CheMLT-F: multitask learning in biochemistry through transformer fusion

Vladislav Mun, Siamac Fazli · 2026

Read CheMLT-F: multitask learning in biochemistry through transformer fusion by Vladislav Mun, Siamac Fazli (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-13217713.txt.

Description

Drug hunting needs predictions for many properties at once: will a molecule heal, harm, dissolve, or stick to its target? Vladislav Mun and Siamac Fazli build one model for the whole panel in a 2026 paper, fusing readers for molecule strings and protein strings into a single transformer. The shared brain serves more than six hundred eighty prediction tasks.

Across thirteen public benchmarks the model matched the best poison detectors and set new marks for physical property prediction, while staying competitive on drug-target strength. Sharing parts across tasks keeps it small and quick, and new tasks bolt on as fresh heads. The design replaces a shelf of single-purpose models.

Catalog record

CREATORVladislav Mun, Siamac Fazli
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE161 KB
VERSION1.0
LICENSECC BY 4.0
RIGHTSCC BY
SOURCEEurope PMC

Source: https://europepmc.org/articles/PMC13217713

Browse the full library · About the Tower of Babel