11-04, 11:30–12:00 (Asia/Jerusalem), Green Track
In the ever-evolving landscape of healthcare, doctors face an ongoing challenge: how to access vital medical information about their patients buried deep within databases. Traditional methods have proven time-consuming and often fall short of providing the comprehensive answers doctors need. But what if I told you that AI, SQL, and GraphQL have walked into fertility clinics, offering a groundbreaking solution?
In my presentation I explore the innovative use of Large Language Models (LLMs) in medical feature development. I introduce a novel approach that leverages LLMs to translate doctors' intricate questions into SQL and GraphQL queries, enabling prompt and accurate retrieval of patient data. The result? A revolution in the way doctors access and utilize critical information to make informed decisions.
Join me at the development table as we uncover the objectives behind crafting the "chatting with my medical database" feature. Together, we'll unravel how LLM-based Python chains became integral to this feature and how GraphQL emerged as the superhero, leaving SQL in the dust. We will dive deep into the key development considerations that influenced our choices, encompassing security, flexibility to handle diverse inputs, and reliability in providing doctors with answers to their questions.
Session outline:
-
Main challenges of doctors while working with patients' medical data in databases.
-
Goals of “chatting with my data” feature for fertility clinics.
-
Going with the audience step by step on the architecture flow of this kind of feature - showing each step in the flow, what is the input and output of each step, what are our main concerns in every step
-
Sharing with the audience the solutions we discussed in our team to achieve the feature goals, and that are adequate to the feature flow. Presenting the pros and cons of each solution with respect to: security, flexibility in input and output, development time and cost, explainability and reliability of answers given back to the doctor. In this part I share a Python code that includes LLM-based chains for each of the solutions. I show how the code enables us to seamlessly test each of the solutions and compare them. The three solutions are: LLM+SQL, LLM+GraphQL and LLM+RestAPI.
-
Discussing why we chose the “GraphQL solution”, how we implement it and present additional challenges in this development section (leaving some of them not solved yet… :))
-
Summary
Dr. Shirli Di-Castro Shashua is a professional in machine learning and AI technologies. She earned her PhD from the Technion in the Faculty of Electrical and Computer Engineering, specializing in reinforcement learning, following her BSc in Biomedical Engineering from Ben Gurion University. Currently, Shirli holds the role of Senior Data Scientist at Embie, where she develops innovative solutions to fertility clinics using advanced generative AI capabilities.