Active learning for optimum experimental design—insight into perovskite oxides
Bibliographic record
Abstract
Finding the optimum material with improved properties for a given application is challenging because data acquisition in materials science and chemistry is time consuming and expensive. Therefore, dealing with small datasets is a reality in chemistry, whether the data are obtained from synthesis or computational experiments. In this work, we propose a new artificial intelligence method based on active learning (AL) to guide new experiments with as little data as possible, for optimum experimental design. The AL method is applied to ABO3 perovskites, where a descriptor based on atomic properties was developed. Several regressor algorithms were employed: artificial neural network, Gaussian process, and support vector regressor. The developed AL method was applied in the experimental design of two important materials: non-stoichiometric perovskites (Ba(1− x)A x Ti(1− y)B y O3) due to substituting ionic sites with different concentrations and elements (A = Ca, Sr, Cd; B = Zr, Sn, Hf), aiming at the maximization of the energy storage density, and stoichiometric ABO3 perovskites where different elements are changed in the A and B sites for the minimization of the formation energy. AL for experimental design is implemented in the machine learning agent for chemistry and design (MLChem4D) software, which has the potential to be applied in inorganic and organic synthesis (e.g., search for the optimum concentrations, catalysts, reactants, temperatures, and pH to improve the yield) and materials science (e.g., search the periodic table for the proper elements and their concentrations to improve the materials properties). The latter marks the first MLChem4D application for the design of perovskites.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.006 | 0.011 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".