<ul><li>Data-driven approaches such as deep learning can result in predictive models for material properties with exceptional accuracy and efficiency.</li><li>To address the limitations of sparse datasets and weak correlations between properties, the authors propose a data fusion technique.</li><li>By combining the learned molecular embeddings from single-task models, the fused, multi-task models outperform standard multi-task models.</li><li>The experimental results demonstrate the enhanced prediction capabilities of the fused models for data-limited properties.</li></ul>

Data Fusion of Deep Learned Molecular Embeddings for Property Prediction

Discover more