The UAE Launches Its Updated AI Series, the Falcon 2

uae, falcon 2, falcon 2 11B, falcon 2 11B VLM

The UAE takes the global stage, launching 2 new AI models; the Falcon 2 11B and the Falcon 2 11B VLM.

  • The Falcon 2 11B is a text-based model while the Falcon 2 11B VLM is a vision-to-language model.
  • Falcon 2 11B is trained on a massive dataset, allowing for a deeper understanding of language nuances and complex relationships.

The United Arab Emirates (UAE) launched the second iteration of its large language model (LLM), Falcon 2, trying to compete with established players like Meta and OpenAI.

UAE’s Technology Innovation Institute (TII) released 2 models in this series. The first, Falcon 2 11 B, is a text-based model. It is trained on a massive dataset of text and code. Specifically, on 5.5 trillion tokens and has 11 billion parameters. The high number of tokens, which are the learning material for an AI system, indicates it can understand the nuances of language and the relationships between words better than other systems. The high number of parameters, that the model tunes to learn complex relationships between the input data and the desired outcome indicates that the Falcon 2 11B can learn very complex relationships. However, its drawback is that the more parameters it has, the more power it requires to run. The text-based LLM is multilingual, processing tasks in various languages including French, English, Spanish, and many others.

Meanwhile, the second, Falcon 2 11B VLM, is a vision-to-language model. It’s TII’s first multimodal model, which can process and interpret more than one type of data input. This capability allows the Vision-to-Language LLM to understand the relationship between what it sees and what it reads. It can then analyze an image and generate a textual description of what it sees. Currently, it is the only one in the top-tier market that has this image-to-text conversion capability.

Performance-wise, the Falcon 2 11B surpasses the performance of Meta’s new Llama 3, which has 8 billion parameters, according to the institute’s blog post. It also “performs on par with Google’s Gemma 7B at first place, with a difference of only 0.01 average performance (Falcon 2 11B: 64.28 vs Gemma 7B: 64.29)”

Faisal Al Bannai, the secretary general of the Advanced Technology Research Council which funds TII, emphasized the LLM’s open-source nature. He said, “With other multimodal models soon coming to the market in various sizes, our aim is to ensure that developers and entities that value their privacy, have access to one of the best AI models to enable their AI journey.”

We can all agree that the Falcon 2 series is very promising. However, will the UAE be able to achieve its goal of competing with the major players?

The major AI players with which the UAE is trying to compete have a significant head start. Google, for example, has been in the AI game since the early days of its core products around 2001. Microsoft has been focusing on AI as a core area of investigation since it founded Microsoft Research in 1991. In contrast, the UAE did not take an interest in the sector until much later. In 2017, the UAE established a national AI strategy, “AI Strategy 2031.

These major AI companies also have immense financial resources to invest in hardware, talent, and datasets. Estimates suggest that Google has invested the most in AI, at a staggering $30.7 billion. A conservative estimate puts Microsoft’s investments in the sector over $16 billion. And there’s probably more where that came from, too. It’s difficult to pinpoint the exact number, but the UAE has invested around $2.15 billion in AI over the past decade. While it’s the region’s second-highest investor, its investments pale in comparison with the others.

There’s also the matter of marketing. If people don’t know about your services, their performance alone won’t sustain them. While the exact numbers of Google’s AI marketing budget are not publicly available, its marketing budget is a significant portion of its overall expenses. According to their Q1 2024, Google spent $6,426 million on sales and marketing. In contrast, the UAE’s AI initiatives, the Falcon 2 series, are still so new that there is less emphasis on marketing. It’s obvious that they are putting all their eggs in the ‘development’ basket instead.

Inside Telecom provides you with an extensive list of content covering all aspects of the tech industry. Keep an eye on our Intelligent Tech sections to stay informed and up-to-date with our daily articles.