Google has recently introduced a new family of Artificial Intelligence models that has medicine as its main focus. Named Med-Gemini, these AI models are not available to the public, but Google has published a pre-print version of the research paper that focuses on its capabilities and methodologies. Google claims that this new AI model will surpass GPT-4 in benchmark testing and its notable feature will be the long-context ability to process and analyze health records and research papers.
ALSO SEE: Best Mileage Cars in India
Med-Gemini Research Paper
ALSO SEE: Headphones Under ₹1,500
The research paper is currently in the pre-print stage and published on an open-access online repository of scholarly papers. Google DeepMind and Research’s Chief Scientist, Jeff Dean shared a post on microblogging platform X, claiming, “I’m very excited about the possibilities of these models to help clinicians deliver better care, as well as to help patients better understand their medical conditions. AI for healthcare is going to be one of the most impactful application domains for AI, in my opinion.”
ALSO SEE: Sony IMX Sensor Mobile Phones
The Med-Gemini Research Models have been built on top of Gemini 1.0 and Gemini 1.5 LLM. There are four models in totality, namely the Med-Gemini-S1.0, Med-Gemini-M 1.0, Med-Gemini-L 1.0 and Med-Gemini-M 1.5. All four of these can offer text, video, and image outputs and are integrated with web search, which Google claims have been made better through self-training to make the models “more faculty accurate, reliable and nuanced” when showing results for complex clinical reasoning tasks.
The AI model has been fine-tuned for better performance during long-context processing, claims Google. This would mean more accurate and pinpointed answers by chatbot even when questions are not perfectly framed or queried or need a documented medical record.
ALSO SEE: Tablets With Qualcomm Snapdragon CPU
ALSO SEE: Mobile Phones With UFS 4 Storage
According to Google’s data, Med-Gemini AI models have performed better than OpenAI’s GPT-4 models in the GeneTuring dataset on text-based reasoning tasks. Med-Gemini-L 1.0 has scored 91.1% accuracy on MedQA which betters its predecessor Med-PaLM 2 by 4.5%.
These new AI models are not available for public or any kind of beta testing as of now. Google is expected to improve these even more before launching them into the public arena.
You can follow Smartprix on Twitter, Facebook, Instagram, and Google News. Visit smartprix.com for the latest tech and auto news, reviews, and guides.