- | 1:00 pm
Abu Dhabi-based TII launches its own version of ChatGPT
Falcon LLM is a foundational large language model with 40 billion parameters, outperforming ChatGPT 3.
With its ability to answer questions, write code and create content cleverly, OpenAI’s ChatGPT continues to make waves, and now, the race to beat the chatbot is on.
Entering the competitive realm is Abu Dhabi as it builds its homegrown rival to ChatGPT.
The Technology Innovation Institute (TII), the applied research pillar of Abu Dhabi’s Advanced Technology Research Council (ATRC), has launched Falcon LLM, a foundational large language model (LLM) with 40 billion parameters, outperforming ChatGPT 3.
Falcon LLM is trained with over one trillion collected web data tokens. The innovation is built by the AI and Digital Science Research Center’s (AIDRC) AI Cross-Center Unit, the team behind building Noor, the world’s largest Arabic AI language processing model.
Today’s LLMs, which can understand and generate texts, have achieved major advancements and can be used across various applications, such as chatbots, virtual assistants, language translation, content generation, and sentiment analysis. It also offers countless business possibilities – from streamlining customer service to improving efficiencies, cutting labor costs, and identifying new revenue streams.
Dr. Ray O. Johnson, CEO of TII, commended the UAE-born innovation, “This is just the beginning. By the end of the year, we will be sharing news on a huge increase in capabilities in this space. We understand that this is the start of a momentous journey, [and] we will press on to give this region its own AI success stories, well-aligned with the UAE’s National AI Strategy.”
Prof. Mérouane Debbah, Chief Researcher AI and Digital Science Research Center, said the milestone significantly impacts various fields such as education, healthcare, film production, and CGI.
Meanwhile, Dr. Ebtesam Almazroui, Director of the AI Cross-Center Unit at AIDRC, pointed out that the Falcon LLM 40B model outperforms other models in the market and improves upon new-generation models. “For instance, it matches the performance of Chinchilla (from DeepMind) and PaLM-62B (from Google), at considerably lower training costs in comparison.”
Though the model isn’t commercially available, the plan is to eventually offer it to government entities, startups, and the private sector.