Learning on compressed molecular representations
It was proposed that a k -nearest neighbour classifier is able to outperform large-language models using compressed text as input and normalised compression distance as a metric. We successfully applied this method to cheminformatics tasks.
Authors
- Jan Weinreich (ORCID: https://orcid.org/0000-0002-9332-4543)
- Daniel Probst (ORCID: https://orcid.org/0000-0003-1737-4407)
Institutions
- University of Vienna (AT)
- École Polytechnique Fédérale de Lausanne (CH)
Publication Details
- Journal
- Digital Discovery
- Published
- 2024-11-04
- DOI
- https://doi.org/10.1039/d4dd00162a
- Primary Topic
- Fractal and DNA sequence analysis
- Type
- article
- Field-Weighted Citation Impact
- 0.00