Back

An improved framework for grey-box identification of biological processes

Jagadeesan, P.; Raman, K.; Tangirala, A. K.

2025-02-20 systems biology
10.1101/2025.02.14.638382 bioRxiv
Show abstract

Developing models from observations is at the heart of empirical science. Grey-box Modeling combines the insights gained from the results obtained from first principles with observational data. When the model turns out to be unsatisfactory, the goodness of such grey-box models in terms of predictability and parameter estimates largely depends on either modifying the model structure obtained from the first principles or conducting new experiments. Unfortunately, in the context of biological models, where the model structures are usually nonlinear ODEs with a large number of states and parameters along with sparse and noisy experimental data, traditional identification protocols have to go through several iterations to identify the source of the issue. Even after multiple iterations, they may still arrive at sub-optimal solutions. In this work, we propose an improved framework with a new set of tools to resolve this issue unambiguously with a minimum round of iterations.

Matching journals

The top 8 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.