- Writer Inc. has launched two specialized large language models: Palmyra-Med-70b for healthcare and Palmyra-Fin-70b for financial services.
- Palmyra-Med-70b is noted for its high accuracy of 85.9% in zero-shot tests, surpassing Google’s Med-PaLM-2.
- Palmyra-Fin-70b is the only financial model available on Nvidia’s NIM inference microservice, offering capabilities in financial forecasting and risk evaluation.
- Both models are available on Nvidia’s NIM interference microservice, Baseten, Hugging Face, and Writer’s platform.
- Domain expertise was crucial in the development of these models, with industry experts assisting in data categorization and model evaluation.
- Palmyra-Med is used by major healthcare organizations, including Vizient Inc., CirrusMD Inc., and Medisolv Inc.
- Palmyra-Fin assists financial analysts with in-depth analysis of Form 10-K statements.
- Both models can be accessed by developers through the Writer Framework, an open-source platform for AI app development.
- Writer Inc. plans to update these models biannually and is exploring multimodal capabilities for future enhancements.
Main AI News:
Writer Inc., a leader in full-stack generative AI platforms, has unveiled two new large language models tailored for the healthcare and financial services sectors. The company is now offering Palmyra-Med-70b and Palmyra-Fin-70b, specialized LLMs designed for medical and financial applications, on Nvidia Corp.’s NIM interference microservice, Baseten, Hugging Face, and Writer’s own platform.
Palmyra-Med-70b, the newest addition to Writer’s medical AI model lineup, has achieved the highest benchmark accuracy in the market, with an impressive 85.9% accuracy in zero-shot tests. This surpasses Google LLC’s Med-PaLM-2, which achieved 84% accuracy after multiple attempts. The model’s accuracy makes it a crucial tool for medical legal reviews, a complex task requiring extensive research and precision, especially in a highly regulated industry.
Palmyra-Fin-70b, the financial counterpart, is the sole model available through Nvidia’s NIM inference microservice for finance. It offers capabilities in financial forecasting, investment analysis, risk evaluation, and asset allocation. This model provides financial analysts with detailed insights and summaries from complex documents like Form 10-K statements, showcasing its proficiency in handling extensive financial data.
Waseem Alshikh, co-founder and CTO of Writer, highlighted the role of domain experts in developing these models, emphasizing their contribution to the data categorization and model evaluation processes. Alshikh noted that Palmyra-Med’s application in medical legal reviews exemplifies its effectiveness, providing detailed evaluations and error analyses of pharmaceutical press releases.
Palmyra-Med is already in use by major healthcare organizations, including Fortune 50 companies and firms such as Vizient Inc., CirrusMD Inc., and Medisolv Inc., aiming to enhance patient outcomes through advanced generative AI. Similarly, Palmyra-Fin has demonstrated its value in financial applications, offering comprehensive analyses of corporate strategies.
Both models are accessible to developers via the Writer Framework, an open-source platform enabling the creation of AI applications with minimal coding. Writer Inc. plans to continuously update these models every six months and is exploring the addition of multimodal capabilities, including image, video, and audio processing.
Conclusion:
Writer Inc.’s release of Palmyra-Med-70b and Palmyra-Fin-70b represents a significant advancement in specialized AI tools for the healthcare and financial sectors. These models set a new standard for accuracy and capability, addressing complex industry-specific needs with high precision. The introduction of these advanced AI models not only enhances operational efficiency in these industries but also demonstrates Writer Inc.’s leadership in the evolving AI landscape. Their commitment to continual updates and exploration of multimodal functionalities positions them favorably in a competitive market, likely influencing other players to enhance their own AI offerings.
