Thrissur native designs AI model to tackle challenges of rare diseases

Kochi: As artificial intelligence is making inroads into various fields, researchers have also begun to parallelly develop AI data for professionals in different areas of specialisation.Karthik Soman, a native of Thrissur and a research scientist at the University of California (UCSF), has developed a ground-breaking AI model ‘Zebra-Llama,' designed to tackle the challenges of rare diseases."Since I work in the healthcare sector, we looked at the challenges faced by the medical fraternity in finding accurate data on rare diseases. Often the data that springs up has both general as well as specific data on rare diseases. So, our innovation was on developing a model that contains all available information on Ehlers-Danlos Syndrome (EDS). EDS is a group of inherited connective tissue disorders characterised by joint hypermobility, skin hyperextensibility, and tissue fragility.The model is just like ChatGPT, but it has exclusive information on EDS that will help doctors identify it in patients easily. "We presented it at the Stanford Medicine Rare Disease AI Hackathon in San Francisco, USA, recently. Later, we had the privilege of presenting it before Greg Brockman, president of OpenAI (the company behind ChatGPT)," Karthik said in a telephonic interview from the US.He and his team are now focusing on using AI to generate information on other such rare diseases, which will help the medical world.Karthik said that Zebra-Llama analyses over 4,000 EDS research papers and more than 13,000 patient experiences from Reddit and Inspire forums, providing a comprehensive view of both scientific understanding and real-world patient insights. "It also allows patients to update themselves with their condition and records newer developments," he added.By structuring the training data, Zebra-Llama learns to consider relevant context when formulating answers, mimicking the way medical professionals approach patient cases.