Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

AKPIK: Arbeitskreis Physik, moderne Informationstechnologie und Künstliche Intelligenz

AKPIK 5: Poster

AKPIK 5.5: Poster

Donnerstag, 12. März 2026, 15:00–16:30, P5

Supporting Physical and Computational Biology with AI-Powered Multi-agent Model Generation — •Prerana Chandratre, Anjali Sharma, and Justin Bürger — TUD Dresden University of Technology, Dresden, Germany

Many open biological questions, from human embryogenesis to complex diseases like cancer, cannot be solved by experiments alone but require integration with biophysical and computational modeling. To address this, our group developed the software Morpheus (https://morpheus.gitlab.io/) which has become a widely used open-source platform (Starruß et al., 2014). Morpheus is based on a declarative modeling language, MorpheusML, and such models together with their biological context are collected in the MorpheusML model repository. Yet, creating MorpheusML models remains a barrier, especially for wet-lab researchers and students without programming expertise. To resolve this bottleneck, we are developing a model generation tool that uses a multi-agent workflow based on large-language models. Planned enhancements include a simulation-in-the-loop architecture enabling iterative, agent-driven model refinement and validation, expansion of the training dataset with model-text pairs from the MorpheusML model repository, and integration of automated validation and benchmarking metrics. The outcome will be the Morpheus.AI modeling assistant that enables robust generation of valid MorpheusML models from textual input sources, including PDF manuscripts. These AI-generated models will empower both research and education in physical and computational biology.

Keywords: AI agent; Large Language Models (LLM); Modeling and Simulation; Tissue Morphogenesis; Biophysics