Google's AMIE ( Articulate Medical Intelligence Explorer) a research AI which utilizes LLM model and optimized for conversational diagnostics .
pic source: Google Research blog
they used trained AMIE using different dimensions which reflected real world clinical conversations from both specialist doctors and patients point of view . AMIE research Ai is trained to communicate with empathy and respecting privacy/trust of patients to scale AMIE through diverse medical conditions, scenarios etc.
then designed crossover study of text-based consultations with validated patient actors interacting either with board-certified primary care physicians (PCPs) or the AI system optimized for diagnostic dialogue. They set up their consultations in the style of an objective structured clinical examination (OSCE), a practical assessment commonly used in the real world to examine clinicians’ skills and competencies in a standardized and objective way.
they developed simulated self play diagnostic conversation environment with automatic feedback system to improve and accelerate its learning process.
pic source : Google Research blog
The above picture shows AMIE research AI's diagnostic conversation with an trained patient actor
It analyzes the context of the information from questions asked during clinical conversation , then generates response later refines the response with empathy and nurtures relationship to provide information precisely.
Additionally, in order to enhance AMIE's diagnostic precision and conversation quality, they implemented an inference time chain-of-reasoning technique.
Finally, they simulate consultations with trained actors to test AMIE prospectively in real-world scenarios of multi-turn dialogue.
There is Limitations - it only utilized text based conversation to scale up LLM patients interactions but doesn't include real life clinical interactions.
There are many important limitations to be addressed, including experimental performance under real-world constraints and dedicated exploration of such important topics as health equity and fairness, privacy, robustness, and many more, to ensure the safety and reliability of the technology.
In future AMIE Conversational Diagnostic AI could help as aid to Doctors/clinicians
Posted by Alan Karthikesalingam and Vivek Natarajan, Research Leads, Google Research The physician-patient conversation is a cornerstone of medicin...
blog.research.google
www.cnbc.com
