Lavita

?? We’re thrilled to announce that Lavita AI’s research on a new, publicly available Medical QA Benchmark has been accepted to AIM-FM: Advancements in Medical Foundation Models at #NeurIPS 2024! ??? We'll be in Vancouver this December to share our work. ???? ?? Preprint: https://lnkd.in/gDP8M6wJ ?? Code and Data: https://lnkd.in/g5AyKM9w ?? Medium post: https://lnkd.in/gvHSwhYB ?? Lavita provides a DeSci Platform for producing high-quality research? Our benchmark leverages real-world consumer questions collected directly via Lavita Medical AI Assist (https://assist.lavita.ai/) and features open-access annotations from medical doctors. ??????????? This effort fosters transparency, collaboration, and progress in understanding the strengths and limitations of foundation models in healthcare. ? As Generative AI transforms healthcare, rigorous evaluation is essential—especially for consumer-facing applications where precision matters and errors can carry significant risks. At Lavita AI, we view this benchmark as a step forward in our groundbreaking initiative: the Medical Evaluation Sphere. ?? This pioneering DeSci platform advances medical AI evaluations across diverse tasks, scaling efforts to meet real-world needs. ?? Top AI Institutes and Researchers as Partners? This work builds on our collaboration with the Dartmouth Center for Precision Health and Artificial Intelligence (CPHAI), co-led by our advisor, Prof. Saeed Hassanpour, and an incredible team of medical doctors at Dartmouth Hitchcock Medical Center. ???? Their insights have been instrumental in pushing the boundaries of medical AI evaluation. If you’re as passionate about advancing AI in healthcare as we are—or if you’re a medical professional interested in contributing to future projects—let’s connect! ?? We’d love to meet during #NeurIPS in Vancouver. Drop us a message! ??? #AI #NeurIPS2024

??? Mark your calendar! ? Join us on Friday, Nov 22nd at 1pm Pacific Time for a Lavita Insider Update ?? —a 30-minute live webinar where our team shares exciting updates on what we’ve been working on and what’s next for Lavita AI, including a sneak peek at our Medical QA Benchmark! Don’t miss the latest from Lavita AI! ?? ?? Webinar link: https://lnkd.in/g96AzQbe

Join us for Lavita AI's first AMA, ft. CEO Ali Farahanchi and AI Lead Pedram Hosseini. We’ll discuss our latest platform developments, including Medical Evaluation Sphere and EHR integration into the Lavita Platform. Submit your questions below and tune in on July 31 1pm PT for the livestream at https://Theta.tv! Don't?miss?out! ?? Medical Evaluation Sphere: https://lnkd.in/gsEsqKnn

We are excited to share that Lavita AI has joined the Microsoft AI Cloud Partner Program. This partner program will enhance our AI capabilities in healthcare, providing us with advanced tools, training, and support to drive innovation. We'll have access to Azure credits, cloud services, and developer tools like Visual Studio and GitHub, helping us reach new markets more efficiently. ?? https://lnkd.in/gnS2WARz Stay tuned for more updates as we continue our journey!

?? We've just launched Lavita AI's Medical Evaluation Sphere, an initiative for the evaluation of foundation models across medical tasks. ?? ?? Link: https://lnkd.in/gsEsqKnn Our goal is to create a collaborative space for conducting evaluations of AI models on medical tasks at scale by engaging both medical experts and general users. Truly excited about this work. We’re planning to release all votes publicly and we invite everyone to join this effort. Read more here: https://lnkd.in/gSm9Ae9h We have a list of dedicated and serverless endpoints to serve our models on the Medical Evaluation Sphere. Our current pool includes a mix of open(-source) models for the medical and biomedical domains, as well as commercial models: ? AlpaCare-13B (https://lnkd.in/guG8Brwa) ? BioMistral-7B-DARE (https://lnkd.in/eZPfUk2e, by Mistral AI) ? Mixtral-8x7B-Instruct-v0.1 (by Mistral AI) ? Meta-Llama-3-70B-Instruct (by AI at Meta) More models, modalities, and tasks coming soon! Stay tuned as we gradually release the results on our leaderboard. Also, shoutout to our friends at https://lmsys.org/ for inspiring this initiative and their great work on Chatbot Arena! #MedicalAI #GenerativeAI

Thrilled to announce that Lavita AI won the Health & Wellness track at the Samsung Next 2024 Generative AI Hackathon in NYC! It was my pleasure to represent our team in the competition. Lavita was selected as the winner by a panel of expert technology investors and VCs from among more than 200 applicants. A huge thank you to Samsung Next for organizing such an incredible event. It was a fantastic opportunity to meet and connect with so many passionate builders and investors who share our belief in the power of AI to address real-world challenges. Grateful for this experience and excited about the future of AI in healthcare!