Presenting our work on Large Language Models in ophthalmology
I recently attended the Association for Research in Vision and Ophthalmology (ARVO) Annual Meeting in New Orleans, to present some of Ufonia’s work exploring the use of Language Model Models (LLMs) in healthcare.
For those who may not be familiar, ARVO is the world’s largest annual vision sciences research conference. It brings together leading experts in ophthalmology, optometry, and everything from bench to bedside, to share the latest research and advancements in eye health.
It’s an incredible forum to connect with leading experts in the field and share our enthusiasm for exploring the possibilities of advanced technology in healthcare. I was fortunate to spend time with numerous clinicians and researchers working on cutting edge clinical-AI applications.
The work I presented focused on how LLMs could potentially address patient queries following routine cataract surgery, and adapted a human evaluation framework from Singhal et al at Google with some real-world patient scenarios facing Dora, our autonomous telemedicine platform.
Compared with my last ARVO experience in Vancouver in 2019, it feels like there has been an explosion of work in AI for ophthalmology. We heard from world-class speakers about mind-blowing research, demonstrating AI’s ability to do things like predict blood biomarkers from a simple external eye photograph, or algorithms that predict the risk of dementia’s and heart disease from a retinal scan.
One of my favourite talks is a keynote session from Professor Tien Yin Wong. Prof Wong is a Singaporean ophthalmologist who played a leading role in first setting up the Singaporean national diabetic retinopathy screening programme, and then being instrumental in also one of the world-first national deployments of an AI system to create a more sustainable and effective screening programme. The success story spanning decades of work and seeing how AI has the potential to make an outsized, real-world impact at the level of entire nations was incredible.
Beyond the sessions, I’d also be remiss not to mention some of the city’s charms! From it’s hearthy southern food (I loved the gumbo and beignets), to the streets overflowing with jazz and blues. I had a fantastic time making many new friends and catching up with old connections, and am even more excited about our work in applying AI to improving healthcare!