<aside>

Introduction

With the Axiomatic tools, researchers can automatically turn plot from scientific articles into physics-grounded, predictive models. The workflow begins with automated data extraction from the article and extraction of data from figure panels, followed by a physics-based model proposal, fitting the model to the observed data, and generating a validation report that includes both statistical assessment and model evaluation. Beyond numerical fitting, the system provides a mathematical derivation from first principles, a reasoning graph that makes the derivation transparent, and explicit checks of constraints and assumptions. This end-to-end process enables scientists to recover parameters not directly reported in publications, ensure that models are physically consistent and predictive, and support reproducibility, cross-validation, and downstream modeling across a wide range of experimental studies.

</aside>

Main goals

• The Axiomatic MCP tools assist in automatically finds a physics-grounded and predictive model based on one the figures in a paper on bilayer graphene.
  • Automated data extraction from the figure panels
  • Automated physics-based model proposal
  • Automated model-fitting with the data
  • Automated Verification and validation, including validation report.
  • Generation of a report with statistical assessment and model evaluation
• Full mathematical derivation of the model from first principles
  • Reasoning graph of the derivation steps
  • Explicit validation of the constraints/assumptions

Axiomatic Tools used

• Ax Equations Explorer
• Ax Plot To Data
• Ax Documents Parser
• Ax Model Fitter

The scientific paper

Yin, J., Tan, C., Barcons-Ruiz, D., Torre, I., Watanabe, K., Taniguchi, T., Song, J.C., Hone, J. and Koppens, F.H., 2022.

Tunable and giant valley-selective Hall effect in gapped bilayer graphene. Science, 375(6587), pp.1398-1402.

User Workflow