What question did this study set out to answer?

This research aims to develop an improved model for identifying aquatic toxic chemicals using multi-task learning and graph convolutional networks.

July 24, 2026Global Environment ScienceOpen Access

Multi-Task Graph Convolutional Network Model with Improved Performance and Broad Applicability Domains for Identifying Aquatic Toxic Chemicals

View Full Paper

Ask AI

Bookmark

Authors

Randomized trial demonstrates improved identification of aquatic toxic chemicals in a large database, indicating enhanced screening methods.

This research aims to develop an improved model for identifying aquatic toxic chemicals using multi-task learning and graph convolutional networks.
Curated a dataset of 17,514 chemicals covering six ecotoxicological endpoints.
Constructed a multi-task learning graph convolutional network (MTL-GCN) for joint modeling of toxicity endpoints.
Utilized the structure-activity landscape (ADSAL) analysis for applicability domain characterization.
MTL-GCN improved area under the curve by 0.4~13.4% compared to single-task learning (STL-GCN).
94.7% of 783,681 chemicals were found in the applicability domain of at least one endpoint.
Utilized weight similarity parameter indicated effective correlation capture among endpoints.

Liu et al. (2026) studied this question.

View Full Paper

Ask AI

Bookmark

Randomized trial demonstrates improved identification of aquatic toxic chemicals in a large database, indicating enhanced screening methods.