用图重建法学习3D分子全局结构，提升性质预测能力
GeoRecon: Graph-Level Representation Learning for 3D Molecules via Reconstruction-Based Pretraining
- 以整个分子为单位进行图级重建预训练
- 在QM9、MD17等数据集上超越基线模型
- 适合需要全局结构感知的分子性质预测任务
预训练-微调范式在自然语言处理和计算机视觉中推动了重大进展，典型如掩码语言建模和下一词预测。但在分子表示学习中，预训练任务仍主要局限于节点级去噪，虽能捕捉局部原子环境，却常不足以编码对图级属性预测（如能量估算和分子回归）至关重要的全局分子结构。为弥补这一差距，我们提出GeoRecon，一种聚焦于分子整体而非单个原子的图级预训练框架。GeoRecon构建了一个图级重建任务：预训练时，模型被训练生成能指导几何重建的有信息量图表示，同时诱导更平滑、更具迁移性的潜在空间。这促使模型学习超越孤立原子细节的连贯全局结构特征。无需外部监督，GeoRecon在多个分子基准测试（包括QM9、MD17、MD22和3BPA）上普遍优于基线模型，证明了图级重建在生成整体性、几何感知分子嵌入方面的有效性。
原文摘要 · Abstract (English)
The pretraining-finetuning paradigm has powered major advances in domains such as natural language processing and computer vision, with representative examples including masked language modeling and next-token prediction. In molecular representation learning, however, pretraining tasks remain largely restricted to node-level denoising, which effectively captures local atomic environments but is often insufficient for encoding the global molecular structure critical to graph-level property prediction tasks such as energy estimation and molecular regression. To address this gap, we introduce GeoRecon, a graph-level pretraining framework that shifts the focus from individual atoms to the molecule as an integrated whole. GeoRecon formulates a graph-level reconstruction task: during pretraining, the model is trained to produce an informative graph representation that guides geometry reconstruction while inducing smoother and more transferable latent spaces. This encourages the learning of coherent, global structural features beyond isolated atomic details. Without relying on external supervision, GeoRecon generally improves over backbone baselines on multiple molecular benchmarks including QM9, MD17, MD22, and 3BPA, demonstrating the effectiveness of graph-level reconstruction for holistic and geometry-aware molecular embeddings.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。