- M42 introduces upgraded version of Med42 AI model at Abu Dhabi Global Healthcare Week.
- Med42 aims to set new standards in healthcare AI accuracy, efficiency, and accessibility.
- Second iteration boasts 70-billion parameter finetuned model based on Llama-3.
- Performance rivals proprietary models with zero-shot accuracy of 85.1% and max accuracy of 87.3% on USMLE.
- Med42 streamlines healthcare processes, aids in crafting tailored treatment strategies, and serves as clinical decision support tool.
- Accessible via HuggingFace platform, enabling global testing and innovation.
Main AI News:
M42 proudly announces the launch of the upgraded iteration of its revolutionary open-access generative AI Clinical Large Language Model (LLM), Med42, at the prestigious Abu Dhabi Global Healthcare Week. This unveiling marks a significant leap forward in AI-driven healthcare innovation, poised to revolutionize patient care while alleviating administrative burdens for clinicians.
Following the success of its initial release, Med42, introduced in October 2023, aimed to establish a new benchmark for accuracy, efficiency, and accessibility in healthcare AI. With an impressive score of 72 percent on the United States Medical Licensing Examination (USMLE) Sample Exam questions, Med42 outperformed industry giants such as OpenAI’s GPT3.5 and Google’s MedPaLM models. Garnering over 8,200 downloads from independent developers who have rigorously tested and validated its capabilities, Med42 has garnered global recognition and acclaim.
The second iteration represents a fully optimized 70-billion parameter finetuned model, built upon Llama-3, Meta’s latest version of open-access LLMs. Demonstrating performance comparable to proprietary (closed) models like OpenAI’s GPT-4 and Google’s Med-Gemini, Med42 boasts a zero-shot accuracy score of 85.1 percent and a maximum accuracy score of 87.3 percent using specialized prompting on the USMLE. This underscores Med42’s efficiency and accessibility relative to larger, closed models, solidifying its status as a frontrunner in open-access clinical LLM technology.
Ashish Koshy, Group Chief Operating Officer of M42, expressed enthusiasm about the release, stating, “We are thrilled to introduce the second iteration of Med42, marking a significant milestone in our mission to ethically leverage AI for the betterment of patients and healthcare professionals. This release brings us closer to our vision of a versatile multimodal AI platform that integrates expert models across various health data modalities. By fostering collaboration among these models and extracting insights from diverse data sources, we aim to address complex healthcare challenges that demand a multimodal approach. Our ongoing efforts in advancing the Med42 platform underscore our unwavering commitment to realizing this vision.”
Koshy emphasized the importance of real-world performance over benchmarks like the USMLE, highlighting ongoing research studies conducted by interdisciplinary teams of AI experts and physician-scientists to comprehensively evaluate and refine Med42’s capabilities in clinical settings.
Med42 exemplifies the potential of automation in healthcare by simplifying documentation processes, easing administrative burdens, and supporting healthcare professionals. The model facilitates the retrieval of pertinent insights, sifts through patient histories, and condenses records. Furthermore, Med42 is undergoing evaluation for its capacity to formulate tailored treatment strategies by analyzing patients’ medical histories and determining optimal courses of action. Additionally, the model shows promise in serving as a clinical decision support tool for physicians, aiding pharmacists in accurate dosage decisions, and assisting researchers in efficiently reviewing treatment literature.
Accessible via the HuggingFace platform, Med42 enables extensive testing, contribution, review, and assessment by the global scientific and developer community. By democratizing access to Med42, M42 aims to catalyze global innovation in AI, empowering experts worldwide to refine and broaden the model’s applications across diverse medical domains, thereby propelling advancements in patient care and medical research.
Conclusion:
The unveiling of Med42’s enhanced AI model at Abu Dhabi Global Healthcare Week signifies a significant advancement in healthcare AI technology. With its impressive performance metrics and accessibility through the HuggingFace platform, Med42 is poised to disrupt the market, empowering healthcare professionals worldwide to streamline processes, improve patient care, and drive innovation in medical research.
