AJIN machine learning predicts atomic layer deposition outcomes from experimental precursor-substrate data
Archived as a method-development and transfer benchmark. AJIN predicts adsorption and deposition outcomes for atomic layer deposition. The November 5, 2026 preprint compares predictions with literature data and the authors’ laboratory cases, but the reviewed abstract and results section 3.7 do not identify a separate new material, deposition chemistry or scientific finding enabled by the model. Laboratory comparison alone does not turn this general prediction method into a qualifying discovery. The historical ID and source are retained outside discovery counts.
Technology: AJIN (ALD Joint Interpretable Network) with graph-neural-network embeddings (CHGNet-MLP best model)
Publication: Preprint · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Preprint
- Test type: direct test
- Tested: Literature-data classification and comparison with laboratory transfer cases; no separate new chemistry result identified in reviewed sections.
- Still unknown: Retained outside discovery counts as a historical method demonstration. This is an admission decision, not a claim that experiments were absent. Physical prediction-before-assay timing was not established.
- AI role: Machine-learning framework predicting whether a precursor-substrate-reactant combination yields ALD, with SHAP and ablation analyses of the drivers.
Evidence sources: https://arxiv.org/abs/2611.10024 · https://arxiv.org/pdf/2611.10024
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: other. Paper: not linked in this catalog; this does not mean no paper exists.
Review scope and date are not recorded; source checked does not establish full-text review.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.