Details
Focus
We develop AI-assisted workflows that integrate synthesis knowledge extracted from the literature, physics- and simulation-informed descriptors, and experimental data to identify promising process conditions for experimental validation. Our research focuses on transforming heterogeneous knowledge from publications, simulations, and experiments into structured synthesis datasets; developing predictive models for synthesis outcomes and material properties; and prioritizing experimentally testable process conditions.
Methods
- Synthesis Knowledge Structuring: Extract and standardize materials, processing conditions, characterization results, and synthesis outcomes from publications and experimental records.
- Synthesis and Property Modeling: Develop and validate data-driven models that relate processing variables to synthesis feasibility and target properties.
- Candidate Prioritization: Combine predictive search, domain knowledge, and optimization methods to rank candidate process conditions for experimental validation.
What we deliver
- Prioritized Candidates: Ranked process-condition candidates for targeted materials systems.
- Interpretable Design Insights: Quantitative relationships between synthesis variables, predicted outcomes, and material-performance trends.
- Iterative Experiment-Model Learning: Workflows that incorporate experimental feedback to refine models and guide subsequent experiments.
Related Papers
- AI-driven quantitative review of mobility-stability trade-off in oxide semiconductors (Nano Convergence, 2026)
- Machine learning-enhanced design of lead-free halide perovskite materials using density functional theory (Current Applied Physics, 2024)
- Predicting the synthesizability of double perovskite halides via interface reaction pathfinding (Chemistry of Materials, 2024)