Connect with us

Technology

DeepMind AI gets silver medal at International Mathematical Olympiad

Published

on

DeepMind AI gets silver medal at International Mathematical Olympiad

DeepMind’s AlphaProof AI can tackle a range of mathematical problems

Google DeepMind

An AI from Google DeepMind has achieved a silver medal score at this year’s International Mathematical Olympiad (IMO), the first time any AI has made it to the podium.

The IMO is considered the world’s most prestigious competition for young mathematicians. Correctly answering its test questions requires mathematical ability that AI systems typically lack.

Advertisement

In January, Google DeepMind demonstrated AlphaGeometry, an AI system that could answer some IMO geometry questions as well as humans. However, this was not from a live competition, and it couldn’t answer questions from other mathematical disciplines, such as number theory, algebra and combinatorics, which is necessary to win an IMO medal.

Google DeepMind has now released a new AI, called AlphaProof, which can solve a wider range of mathematical problems, and an improved version of AlphaGeometry, which can solve more geometry questions.

When the team tested both systems together on this year’s IMO questions, they answered four out of six questions correctly, giving them a score of 28 out of a possible 42 points. This was enough to win a silver medal and just one point under this year’s gold medal threshold.

At the contest in Bath, UK, last week, 58 entrants won a gold medal and 123 won a silver medal.

Advertisement

“We are all very much aware that AI will eventually be better than humans at solving most mathematical problems, but the rate at which AI is improving is breathtaking,” says Gregor Dolinar, the IMO president. “Missing the gold medal at IMO 2024 by just one point a few days ago is truly impressive.”

At a press conference, Timothy Gowers at the University of Cambridge, who helped mark AlphaProof’s answers, said the AI’s performance was surprising and it appeared to find “magic keys” to answer problems in a similar way to humans. “I thought that these magic keys would probably be a little bit beyond what it could do, so it came as quite a surprise in one or two instances when the program had indeed found these keys,” said Gowers.

AlphaProof works similarly to Google DeepMind’s previous AIs that can beat the best humans at chess and Go. All of these AIs rely on a trial-and-error approach called reinforcement learning,  where the system finds its own way to solve a problem over many attempts. However, this method requires a large set of problems written in language that the AI can understand and verify, whereas most IMO-like problems are written in English.

Advertisement

To get around this, Thomas Hubert at DeepMind and his colleagues used Google’s Gemini AI, a language model like the one that powers ChatGPT, to translate these problems into a programming language called Lean so that the AI could learn how to solve them.

“At the beginning, it will be able to solve perhaps the simplest problems, and learn from solving those simpler problems to attack harder and harder problems,” Hubert said at the press conference. It also produces its answers in Lean, so they can be instantly verified as correct.

While AlphaProof’s performance is impressive, it works slowly, taking up to three days to find some solutions instead of the 4.5 hours per three questions that competitors are allowed. It also failed to answer both questions on combinatorics, which is the study of counting and arranging numbers. “We are still working to understand why this is, which will hopefully lead us to improve the system,” says Alex Davies at Google DeepMind.

It is also not clear how AlphaProof arrives at its answers or whether it uses the same kind of mathematical intuitions that humans do, said Gowers, but its ability to translate proofs from Lean into English makes it easy to check they are correct.

Advertisement

The result is impressive and a significant milestone, says Geordie Williamson at the University of Sydney, Australia. “There have been many previous attempts to do reinforcement learning on formal proofs and none have had much success.”

While a system like AlphaProof could be useful for working mathematicians in helping develop proofs, it obviously can’t help with identifying problems to solve and work on, which takes up a large portion of researchers’ time, says Yang-Hui He at the London Institute for Mathematical Sciences.

Hubert said his team hopes that AlphaProof will be able to help improve Google’s large language models, like Gemini, by reducing incorrect responses.

The trading company XTX Markets has offered a $5 million prize – called the AI Mathematical Olympiad – for an AI capable of achieving a gold medal at the IMO, but AlphaProof is not eligible because it is not publicly available. “We hope that DeepMind’s advances will inspire more teams to enter the AIMO Prize, and would of course welcome a public entry from DeepMind themselves,” says Alex Gerko at XTX Markets.

Advertisement

Topics:

Source link

Continue Reading
Advertisement
Click to comment

You must be logged in to post a comment Login

Leave a Reply

Technology

Data platform Airbyte can now create API connectors by reading the docs

Published

on

Data platform Airbyte can now create API connectors by reading the docs

If your startup is only remotely related to working with data pipelines, you’re probably trying to figure out how to capitalize on the current moment: Enterprises are trying to figure out how to best use data to power generative AI products, and to do that, they need robust data services. Airbyte, which launched in 2020, started with a focus on building a low-code/no-code open source data integration platform. Since then, Airbyte raised a total of $181.2 million, including a massive $150 million Series B round during the somewhat anomalous days of late 2021.

After four years, the company is now launching Airbyte 1.0 — and the focus, of course, is on AI, both as an addition to Airbyte’s own tools and to help its users build their own AI-based services.

Indeed, the company is now leveraging AI in a clever way to expand on its overall low-code/no-code philosophy: Its model will be able to look at the documentation for an API and automatically create a connector based on that. You simply point it at the documentation, and it’ll handle the rest (at least in theory; time will tell how well that works in practice, of course.)

As Airbyte co-founder and CEO Michel Tricot told me, he believes that one area where large language models are transforming how enterprise use their data is by making unstructured data far more useful — and usable.

Advertisement

“Structured data is just the tip of the iceberg when it comes to leveraging data’s full potential,” he said. “With the rise of LLMs, we can now efficiently tap into previously untouched unstructured data. … We’ve seen massive demand for handling multi-modal data. Our latest developments have been geared toward supporting intelligent, context-aware pipelines, optimizing frameworks like RAG, and automating pipeline creation based on customer data workflows. These innovations are crucial to unlocking advanced use cases and enhancing LLM performance.”

Because Airbyte is now so much better at managing unstructured data, its users can now leverage their existing pipelines to do that, without having to rely on additional tools.

In non-AI news, Airbyte’s connector now also supports GraphQL, which should help users access many additional datasets without even having to build custom pipelines.

With this release, Airbyte is also making its self-managed enterprise service generally available. Like with so many open source companies, the enterprise version, which is available on the AWS and GCP marketplaces, will offer features like single sign-on (SSO) and role-based access control (RBAC), as well as Airbyte-specific features like sensitive data masking and advanced observability.

Advertisement

Airbyte says it has 7,000 enterprise customers and has seen over 170,000 deployments by now. Its customers range from Calendly and Coupa to Perplexity AI and Siemens.

“Every company is a data company — to drive decision-making and as the foundation for AI initiatives,” Tricot said. “Only Airbyte, with our open source strategy enabling hundreds of connectors, can give enterprises the ability to leverage any data they choose. As AI continues to drive transformation, we’re delivering the technology and ecosystem required for organizations to build the data infrastructure needed for AI-driven innovation.”

Source link

Advertisement
Continue Reading

Servers computers

Types of Server Machines | Tower Servers | Rack Servers | Blade Servers

Published

on

Types of Server Machines | Tower Servers | Rack Servers | Blade Servers



What are the various types of server machines ? Which one is suitable for your business ?

tower server vs rack server vs blade server

सर्वर मशीन्स कितने प्रकार के होते हैं ? आपके बिज़नेस के लिए कौन सा सर्वर मशीन सबसे अच्छा रहेगा ?

Facebook: https://www.facebook.com/ITSimplifiedInHindi

Advertisement

Image Credits:

1. https://goo.gl/wvx7ve
2. https://goo.gl/d6SNzN
3. https://goo.gl/hKzST9
4. https://goo.gl/9DhEzX
5. https://goo.gl/rWxD4X
6. https://goo.gl/USC5Ro
7. https://goo.gl/qyA2cV
8. https://goo.gl/9VCama
9. https://goo.gl/SnmvDp
10. https://goo.gl/Lhxijy
11. https://goo.gl/ij9G2t
12. https://goo.gl/7PFPrv
13. https://goo.gl/a9vcMy
14. https://goo.gl/KEDGfh
15. https://goo.gl/foacPQ
16. https://goo.gl/5gXYaZ
17. https://goo.gl/JRUPWN
18. https://goo.gl/qBSkRQ
19. https://goo.gl/MoS716
20. https://goo.gl/e8ak1h
21. https://goo.gl/YJWKS9
22. https://goo.gl/DRmV9h
23. https://goo.gl/3C8nwP
24. https://goo.gl/nGWLXc
25. https://goo.gl/LFCmcg
26. https://goo.gl/DrPruZ
27. https://goo.gl/av7WPw
28. https://goo.gl/8NYyBD
29. https://goo.gl/MmPUko .

source

Continue Reading

Technology

Ghost of Yotei is the next entry in Sucker Punch’s Ghosts series

Published

on

Ghost of Yotei is the next entry in Sucker Punch’s Ghosts series

To close out Sony’s State of Play presentation, Sucker Punch debuted a first look at the next game in its award-winning Ghosts series. Dubbed Ghost of Yotei, players will inhabit a new female protagonist named Atsu, who wields a shamisen in addition to a samurai’s traditional sword. According to the trailer description, Atsu will travel the wilds around Mount Yotei accompanied by a mysterious wolf companion.

Sucker Punch Studios, the developers behind the Sly Cooper and Infamous series, released Ghosts of Tsushima back in 2020. The game was praised for its quiet, contemplative moments as much as it’s Akira Kurosawa-style samurai action. Earlier this year, Sucker Punch released a PC port of Tsushima and confirmed that a movie based on the game is in the works, with John Wick director Chad Stahelski tapped to lead the project.

In the blog post announcing the game, Sony stated that Ghost of Yotei was Sucker Punch’s first game built specifically around the PS5. According to the developers, Ghost of Yotei will offer the same kind of beautiful observation moments as its predecessor while adding new features as well.

Ghost of Yotei
Image: Sony
Advertisement

“We have massive sightlines that let you look far across the environment, whole new skies featuring twinkling stars and auroras, even more believable movement from wind on grass and vegetation,” said Andrew Goldfarb, communications manager at Sucker Punch. “Our new setting also gives us the opportunity to introduce new mechanics, gameplay improvements, and even new weapons.”

Sony didn’t share release date details, only pegging the game to a 2025 release window.

Source link

Advertisement
Continue Reading

Technology

Watch out Dyson: these 3 radically different hair dryers are making haircare exciting again

Published

on

Three hair dryers from IFA 2024 that impressed me; the Dreame Pocket (right) Laifen Mini (middle) and Shark FlexFusion (left)

There were plenty of devices at IFA 2024 that showed how technology in general is getting creative with form and function, but a lot of eyes were on the beauty space as it continues to radically reinvent what hair styling devices look and feel like. 

Gone are the days of conventional hair dryers being the market norm; the once iconic pistol-shaped handheld dryers are shifting into new form factors, making the best hair dryers interesting once again. 

Source link

Continue Reading

Servers computers

Dell Rack Servers | Price List | Refurbprice

Published

on

Dell Rack Servers | Price List | Refurbprice



#Dell #Rack #Servers | #PriceList | #Refurbprice

https://www.refurbprice.com/refurbished-dell-rack-servers-price-list

source

Continue Reading

Technology

Why Machines Learn: A clever primer makes sense of what makes AI possible

Published

on

Members of the medical staff at the Elithair clinic conduct an artificial Intelligence analysis of Felix Hofmann
Members of the medical staff at the Elithair clinic conduct an artificial Intelligence analysis of Felix Hofmann's scalp and his hair roots.

Machine learning is key to developments in medical diagnostics

Mario Heller/Panos Pictures

Why Machines Learn
Anil Ananthaswamy (Allen Lane (UK); Penguin Random House (US))

As someone who writes for a living, I routinely feel assaulted by the onslaught of generative artificial intelligence. How long before I become a mere massager of prompted paragraphs, the joy of creation abandoned in favour of more, faster, cheaper?

Advertisement

Anil Ananthaswamy’s Why Machines Learn: The elegant maths behind modern AI won’t tell me or you about the future of AI in our society, nor what we should do about it. But whether you regard the algorithms used in facial recognition,…

Source link

Advertisement
Continue Reading

Trending

Copyright © 2017 Zox News Theme. Theme by MVP Themes, powered by WordPress.