This publication contains three exemplary data sets generated via active learning and numerical simulations. The active learning approach used is query-by-committee. For comparison, data is also generated using classical sampling approachs. The first data set originates from a toy example that is based on an appoximated Dirac delta function, for which data was generated randomly and via query-by-committee. The second example is part of a parameter identification problem in materials modeling, for which data was generated via Latin Hypercube design, a knowledge-based approach and query-by-committee (cf. [1]). The third example is about generating artificial bcc rolling textures (cf. [2]), for which data was generated via Latin Hypercube design, query-by-committee and an extended query-by-committee approach that prevents sampling in regions out of scope.
The publication contains
- sampled Direc delta function
* 1_dirac_rd_X.txt - Inputs generated by a random sampling approach
* 1_dirac_rd_y.txt - Corresponding values of the delta function
* 1_dirac_qbc_X.txt - Inputs generated by query-by-committee
* 1_dirac_qbc_y.txt - Corresponding values of the delta function
- sampled hardeing model for parameter identification (cf. [1])
* 2_paramident_lhs_params.txt - Hardening parameters generated by Latin Hypercube design
* 2_paramident_lhs_curves.txt - Corresponding hardening curves
* 2_paramident_kbs_params.txt - Hardening parameters generated by the knowledge-based approach from [1]
* 2_paramident_kbs_curves.txt - Corresponding hardening curves
* 2_paramident_qbc_params.txt - Hardening parameters generated by query-by-committee
* 2_paramident_qbc_curves.txt - Corresponding hardening curves
- generated artificial bcc rolling textures (cf. [2])
* 3_rolling_lhs_params.txt - Parameters of the texture description model generated by Latin Hypercube design
* 3_rolling_lhs_props.txt - Corresponding properties
* 3_rolling_qbc_params.txt - Parameters of the texture description model generated by query-by-committee
* 3_rolling_qbc_props.txt - Corresponding properties
* 3_rolling_qbc_Rle5_params.txt - Parameters of the texture description model generated by query-by-committee, forced to sample in regions with R-values less than or equal to 5.0
* 3_rolling_qbc_Rle5_props.txt - Corresponding properties
References
[1] Lukas Morand, Dirk Helm. A mixture of experts approach to handle ambiguities in parameter identification problems in material modeling, Computation Materials Science 167, 85-91 (2019)
[2] Tarek Iraki, Lukas Morand, Johannes Dornheim, Norbert Link, Dirk Helm. A multi-task learning-based optimization approach for finding diverse sets of material microstructures with desired properties and its application to texture optimization, arXiv:2111.00916 (2021)