| Speaker: Ludvig Hult, Uppsala University
About the Webinar
All of society is in rapid change as AI becomes more prolific. Generative models for text and images have disrupted sectors like marketing, publishing and software. Healthcare seems to be lagging behind, largely because many obstacles for AI in healthcare are poorly addressed by well known commercial providers like ChatGPT and Gemini.
This webinar highlights the unique challenges in healthcare, such as data security and model safety. Topics covered include applications of AI in healthcare (transcription, summarization), and key challenges (data privacy, algorithmic fairness).
The EmergAI project at Uppsala University explores other types of AI to face these challenges. One of those models, the healthcare event foundation model is introduced, which can be thought of as a specialized LLM. Other approaches such as use of multimodal models in this context are also presented.
Who is the webinar for
This talk is ideal for healthcare professionals, data scientists, researchers in life sciences, and those in related fields. Additionally, it offers inspiration to anyone with a general interest in AI and machine learning, highlighting the importance of developing specialized models to ensure safe and reliable systems.
Key takeaways for participants:
- Explore how AI can be employed in transcription, summarization, and image recognition within healthcare settings.
- Understand the key challenges facing AI implementation in healthcare, including data limitations, security concerns, safety issues, fairness considerations, and compliance with Medical Device Regulations (MDR).
- Discover EmergAI’s innovative approach through the healthcare event foundation model, multimodal models and other specialized models, including applications in ECG analysis and self-reported data.
- Understand the importance of tailored AI models like sequence models in healthcare and how they compare to generic large language models (LLMs).