The precise prediction of molecular properties holds paramount importance in
facilitating the development of innovative treatments and comprehending the
intricate interplay between chemicals and biological systems. In this study, we
propose a novel approach that integrates graph representations of individual
molecular structures with multi-domain information from biomedical knowledge
graphs (KGs). Integrating information from both levels, we can pre-train a more
extensive and robust representation for both molecule-level and KG-level
prediction tasks with our novel self-supervision strategy. For performance
evaluation, we fine-tune our pre-trained model on 11 challenging chemical
property prediction tasks. Results from our framework demonstrate our
fine-tuned models outperform existing state-of-the-art models.

本文提出了一种新的方法，将单个分子结构的图形表示与生物医学知识图的多领域信息相结合，利用自监督策略进行预训练，成功地应用于化学性质预测任务，取得了超过现有最先进模型的表现。

Gode：将生化知识图谱整合到预训练分子图神经网络中

Gode -- Integrating Biochemical Knowledge Graph into Pre-training  Molecule Graph Neural Network

We propose a 2-WL-like geometric graph isomorphism test and prove it is
complete when applied to Euclidean Graphs in $\mathbb{R}^3$. We then use recent
results on multiset embeddings to devise an efficient geometric GNN model with
equivalent separation power. We verify empirically that our GNN model is able
to separate particularly challenging synthetic examples, and demonstrate its
usefulness for a chemical property prediction problem.

该论文提出了一种基于 2-WL 的几何图同构测试方法，证明其对三维欧几里得图具有完备性。然后利用最新的多重集嵌入结果来设计一个具备等效分离能力的有效几何 GNN 模型，并经验证实验表明 GNN 模型能在特别具有挑战性的综合实例上进行分离，证明了它在化学性质预测问题中的实用性。

欧几里得图的完全神经网络

Complete Neural Networks for Euclidean Graphs

With access to large datasets, deep neural networks (DNN) have achieved
human-level accuracy in image and speech recognition tasks. However, in
chemistry, data is inherently small and fragmented. In this work, we develop an
approach of using rule-based knowledge for training ChemNet, a transferable and
generalizable deep neural network for chemical property prediction that learns
in a weak-supervised manner from large unlabeled chemical databases. When
coupled with transfer learning approaches to predict other smaller datasets for
chemical properties that it was not originally trained on, we show that
ChemNet's accuracy outperforms contemporary DNN models that were trained using
conventional supervised learning. Furthermore, we demonstrate that the ChemNet
pre-training approach is equally effective on both CNN (Chemception) and RNN
(SMILES2vec) models, indicating that this approach is network architecture
agnostic and is effective across multiple data modalities. Our results indicate
a pre-trained ChemNet that incorporates chemistry domain knowledge, enables the
development of generalizable neural networks for more accurate prediction of
novel chemical properties.

本研究旨在提出一种基于规则知识训练的 ChemNet 神经网络，从大规模未标记化的化学数据库中进行弱监督学习，通过迁移学习的方法对化学性质进行预测，并证明其相比于常规监督学习的深度神经网络模型具有更高的准确性，同时不受网络结构类型限制。该预训练 ChemNet 的应用可以通过整合化学领域知识，实现对新型化学性质预测的更为精确和系统化。