You can’t attend Transform 2022? View all summit sessions in our on-demand library now! See here.
AI revelations, innovations, and questions come in this week’s VentureBeat news. Deep learning turns 10 years old and insights from top leaders of the field like Yann LeCun and Geoffrey Hinton predict that deep learning shows no signs of slowing down.
Meanwhile, Melanie Mitchell, a professor at the Santa Fe Institute, warns engineering decision makers that on the whole, AI still needs three essential abilities to continue meaningful advances in the field: Understand concepts, form abstractions, and draw analogies.
In Mitchell’s view, AI can explain is growing and evolving rapidly to address some of these concerns – and MLops is at the helm of a number of solutions, ranging from the likes of Domino Data Lab, Qwak, ZenML and others. other law. Much work has yet to be done in space, but research is continuing.
Speaking of research – this week, Meta announced that AI Research Framework, PyTorch, is moving from under its vision and becoming part of the Linux Foundation. Zuckerberg noted that while the company still plans to fund PyTorch, Meta plans to take steps to separate itself from PyTorch over the next year.
In other news, Apple’s launch of iOS 16 sheds some light on what other tech giants can do in the future going passwordless. In the latest software update, Apple users can now use biometrics on their iPhone, iPad, and Mac devices for easier sign-in – with their biometrics synced via iCloud .
Here’s more from our top five tech stories of the week:
- Ten years later, a deep learning ‘revolution’ broke out, in the words of AI pioneers Hinton, LeCun, and Li.
Artificial intelligence (AI) pioneer Geoffrey Hinton, one of the leaders in the deep learning “revolution” that began a decade ago, says that the rapid advancement of AI will continue to accelerate. .
In an interview ahead of the 10th anniversary of the pivotal neural network research that led to a major AI breakthrough in 2012, Hinton and other leading AI innovators hit back at some of the critics. commented that deep learning has “hit a wall”.
Other AI road map breakers, including Yann LeCun, head of AI and principal scientist at Meta and Stanford University professor Fei-Fei Li, agree with Hinton that the results from the groundbreaking study in 2012 on the ImageNet database has propelled deep learning into the mainstream and has fueled a massive momentum that will be hard to stop.
- Apple iOS 16: Passkeys bring the trend of passwordless authentication
When it comes to security, passwords are often not an asset, but an obligation. They provide cybercriminals with an entry point to protected information that they can exploit with phishing schemes and social engineering efforts to manipulate users into transferring personal information.
With 15 billion passwords exposed online, something needs to change. Many vendors are claiming that the solution to this problem is to get rid of the password altogether.
Now, when Apple iOS 16 launches today with macOS Ventura, users will be able to sign in with Passkey on iPhone, iPad, and Mac, using biometric authentication options like Touch ID and Face ID, which are enabled sync on iCloud keychain.
- 3 Essential Abilities that AI Is Missing
As the AI community places increasing focus and resources on data-driven, deep learning approaches, Melanie Mitchell, a professor at the Santa Fe Institute, warns that what appears to be a performance hit human-like neural networks are, in fact, a shallow imitation that omits key components of intelligence.
Despite the advancement in deep learning, some of its problems remain. Among them, she says, are three essential: Understanding concepts, forming abstractions, and drawing analogies.
What is certain is that as AI becomes more pervasive in the applications we use every day, it will be important to create robust systems that are compatible with human intelligence and work – and fail. – in predictable ways.
- Why the market for explainable AI is growing rapidly
Powered by digital transformation, there seems to be no ceiling to the heights organizations will achieve in the next few years. One of the remarkable technologies that help businesses scale to new heights is artificial intelligence (AI).
As AI advances, there remains a nagging trust issue: AI is still not fully trusted by humans. At best, it’s being closely watched, and we’re still a long way from achieving human-AI synergy.
- PyTorch has a new home: Meta announces independent establishment
Meta announced today that their artificial intelligence (AI) research framework, PyTorch, has got a new home. It transitions to the independent PyTorch Foundation, which is part of the nonprofit Linux Foundation, a technology consortium with a core mission of collaborative open source software development.
Despite being released from direct supervision, Meta said it intends to continue using Pytorch as its primary AI research platform and will “financially support it”. However, Zuckerberg did note that the company plans to maintain a “clear separation between business administration and engineering” of the platform.
VentureBeat’s mission is a digital city square for technical decision-makers to gain knowledge of transformative enterprise technology and transactions. Explore our summary report.