HARMAN announced that its Digital Transformation Solutions business unit has launched what it called the ground-breaking Healthcare Private Language Model (LLM).
Known as HealthGPT, the new technology tool represents a leap in healthcare solutions by leveraging generative AI to empower healthcare professionals, researchers, and institutions through advanced patient care, medical research, and decision-making.
Health GPT brings forward all the benefits of a LLM like natural language interaction and learnability, with a strong knowledge base established from a variety of healthcare data sets. The model, built using the key principles of responsible AI has additionally been tested for accuracy and hallucinations using HARMAN’s own automated LLM testing framework and the results were further validated by the healthcare subject matter experts.
HARMAN HealthGPT is a private LLM, providing enterprises with more control over roadmap, privacy, compliance and security issues at optimized cost.
Key advantages of working with HARMAN HealthGPT include:
- Enhanced Clinical Insights: Provides real-time, context-aware clinical insights, aiding in decision making.
- End-to-end LLM Fine tuning framework: A comprehensive framework for fine-tuning Language Model (LLM) to achieve optimal performance.
- Data generation framework: A framework for creating high quality customized datasets to further fine-tune Language Model (LLM) for enhanced performance.
- Automated LLM evaluation framework: Completely automated solution to validate fine-tuned LLMs for quality of outputs, factual correctness, hallucinations, and toxicity.
- Cost Optimization: Advanced deployment techniques using quantization to significantly reduce model size and thereby processing costs by up to one-tenth.
- Drug Discovery and Research: Accelerate drug discovery and development by extracting valuable insights from clinical trial data.
“The business value of generative AI cannot be overstated,” declares Nick Parrotta, President – Digital Transformation Solutions & chief digital and information officer at HARMAN. “Organizations that scale and implement swiftly will see significant competitive advantages, productivity gains and more – but only if they can unlock their data and move from general purpose applications to more specialized, domain-specific applications,” he asserts. “At HARMAN, it is our mission to help our clients navigate this hurdle and create a competitive advantage long-term.”
According to him, HARMAN HealthGPT is an example of the company’s capabilities that will help organizations utilize specific industry trained models to better solve unique problems and add value for customers. “With our long-standing AI expertise and ability to develop effective LLMs, we’re equipped to help our clients move past challenging roadblocks and fully capitalize on the exciting promise of generative AI,” asserts Nick.
HARMAN is seeing strong early results of the LLM training, and its AI and ML teams are well positioned with an end-to-end, tested framework to fine-tune more private LLMs, beyond the healthcare industry, to solve similar customer problems.
