Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
February 27, 2026

ADFC-ATP: Attention-Guided Dual-View Fusion and Contrastive Pretraining for Robust Aquatic Toxicity Prediction.

View Full Paper
Ask AI
Bookmark
Share

Authors

JJJixuan JiaUniversity of Science and Technology LiaoningXYXin YangBeijing Institute of TechnologyYFYing FangUniversity of Science and Technology Liaoning

Discussion

Loading...

Member takes

Implication

This framework improves aquatic toxicity prediction using dual-view fusion and contrastive learning, implying greater ecological safety measures.

Key Points

  • The aim is to enhance the prediction of aquatic toxicity by improving robustness and interpretability through advanced deep learning techniques.
  • Integrated dual-view molecular graph fusion with contrastive learning based on NT-Xent loss.
  • Utilized structural graph augmentations during pretraining to boost robustness.
  • Employed a graph attention encoder for hierarchical substructure learning.
  • Implemented an adaptive attention-based fusion mechanism for toxicity prediction.
  • ADFC-ATP demonstrated an average relative AUC improvement of 10.2% over classic graph neural network models.
  • Ablation studies confirmed the significance of scaffold preservation in the model's performance.
  • Attention visualizations highlighted the model's capacity to identify toxicophores in line with QSAR principles.

Cite This Study

Jia et al. (2026) studied this question.

synapsesocial.com/papers/69a1351ded1d949a99abea3bhttps://doi.org/10.1111/jcmm.71067
View Full Paper
Ask AI
Bookmark
Share

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

  1. 1Multi-task aquatic toxicity prediction model based on multi-level features fusion2024 · 85 citations
  2. 2Multi-Task Graph Convolutional Network Model with Improved Performance and Broad Applicability Domains for Identifying Aquatic Toxic Chemicals2026
  3. 3EGAT-Pool: A Hierarchical Graph Attention Network with Edge-aware SAG Pooling for Robust Environmental Toxicity Prediction2026
  4. 4GraphADT: empowering interpretable predictions of acute dermal toxicity with multi-view graph pooling and structure remapping2024 · 16 citations
  5. 5AquaTox-Predictor: An MMoE-Enhanced Multimodal Deep Learning Framework for Aquatic Ecotoxicity Prediction2026