Mendel.ai

Mendel is proud to announce its AI-driven Hallucination Detection research led by University of Massachusetts Amherst | Research with its paper accepted to the prestigious AI conference - ACM SIGKDD & Annual KDD Conference. As AI continues to integrate into healthcare, addressing hallucinations in LLM outputs will be vital. Our research developed a robust framework to automatically detect and categorize hallucinations into five types. The research tested OpenAI's GPT-4o and Meta's Llama-3 in a pilot study of 100 summaries. Furthermore, our product Hypercube was used as an option to review the summaries, detect hallucinations, and recommend changes to improve the accuracy. Thank you to our authors: Prathiksha Rumale Vishwanath, Simran Tiwari, Tejas Ganesh Naik, Sahil Gupta, Dung Thai, Wenlong Zhao, Sunjae Kwon, Victor Ardulov, Karim Tarabishy, Andrew McCallum, and Wael Salloum, Ph.D. For more information, please see our press release here: https://lnkd.in/gthYWx5b To try Hypercube, sign-up here: https://www.mendel.ai/#CTA

Karim Galil, M.D. participated in the inaugural Google Cloud Cancer AI Symposium. This was an important venue to bring together the top AI leaders in oncology. Thank you for inviting us to speak on a panel with William H. Morris, MD, MBA, Sanjay Juneja, M.D. from Tempus AI, Robert Reid from Ontada and David Fajgenbaum, MD, MBA, MSc from Every Cure. Information Retrieval Problem was a key theme. "Every 72 days the medical literature doubles and our 40-watt light bulb brain does [not]." - William H. Morris, MD, MBA "80% of the upfront work for an oncologist is... retrieval and processing... what did the pathology report say and hopefully the pathologist was correct... let me order the molecular, what does that mean, where were the nodes positive, that's what the stage is, let me find out the algorithm, what's best and then if I'm really good what medications... what polymorphisms, what pharmacogenomics, but it's all glorified retrieval at first." - Sanjay Juneja, M.D. "We have a lot of systems that are able to index the data, but not necessarily answer questions like a human would answer the question. Is that patient a fit for NCT for this clinical trial or not and that's the problem that Mendel.ai solves." - Karim Galil, M.D. Karim also presented a lightning talk on guarding against hallucinations and unleashing patient level reasoning. Look out for the video we will be publishing in the next week!

Mendel AI Hypercube product is the leading product for clinical researchers. Hypercube has proven results with 30% faster chart review for nurses. Thank you to the Google Cloud and Allytics team for creating this video. Special shoutout to Kashaf Mazhar and former GCP Riju Khetarpal for inviting us to the Springboard program. We are also presenting at Google AI Cancer Conference. Register here: https://lnkd.in/eQT5hiGu Video on Youtube: https://lnkd.in/g3iNhKYt

🤔 "𝐓𝐡𝐚𝐭’𝐬 𝐚 𝐬𝐞𝐫𝐢𝐨𝐮𝐬 𝐡𝐚𝐥𝐥𝐮𝐜𝐢𝐧𝐚𝐭𝐢𝐨𝐧." In AI for healthcare, accuracy is everything. When medical systems misinterpret critical patient data like allergies, the impact on care can be dangerous. As Wael Salloum, Ph.D., Co-Founder and Chief Science Officer at Mendel.ai, puts it, having to double-check AI output defeats the purpose of automation and summarization. Ravi B. Parikh, one of Emory University AI Leaders and advisor to Mendel, highlights LLMs making a particular mistake with prostate cancer. 🧠 Learn more about the challenges and advancements in AI for medicine in this insightful article "AI in Medicine: Are Large Language Models Ready for the Exam Room?" authored by Donavyn Coffey, Medscape, featuring Wael’s expert perspective. 👉 Read the full post here: https://lnkd.in/d8vWY2in #AIinHealthcare #ClinicalAI #MendelAI #ResponsibleAI #MedTech #HealthTech #ClinicalReasoning

🚀 Google Cloud’s release of Vertex AI Search for Healthcare highlights the growing impact of generative AI in transforming healthcare. This aligns closely with what we’re building at Mendel AI—AI-powered solutions that reduce administrative burdens and enhance clinical decision-making through reliable, grounded insights. Thank you, Susanna Vogel, for shedding light on our peer-reviewed research on AI Hallucinations! At Mendel, we’re proud to be at the forefront of addressing challenges like AI hallucinations through rigorous research, ensuring that generative AI becomes a tool healthcare professionals can trust. 🔗 Read the full article: https://lnkd.in/gEwYWYdB #MendelAI #HealthcareTech #GenerativeAI #Innovation

🔈New Blog Post 🔈 𝐈𝐧𝐭𝐫𝐨𝐝𝐮𝐜𝐭𝐢𝐨𝐧 𝐭𝐨 𝐇𝐲𝐩𝐞𝐫𝐜𝐮𝐛𝐞’𝐬 𝐎𝐧𝐭𝐨𝐥𝐨𝐠𝐲 𝐚𝐧𝐝 𝐑𝐞𝐚𝐬𝐨𝐧𝐢𝐧𝐠 𝐄𝐧𝐠𝐢𝐧𝐞 How do we ensure AI helps clinicians make better decisions? Mendel’s Hypercube platform offers a cutting-edge solution, diving deep into advanced ontology and clinical reasoning to unlock insights from both structured and unstructured medical data. In our latest blog, Jose Mena, Sr. Director of Bioinformatics at Mendel.ai, explains how Hypercube is transforming clinical research and diagnostics with precision and speed. 🔍 Key highlights from the blog: Ontology-driven clinical reasoning: Hypercube leverages AI to understand and reason over clinical data, providing actionable insights. Faster patient cohort identification: Identify patient groups more efficiently, even for intricate medical conditions. Reduced query errors: Hypercube ensures error rates under 5%, delivering trustworthy, high-precision results. Seamless data integration: Effortlessly analyze both structured and unstructured clinical data to unlock deeper insights. 💡 “Hypercube empowers professionals to ask complex medical questions and get fast, reliable answers with clinical context,” says Jose Mena. Read the full post here: https://lnkd.in/eRSM-Sjx #ClinicalAI #Ontology #ClinicalReasoning #DataScience #MedicalResearch #HealthcareInnovation #AI

Why does AI in healthcare end of being so expensive and difficult to scale? HealthLeaders published a story that highlights a core symptom and cause of the challenges AI faces in healthcare when it comes to cost and scalability, tied to a challenge we’re very focused on at Mendel—AI hallucinations.   “It’s actually really difficult to respond to these hallucinations by modifying the algorithm,” said @Ravi B. Parikh of Emory University. “You might be able to fine-tune and sort of say ‘Hey, we want to avoid this type of output’ and there's certain reward-based mechanisms to do that, but usually that's not in the health system’s control. Usually it's a developer who's having to respond to feedback that they're getting from the health system and then doing some things behind the hood that we don't honestly know about.”   At Mendel, we’ve been focused on AI that is specifically built for clinical complexity, eliminating hallucinations and opening up opportunities to scale and respond in the ways physicians, like Dr. Parikh, require.   You can read the full article authored by Eric Wicklund here: https://lnkd.in/eHgqT_k5

We are excited to announce our partnership with AWS. Our Hypercube Platform is now powered by AWS and available on the AWS Marketplace. AWS' commitment to supporting generative AI startups has been impressive, and we look forward to accelerating our partnership to transform generative AI for life sciences. Big shoutout to our AWS partner team and many supporters: Sydney Falk Chad Coder Yin He Tina Sayadi Sherry Karamdashti Jared Saul, MD Amit S. Matt Pearson Praveen Haridas Damon Cronkey Bret Borota, FACHE George Softas Caleb Wilkinson Christine C. Link to press release: https://lnkd.in/gmS3C2ax

Professor Mark Braunstein, a leader in health informatics, published a blogpost on AI for Clinical Trials. He features Mendel AI's neuro-symbolic technology for automatic cohort retrieval (ACR) as an approach for clinical trial matching. ACR authors: Dung Thai Victor Ardulov Jose Mena Simran Tiwari Gleb Erofeev, MCS Ramy Eskander, Ph.D. Karim Tarabishy Ravi B. Parikh Wael Salloum, Ph.D.