the health strategist
institute of research and strategy
for continuous health transformation
Joaquim Cardoso MSc.
Chief Research and Strategy Officer (CRSO),
Chief Editor and Senior Advisor
August 17, 2023
What is the message?
Microsoft Research asserts that GPT-4, an OpenAI language model, is proficient for medical tasks, potentially accelerating processes like drug development by structuring complex clinical data.
GPT-4 showcases impressive performance in medical applications, outperforming specialized systems without requiring task-specific fine-tuning, offering promise for evidence-based precision health and bridging the gap between clinical research and care.
Key takeaways:
- GPT-4’s Medical Capabilities: Microsoft Research states that GPT-4, a large language model developed by OpenAI, is proficient enough to handle medical tasks effectively. This is in contrast to specialized medical language models being developed by other companies like Google.
- Acceleration of Medical Processes: Microsoft believes that GPT-4’s capabilities can expedite medical processes, particularly in handling large amounts of unstructured medical data.
- This could lead to more efficient development of medical treatments, such as cancer drugs, where time-consuming manual processing is currently required.
- Potential for Medical Breakthroughs: GPT-4’s ability to structure patient information from complex clinical texts could significantly impact fields like cancer research.
- The model’s competence in abstracting patient data from large datasets could lead to transformative advancements similar to those seen in programming or productivity applications.
- Impressive Performance without Medical Training: Despite being trained on general Internet data and not specific medical information, GPT-4 excels at structuring clinical studies and answering medical questions. It outperforms specialized systems designed for these tasks without requiring fine-tuning.
- Language Models as Universal Annotators: Large language models (LLMs) like GPT-4 can serve as universal annotators, generating labeled examples from unstructured data to train other models.
- They could also identify cause-and-effect relationships in medical contexts.
- Medical Models: LLMs have the potential to process not only text but also medical images containing genetic, protein, and other biological data.
- Microsoft is working on LLaVA-Med, a chatbot for biomedical imaging data, and Google has unveiled Med-Palm M, a multimodal medical model.
- Precision Health Copilots: Microsoft envisions “precision health copilots,” which would provide real-time insights from large health datasets.
- These copilots could enhance decision-making for medical professionals by updating patient health statuses based on the latest evidence.
- The Future of Evidence-Based Precision Health: Microsoft Research sees generative AI, including large language models like GPT-4, as a crucial force in driving the evolution of evidence-based precision health.
- These advancements could revolutionize the medical field, connecting clinical research and care in unprecedented ways.
Infographic:
DEEP DIVE
Microsoft Research says GPT-4 is good enough for medical tasks
Companies like Google are developing language models optimized for medical purposes. Microsoft believes that GPT-4 is sufficient
The Decoder
Matthias Bastian
August 13, 2023
According to Microsoft, large language models can help speed up medical processes by, for example, structuring “large unstructured data” that currently requires time-consuming manual processing.
As an example, Microsoft cites the faster development of cancer drugs, where many clinical trials would have to be abandoned due to insufficient recruitment. Billions of dollars would be wasted in lengthy processes.
Large language models such as GPT-4 could significantly accelerate such processes by efficiently abstracting patient information from large clinical texts. The impact of language models here would be similarly transformative to that of programming or productivity applications.
GPT-4 achieves SOTA results without special medical training
Although GPT-4 was trained only on generic Internet data and not on specific medical data, it was able to structure complex clinical studies according to specified criteria. In this respect, it outperforms current systems such as Criteria2Query, even though they were developed specifically for this task.
OpenAI’s large language model could achieve expert-level performance on medical question-answer datasets such as MedQA (USMLE exam) without requiring “costly task-specific fine-tuning or intricate self-refinement”, according to the report.
Microsoft has also introduced language models such as BioGPT specifically for medical tasks, but is now making it clear that it will rely primarily on GPT-4 in the future.
GPT-4 could also structure patient data sets on a large scale, for example in cancer research. The model could act as a kind of super-organizer, enabling the use of real-world data on an unprecedented scale.
Although pretrained on general web content, GPT-4 has demonstrated impressive competence in biomedical tasks straightaway and has the potential to perform previously unseen natural language processing (NLP) tasks in the biomedical domain with exceptional accuracy.
Microsoft
Toward evidence-based precision medicine
LLMs could also serve as universal annotators, supporting the training of other models by generating labeled examples from unstructured data or finding cause-and-effect relationships.
Another application for AI in medicine would be multimodal models that can process medical images that contain genetic, protein, and other types of biological data in addition to text.
Microsoft is developing LLaVA-Med, a sort of chatbot for biomedical imaging data available to medical professionals. Google also recently unveiled Med-Palm M, a multimodal medical model that can solve medical tasks in many domains and offers a chat mode.
The ultimate goal, according to Microsoft’s research team, is “precision health copilots” that can assist anyone involved in biomedical processes. They would provide a real-time view of large amounts of health data, accelerate care and new discoveries, and ensure a closer connection between clinical research and care.
Any clinical observation could be used immediately to update the patient’s health status. This would enable physicians and caregivers to make decisions based on the latest and most comprehensive evidence.
“This vision embodies the dream of evidence-based precision health. Generative AI, including large language models, will play a pivotal role in propelling us towards this exciting and transformative future.”
About the Author
Matthias Bastian is an online journalist Matthias and the co-founder and publisher of THE DECODER. He believes that artificial intelligence will fundamentally change the relationship between humans and computers.
Originally published at https://the-decoder.com
