ncibtep@nih.gov

Bioinformatics Training and Education Program

Generative AI, it's all about Embeddings

Generative AI, it's all about Embeddings

 When: Jul. 9th, 2025 11:00 am - 12:00 pm

Learning Level: Any

To Know

Where:
Building 35A, Room 610 (NIH Bethesda Campus)
Organizer:
CDSL
Presented By:
Jonas Almeida (NCI DCEG)

About this Class

From image analysis to federated learning and multimodal modeling, generative AI can be used to study observational succession across scales and data types. At DCEG's Data Science and Engineering Research Group (DSERG), we are developing user-facing, FAIR, privacy-preserving infrastructure for cancer research based on numerical embedding shared across, and between data types. This presentation is configured as a show-and-tell activity with live applications ranging from digital pathology to real-time epidemiology trackers. Particular focus will be placed on showing where and how does genAI modeling expose a numeric representation of the underlying latent space.

 CDr. Jonas Almeida leads a multidisciplinary program of data science and engineering research that combines systems biology, computational statistics, and software engineering for biomedical applications. The primary focus of his research is to accelerate the investigation of epidemiologic and genetic causes of cancer by developing innovative digital methods that advance the computational research infrastructure for precision prevention.