Meta Platforms Inc Unveils AI Language Model LLama to Help Researchers Improve AI Tools
Meta Platforms Inc (META.O) recently announced the release of a new large language model based on Artificial Intelligence (AI). The model, called Large Language Model Meta AI (LLaMA), was developed by Meta’s Fundamental AI Research (FAIR) team, and it is aimed at the research community. It is the latest addition to the AI race, with large tech companies and well-capitalized startups alike vying to tout advances in AI technology and integrate it into commercial products.[0]
LLaMA is a subset of AI deep learning and uses artificial neural networks to attempt to simulate the behavior of the human brain.[1] These neural networks can learn from large amounts of data and demonstrate widespread knowledge.[1] Language models have become one of the fastest-growing fields in AI and are being used for tasks such as summarization, writing stories, answering questions, and completing codes.[2]
Meta’s research team has developed a new AI language model called Toolformer that can teach itself how to choose and use external tools such as search engines, calculators, and translation systems via API calls to boost its performance on downstream tasks.[3] Toolformer requires a minimal amount of human annotation and can choose which tool to use without sacrificing generality. Experiments conducted by the researchers on various tool-using tasks show that Toolformer can outperform the much larger GPT-3 model.[4]
Meta has committed to making LLaMA available to the research public, and the company is taking applications from researchers and institutions. Unlike ChatGPT and Bing, which are open to everyone, Meta is limiting the use of LLaMA by releasing it under a noncommercial license, primarily for research purposes.
Meta has said that its new model can help researchers improve and fix AI tools that promote “misinformation.” The company believes that further research into the field of AI can help solve issues such as bias, toxicity, and the potential for generating misinformation that generative AI tools can pose. For now, the model is not being used in Meta’s products, but the company plans to make the technology available to AI researchers.[5]
0. “Meta rolls out new language model amid Big Tech's AI push” Vom News – Voice Of Masses, 24 Feb. 2023, https://vomnews.in/meta-rolls-out-new-language-model-amid-big-techs-ai-push
1. “Meta releases LLaMA to democratize access to large language AI models” SiliconANGLE News, 24 Feb. 2023, https://siliconangle.com/2023/02/24/meta-platforms-releases-llama-democratize-access-large-language-ai-models/
2. “A New AI Research Proposes Compositional Exemplars for In-context Learning (CEIL): A Novel Approach That Uses Determinantal Point Process (DPP) for More Efficient In-context Learning” MarkTechPost, 22 Feb. 2023, https://www.marktechpost.com/2023/02/21/a-new-ai-research-proposes-compositional-exemplars-for-in-context-learning-ceil-a-novel-approach-that-uses-determinantal-point-process-dpp-for-more-efficient-in-context-learning/
3. “Meta develops an AI language tool: Toolformer” Research Snipers, 16 Feb. 2023, https://researchsnipers.com/meta-develops-an-ai-language-tool-toolformer
4. “Meta AI & UPF’s Toolformer: Enabling Language Models to Teach Themselves to Use External Tools” Synced, 16 Feb. 2023, https://syncedreview.com/2023/02/16/meta-ai-upfs-toolformer-enabling-language-models-to-teach-themselves-to-use-external-tools
5. “Meta Is Introducing AI Large Language Model Called ‘Llama'” Yahoo! Voices, 24 Feb. 2023, https://www.yahoo.com/now/meta-introducing-ai-large-language-162336151.html
