
by MIT
Generative deep-learning model for creating stable crystal structures with targeted materials properties
Web appFree
Step 1: Begin with a small reproducible benchmark for generative deep-learning model for creating stable crystal structures with targeted materials properties. Use a dataset or example with known expected behavior, pin the model or package version, and document the compute environment before applying the tool to novel research.
Step 2: Use CDVAE to generate a first set of candidates, analyses, structures, literature findings, segmentations, or experimental suggestions. Inspect the underlying evidence and intermediate outputs instead of accepting only the final result. Where possible, compare against a trusted baseline or an independent method.
Step 3: Validate the result using the standards of the domain: experimental measurement for chemistry or biology, held-out benchmarks for predictive models, expert review for literature synthesis, or quantitative ground truth for imaging. Record failed predictions and negative results as well as successful examples.
Step 4: Before scaling the workflow, confirm licensing and data-use terms, model and dataset versions, reproducibility, privacy or biosafety requirements, and the limits of the training distribution. Re-check the primary documentation when models or hosted services change.
ChempropChemprop is a PyTorch-based framework for training and evaluating message-passing neural networks that predict molecular properties. The site hosts its documentation for version 2.3.1.
ASKCOSASKCOS is a hosted set of computer-aided design tools for organic synthesis. You enter or draw a molecule or reaction SMILES and see the tasks available for it, such as synthetic planning modules.
Therapeutics Data CommonsTherapeutics Data Commons is a collection of AI-ready datasets, tasks and benchmarks for drug discovery. Everything is accessed through an open Python library.
DeepChemDeepChem is an open-source Python library for machine learning and deep learning on molecular and quantum datasets. It builds on PyTorch and other popular frameworks.
Boltz-1Boltz builds frontier AI models for biomolecular structure prediction and design, offered through Boltz Lab and the Boltz API. You can design small molecules and proteins and predict structures.