Meta already has AI in its DNA – said Naila Murray, research lead at FAIR (Fundamental AI Research), at a roundup on AI organised by Meta in Berlin. The event includes panels, keynotes and workshops on red teaming. This is the name given to checking AI models for sources of errors and risks.
Advertisement
Murray says that AI has moved from the back seat to the front seat a bit. So instead of sitting in the background and providing Meta’s services, people can now use AI directly themselves. For example, Murray explains how Meta worked step by step to develop automatic detection and clipping for images. You know how it is, you can turn image sections into stickers in the Metas application; the boundaries of an object are automatically recognized; that’s what the AI does. However, it often has to be said that it is not always generative AI that has caused the hype since ChatGPT. Such applications have been in the works for a long time. Meta has been investing and researching in the field of AI for years.
Funding programs for start-ups and new infrastructure
The keynote is also about how you can participate in the PhD program at Meta. There is also a brand new AI Accelerator program for start-ups. To achieve this, Meta works with Hugging Face and Scaleway. According to Meta, the core of the program is access to the latest open models and resources in the field of AI, as well as guidance and access to the Startup Campus in Paris, Station F. You can apply till mid August,
US Supreme Court lifts ban on contacting social media companiesMeta has decided to offer its own AI model as open source. This was also an issue. Open source has advantages for the security of AI systems, it is said repeatedly. And open source will also benefit competition and progress. Basically, Meta is a company that stands for openness. And AI is a research field in which there has always been a close exchange between researchers, says Murray. An announcement that is relatively thin. The scientist says that a new infrastructure is needed to be able to use AI widely. More information will be available shortly.
German AI research
The participants on the forum changed; for example, Meta invited Jonas Gipping from the ALICE Institute in Tübingen, which was founded only last week. Professor Georg Rehm from the DFKI is there. For example, he says that the German Research Center for AI was founded in 1988 and that, given the current hype, one can hardly imagine such an old AI facility. Murray emphasizes that Germany is a place where a lot of important basic research in the field of AI has been done for a long time. If you look at the panel, you can really hope that Germany and Europe will not be left behind, as is often feared.
Read this also
However, Rehm also has a problem with the fact that the financing of projects in Germany is always very specific and short-term. On the other hand, in the United States there are projects that last ten or more years and receive a lot of money despite the high risk, i.e. a really abstract idea. In particular, he cites Apple’s Siri as an example, which essentially emerged from such a project. Funded by the US Defense Advanced Research Projects Agency, DARPA. Meta’s public policy director for Europe, Dr. Julia Reiss with Richard Wurz of the venture capital company Redstone. He assures Germany of great prospects.
When asked about artificial general intelligence, everyone agrees that a lot of research is still needed. However, it is also not clear how one can achieve AGI. Rehm explains that he believes large language models are not the way forward. And Jonas Gipping of the Ellis Institute also says that large language models are just language models, meaning they can handle language.
Multi agent systems are currently the focus of development at Meta when it comes to AGI. So there is no real general intelligence in the foreground, but rather agents that then learn from each other. However, what is not visible: social networks, which still represent the main source of income of Meta. Of course, they have AI and a lot of other research, which until now was not as obvious because AI is now in focus.
(EMW)
Sports industry: Federal Audit Office cuts off sports funding
