Autonomous design of new chemical reactions using a variational autoencoder
Training datasets for artificial intelligence based chemistry models typically rely on experimental data from optimised synthetic conditions only, leading to an inherited bias in the model predictions. Here, the authors develop an artificial intelligence model based on a variational autoencoder to s...
| Principais autores: | , |
|---|---|
| Formato: | Artigo |
| Idioma: | English |
| Publicado em: |
Nature Portfolio
2022-03-01
|
| coleção: | Communications Chemistry |
| Acesso em linha: | https://doi.org/10.1038/s42004-022-00647-x |