Digital Science are Gold Sponsors of this year’s event. Join us for our presentations in the ‘Data Integration and FAIR’, and ‘AI in Drug Discovery and Development’ tracks, and visit us in the exhibit hall!
Data Integration and FAIR track
The Foundation for Trustworthy AI: From FAIR Data Principles to Enterprise-Ready AI in Pharma
Speaker:
Mark Hahnel, VP of Open Research, Digital Science
Abstract:
The pharmaceutical industry faces a “scientific content crisis” where over half of professionals cite poor data quality and non-FAIR data as major barriers to AI implementation. A significant governance gap exists: 27% of professionals cannot trace the evidence chain powering their AI, and 38% have unclear copyright policies, leading to legal and regulatory risk (e.g., under the EU AI Act). This session provides an actionable blueprint to solve this crisis. We demonstrate how FAIR principles offer the necessary policy and metadata foundation, and how knowledge graphs serve as the operational layer to make that foundation actionable. Attendees will leave with five strategic recommendations, including the need to treat data infrastructure as a strategic investment, embrace neuro-symbolic architectures (like GraphRAG) to eliminate hallucinations, and build for machine actionability from day one to ensure AI systems are not only powerful but audit-ready and trustworthy.
AI in Drug Discovery and Development
A Practical Path to AI-Grounded Drug Discovery with Knowledge Graphs
Speaker:
Mark Hahnel, VP of Open Research, Digital Science
Abstract:
Large Language Models (LLMs) often suffer from a critical failure mode in drug discovery: hallucination and a lack of semantic intelligence. This presentation moves beyond theoretical promise to showcase practical, knowledge-grounded AI applications that deliver measurable value across the drug discovery pipeline today. We will examine proven use cases, including target identification, drug repurposing, polypharmacy prediction, and adverse drug reaction prediction, where the systematic integration and traversal of interconnected biological knowledge via knowledge graphs have surfaced non-obvious, clinically relevant connections. The core solution is the neuro-symbolic approach, specifically Graph-based Retrieval Augmented Generation (GraphRAG), which uses the knowledge graph as a verifiable grounding layer to transform probabilistic LLMs into reliable reasoning systems. Learn how to connect your internal knowledge with the broader research intelligence ecosystem using persistent identifiers (ORCID, DOI, ROR) to accelerate your pipeline.
