.

In 2006, though, Nvidia released CUDA, a programming language that allowed for the use of GPUs as general-purpose supercomputers. In 2009, Stanford AI researchers introduced Imagenet, a collection of labeled images used to train computer vision algorithms. In 2012, AlexNet combined CNNs trained on GPUs with Imagenet data to create the best visual classifier the world had ever seen. Deep learning and AI exploded from there.

CNNs, the ImageNet data set, and GPUs were a magic combination that unlocked tremendous progress in computer vision. 2012 set off a boom of excitement around deep learning and spawned whole industries, like those involved in autonomous driving. But we quickly learned there were limits to that generation of deep learning. CNNs were great for vision, but other areas didnt have their model breakthrough. One huge gap was in natural language processing (NLP)i.e., getting computers to understand and work with normal human language rather than code. Advertisement

The problem of understanding and working with language is fundamentally different from that of working with images. Processing language requires working with sequences of words, where order matters. A cat is a cat no matter where it is in an image, but theres a big difference between this reader is learning about AI and AI is learning about this reader.

Until recently, researchers relied on models like recurrent neural networks (RNNs) and long short-term memory (LSTM) to process and analyze data in time. These models were effective at recognizing short sequences, like spoken words from short phrases, but they struggled to handle longer sentences and paragraphs. The memory of these models was just not sophisticated enough to capture the complexity and richness of ideas and concepts that arise when sentences are combined into paragraphs and essays. They were great for simple Siri- and Alexa-style voice assistants but not for much else.

Getting the right training data was another challenge. ImageNet was a collection of one hundred thousand labeled images that required significant human effort to generate, mostly by grad students and Amazon Mechanical Turk workers. And ImageNet was actually inspired by and modeled on an older project called WordNet, which tried to create a labeled data set for English vocabulary. While there is no shortage of text on the Internet, creating a meaningful data set to teach a computer to work with human language beyond individual words is incredibly time-consuming. And the labels you create for one application on the same data might not apply to another task. Page: 1 2 3 4 5 6 7 8 Next → reader comments 38 with 0 posters participating Share this story Share on Facebook Share on Twitter Share on Reddit Advertisement Channel Ars Technica ← Previous story Next story → Related Stories Today on Ars

Continue Reading

Business

Magnum debut suffers a chill as Ben & Jerry’s row lingers

Published

on

By

Magnum debut suffers a chill as Ben & Jerry's row lingers

Shares in The Magnum Ice Cream Company (TMICC) have fallen slightly on debut after the completion of its spin-off from Unilever amid a continuing civil war with one of its best-known brands.

Shares in the Netherlands-based company are trading for the first time following the demerger.

It creates the world’s biggest ice cream company, controlling around one fifth of the global market.

Primary Magnum shares, in Amsterdam, opened at €12.20 – down on the €12.80 reference price set by the EuroNext exchange, though they later settled just above that level, implying a market value of €7.9bn – just below £7bn.

The company is also listed in London and New York.

Money latest: The cheapest days to travel by plane

Unilever stock was down 3.1% on the FTSE 100 in the wake of the spin off.

More from Money

The demerger allows London-headquartered Unilever to concentrate on its wider stable of consumer brands, including Marmite, Dove soap and Domestos.

The decision to hive off the ice cream division, made in early 2024, gives a greater focus on a market that is tipped to grow by up to 4% each year until 2029.

Ben & Jerry's accounts for a greater volume of group revenue now under TMICC. Pic: Reuters
Image:
Ben & Jerry’s accounts for a greater volume of group revenue now under TMICC. Pic: Reuters

But it has been dogged by a long-running spat with the co-founders of Ben & Jerry’s, which now falls under the TMICC umbrella and accounts for 14% of group revenue.

Unilever bought the US brand in 2000, but the relationship has been sour since, despite the creation of an independent board at that time aimed at protecting the brand’s social mission.

The most high-profile spat came in 2021 when Ben & Jerry’s took the decision not to sell ice cream in Israeli-occupied Palestinian territories on the grounds that sales would be “inconsistent” with its values.

Unilever responded by selling the business to its licensee in Israel.

A series of rows have followed akin to a tug of war, with Magnum refusing repeated demands by the co-founders of Ben & Jerry’s to sell the brand back.

Please use Chrome browser for a more accessible video player

Sept: ‘Free Ben & Jerry’s’

Magnum and Unilever argue its mission has strayed beyond what was acceptable back in 2000, with the brand evolving into one-sided advocacy on polarising topics that risk reputational and business damage.

TMICC is currently trying to remove the chair of Ben & Jerry’s independent board.

It said last month that Anuradha Mittal “no longer meets the criteria” to serve after internal investigations.

An audit of the separate Ben & Jerry’s Foundation, where she is also a trustee, found deficiencies in financial controls and governance. Magnum said the charitable arm risked having funding removed unless the alleged problems were addressed.

The Reuters news agency has since reported that Ms Mittal has no plans to quit her roles, and accused Magnum of attempts to “discredit” her and undermine the authority of the independent board.

Magnum boss Peter ter Kulve said on Monday: “Today is a proud milestone for everyone associated with TMICC. We became the global leader in ice cream as part of the Unilever family. Now, as an independent listed company, we will be more agile, more focused, and more ambitious than ever.”

Commenting on the demerger, Hargreaves Lansdown equity analyst Aarin Chiekrie said: “TMICC is already free cash flow positive, and profitable in its own right. The balance sheet is in decent shape, but dividends are off the cards until 2027 as the group finds its footing as a standalone business.

“That could cause some downward pressure on the share price in the near term, as dividend-focussed investment funds that hold Unilever will be handed TMICC shares, the latter of which they may be forced to sell to abide by their investment mandate.”

Continue Reading

UK

Ex-footballer Joey Barton sentenced for posting grossly offensive social media messages

Published

on

By

Ex-footballer Joey Barton sentenced for posting grossly offensive social media messages

Retired footballer Joey Barton has been sentenced over X posts he sent to football pundits Eni Aluko and Lucy Ward, along with broadcaster Jeremy Vine.

Barton, 43, had been found guilty of six counts of sending a grossly offensive electronic communication with intent to cause distress or anxiety.

He was sentenced to a six-month prison sentence, suspended for 18 months.

The former Manchester City, Newcastle United and Rangers midfielder had claimed he was the victim of a “political prosecution” and denied his aim was to “get clicks and promote himself”.

But the jury decided Barton, capped once for England in 2007, had “crossed the line between free speech and a crime” with the six posts he made on the social media platform.

The prosecution argued that Barton, who has 2.5 million followers, “may well be characterised as cutting, caustic, controversial and forthright”.

Peter Wright KC continued: “Everyone is entitled to express views that are all of those things.

“What someone is not entitled to do is to post communications electronically that are – applying those standards – beyond the pale of what is tolerable in society.”

Barton denied 12 counts of sending a grossly offensive electronic communication with intent to cause distress or anxiety between January and March last year.

He was found guilty on six counts, but cleared of another six.

In one post in January 2024, Barton compared Aluko and Ward to the “Fred and Rose West of football commentary”, and superimposed the women’s faces on a photograph of the serial murderers.

He also described Aluko as being in the “Joseph Stalin/Pol Pot category”, suggesting that she had “murdered hundreds of thousands, if not millions, of football fans’ ears”.

The jury found him not guilty in relation to the comparison with the Wests, Stalin and Pol Pot, but decided the superimposed image was grossly offensive.

Another message allegedly suggested Vine had a sexual interest in children, after the broadcaster posted a question relating to the posts about the football commentators asking whether Barton had a “brain injury”.

The ex-footballer told the court the posts were “dark and stupid humour” and “crude banter”. He also said he had no intention of implying Vine was a paedophile.

This breaking news story is being updated and more details will be published shortly.

Please refresh the page for the latest version.

You can receive breaking news alerts on a smartphone or tablet via the Sky News app. You can also follow us on WhatsApp and subscribe to our YouTube channel to keep up with the latest news.

Continue Reading

UK

Storm Bram named as weather warnings issued for UK and Ireland

Published

on

By

Storm Bram named as weather warnings issued for UK and Ireland

Storm Bram has been named by the Irish weather service – with warnings for strong winds and heavy rain issued for parts of the UK and Ireland.

More than half a month’s rainfall could hit some parts of the UK in just a 24-hour period, the Met Office has warned.

Yellow and orange warnings are in place across Ireland today and tomorrow, with “very strong to gale force” winds forecast on Tuesday.

Check the weather forecast where you are

The Met Office said strong winds forecast from Monday evening through until Wednesday could cause disruption, with gusts of 50-60mph predicted widely and 70-80mph in some places.

A yellow weather warning for rain comes into effect from 6pm on Monday, and will be in place for 24 hours, covering parts of southwest England and Wales, and stretching to parts of Herefordshire and Hampshire.

The Met Office has also issued a yellow warning for high winds from Dorset to Cornwall and up to north Wales, in place from 10pm on Monday until 4pm on Tuesday.

It said transport networks could face disruption, with delays for high-sided vehicles on exposed routes and bridges, and coastal roads and seafronts affected by spray and large waves. Power outages are also possible.

For 24 hours from 6pm on Monday, up to 40mm of rain could fall in some areas, with 60-80mm of rain over Dartmoor and high ground in South Wales, which would amount to more than half the average monthly rainfall in December.

The predicted rainfall across southwest England and South Wales is expected to hit already saturated ground and could lead to difficult travel conditions.

An amber warning for wind has been issued for northwest Scotland on Tuesday.

Flying debris “could result in a danger to life” – and there could be damage to buildings and homes along with the risk of roofs being “blown off” due to the “very strong and disruptive winds”, the Met Office warned.

Forecasters added there was the potential for large waves and beach material “being thrown” across sea fronts, roads and properties.

There are also yellow warnings for wind and rain on Tuesday across Northern Ireland, Scotland, Wales and northern and southwest England.

Weather warnings issued for Tuesday. Pic: Met Office
Image:
Weather warnings issued for Tuesday. Pic: Met Office

Yellow warnings for wind have been issued for Scotland and parts of northern England on Wednesday.

The Met Office’s deputy chief meteorologist, Steven Keates, said: “A deepening area of low pressure will approach the UK from the southwest later on Monday, bringing with it heavy rain and strong winds, which are likely to affect the UK between late Monday and early Wednesday.

“The exact track, depth and timings of this low are uncertain, which makes it harder to determine where will be most impacted by strong winds and/or heavy rain.

“This system has the potential to cause disruption, and severe weather warnings are likely to be issued over the weekend as details become clearer. We therefore urge people to keep up-to-date with the latest Met Office forecast.”

The Met Office said the rest of the month remained unsettled, with further periods of low pressure predicted.

It also said it is too early to provide an accurate forecast for the Christmas period.

Continue Reading

Trending