New offering arms life sciences with novel clinical AI to democratize analytics across the organization and drive insights at an unprecedented speed
Mendel, a clinical artificial intelligence technology leader, announced the launch of Hypercube, an AI-copilot that enables life sciences and healthcare enterprises to interrogate their troves of patient data in everyday language through a chat-like experience. Using Hypercube, organizations can now command the full spectrum of their data, structured and unstructured alike, to deliver blazing-fast insights and answer previously unanswerable questions.
The current approach of leveraging AI in health data relies solely on machine learning techniques like neural networks and large language models (LLMs). However, this approach is inadequate for clinical applications because of its inability to discern the nuances of clinical language and susceptibility to hallucinations (unpredictable errors or inaccuracies). Hypercube fills this void and can clinically reason through a proprietary hybrid approach that combines large language modeling with symbolic AI. The solution opens complex cohort analysis to the business user through plain English questions, while providing advanced users like data scientists query visualization and programmatic access to run analyses – enabling a broad variety of use cases across the organization.
Also Read: Klick Health Launches First ChatGPT Plugin for Life Sciences Industry in US
“After six years of R&D, Hypercube is not a step forward for real-world data applications, it is a leap to RWD 2.0,” said Dr. Karim Galil, co-founder and CEO of Mendel. “We are freeing life sciences from both the brute force measures previously needed to unlock unstructured data, and the limitations of standard data querying methods like SaaS and structured query language (SQL).”
To test Hypercube’s technology against other LLMs, including GPT-4 and Llama2-7b, Mendel has partnered with a team of oncologists at the University of Pennsylvania School of Medicine. “AI and large language models have the potential to identify patients who are eligible for clinical trials. The problem is that we don’t know which LLMs do best. The results here suggest that domain-specific models can outperform generic LLMs at a scalable cost,” said Dr. Ezekiel Emanuel, Vice Provost for Global Initiatives, University of Pennsylvania. “With further validation, Hypercube could become a critical step in making tasks like our clinical trial screening process and answering questions from patient data much more efficient.”
“Mendel is ahead of the game – they recruited a world-class team to develop a novel AI technology, six years before it became a trend,” said Billy Deitch, Partner at Oak HC/FT, a lead investor in Mendel. “We are thrilled to see Mendel’s efforts come to fruition with this launch and are excited for the impact Hypercube will have on healthcare at large.”
SOURCE: BusinessWire
