Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
Jožef Stefan Institute
4
0
5
67
A handful of chemically diverse unlabeled molecules can boost QSAR model performance more than massive, randomly sampled datasets.
SORT shifts the focus from brittle term selection to adaptive basis design, enabling more robust and accurate equation discovery from noisy data.
Dataset properties dictate how design choices impact model performance, revealing that fine-tuning and architecture dominate in large datasets, while initialization is critical in data-scarce environments.
Spectral learning faces a fundamental noise threshold that can destabilize coefficient estimates, challenging the reliability of functional recovery from noisy data.