Apple one step closer to becoming a major player in AI — research shows work on advanced artificial intelligence model that works with images and text in tandem
Apple’s work on advanced artificial intelligence integrations in its products is continuing to pick up pace, with a newly published paper from its research teams revealing the iPhone maker’s work on MM1, a suite of Multimodal Large Language Models.
The paper, titled ‘MM1: Methods, Analysis & Insights from Multimodal LLM Pre-training’ discusses “building performant Multimodal Large Language Models (MLLMs),” and how “pre-training using a careful mix of image-caption, interleaved image-text, and text-only data is crucial for achieving state-of-the-art few-shot results across multiple benchmarks, compared to other published pre-training results.”
It’s a pretty jargon-filled paper, but at its essence, it describes how Multimodal Large Language Models can produce more advanced and accurate workflows for AI applications by pulling from multiple datasets at once — not just text, but imagery too, as well as potentially audio and video sources.
MM1 is able to interpret information across 30 billion parameters, and as a result, can
“achieve competitive performance after supervised fine-tuning on a range of established multimodal benchmarks.” The researchers call MLLM “the next frontier in foundation models”, with “superior capabilities” to the large language models that have fuelled the breakthrough wave of AI tools in recent months and years.
Apple’s AI ambitions
However, MM1 remains behind closed doors for now, and there’s no guarantee it’ll ever morph into a consumer-facing product, even if the lessons learned from it find their way into other applications of AI.
At present, we know that Apple is doubling down on its AI plans, following shareholder pressure that may have inadvertently led to the cancellation of its driverless Apple car project. In response to shareholder unrest, CEO Tim Cook stated that Apple will “break new ground” in AI, unlocking “transformative opportunities” for users.
Its efforts so far focus on a $1 billion research and development push for a large language model named Ajax, with the company also purchasing a Canadian artificial intelligence startup called DarwinAI.
Apple, at least in a public-facing sense, is playing catch up against competitors like Google and Microsoft, who are already implementing their respective Gemini and CoPilot AI tools in consumer products. iOS 18, set to be revealed at WWDC 2024, is likely to be the platform where Apple pulls back the curtain on its major AI efforts so far, though the company has been keen to stress its products (via their Neural Engine chip elements) are already making use of artificial intelligence principles. It’s calling its recent M3 MacBook Air release “the world’s best consumer laptop for AI.”
More from iMore
Apple’s work on advanced artificial intelligence integrations in its products is continuing to pick up pace, with a newly published paper from its research teams revealing the iPhone maker’s work on MM1, a suite of Multimodal Large Language Models.
The paper, titled ‘MM1: Methods, Analysis & Insights from Multimodal LLM Pre-training’ discusses “building performant Multimodal Large Language Models (MLLMs),” and how “pre-training using a careful mix of image-caption, interleaved image-text, and text-only data is crucial for achieving state-of-the-art few-shot results across multiple benchmarks, compared to other published pre-training results.”
It’s a pretty jargon-filled paper, but at its essence, it describes how Multimodal Large Language Models can produce more advanced and accurate workflows for AI applications by pulling from multiple datasets at once — not just text, but imagery too, as well as potentially audio and video sources.
MM1 is able to interpret information across 30 billion parameters, and as a result, can
“achieve competitive performance after supervised fine-tuning on a range of established multimodal benchmarks.” The researchers call MLLM “the next frontier in foundation models”, with “superior capabilities” to the large language models that have fuelled the breakthrough wave of AI tools in recent months and years.
Apple’s AI ambitions
However, MM1 remains behind closed doors for now, and there’s no guarantee it’ll ever morph into a consumer-facing product, even if the lessons learned from it find their way into other applications of AI.
At present, we know that Apple is doubling down on its AI plans, following shareholder pressure that may have inadvertently led to the cancellation of its driverless Apple car project. In response to shareholder unrest, CEO Tim Cook stated that Apple will “break new ground” in AI, unlocking “transformative opportunities” for users.
Its efforts so far focus on a $1 billion research and development push for a large language model named Ajax, with the company also purchasing a Canadian artificial intelligence startup called DarwinAI.
Apple, at least in a public-facing sense, is playing catch up against competitors like Google and Microsoft, who are already implementing their respective Gemini and CoPilot AI tools in consumer products. iOS 18, set to be revealed at WWDC 2024, is likely to be the platform where Apple pulls back the curtain on its major AI efforts so far, though the company has been keen to stress its products (via their Neural Engine chip elements) are already making use of artificial intelligence principles. It’s calling its recent M3 MacBook Air release “the world’s best consumer laptop for AI.”
More from iMore
Denial of responsibility! Techno Blender is an automatic aggregator of the all world’s media. In each content, the hyperlink to the primary source is specified. All trademarks belong to their rightful owners, all materials to their authors. If you are the owner of the content and do not want us to publish your materials, please contact us by email – [email protected]. The content will be deleted within 24 hours.