Abstract For years, mathematical models have been successfully used to explain biological, chemical, or physical relationships. The enormous advances in artificial intelligence in understanding context-specific content can support humans in generating interaction hypotheses for modeling, enabling models to be generated and modified far more quickly. Our pipeline, DataXflowGen, uses GenAI to create models based on information such as research publications available on the internet. This approach reduces the need for specialized knowledge, as GenAI identifies relevant research papers online and extracts the relevant information. GenAI can build a model that fits the data to test its hypothesis. As a result, humans only need to conduct an in-depth analysis of those model components that do not align with the data, saving time when building a suitable regulatory network based on existing knowledge. In summary, GenAI is used to generate a human-interpretable model as a hypothesis, allowing an understanding of therapy suggestions and an explanation of actions based on a model validated with specific data, avoiding black-box AI decisions. This constitutes an approach to explainable AI that supports humans in analyzing complex relationships.
Crouch et al. (Sat,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: