Molecular Relational Learning (MRL), aiming to understand interactions
between molecular pairs, plays a pivotal role in advancing biochemical
research. Recently, the adoption of large language models (LLMs), known for
their vast knowledge repositories and advanced logical inference capabilities,
has emerged as a promising way for efficient and effective MRL. Despite their
potential, these methods predominantly rely on the textual data, thus not fully
harnessing the wealth of structural information inherent in molecular graphs.
Moreover, the absence of a unified framework exacerbates the information
underutilization, as it hinders the sharing of interaction rationale learned
across diverse datasets. To address these challenges, this work proposes a
novel LLM-based multi-modal framework for Molecular inTeraction prediction
following Chain-of-Thought (CoT) theory, termed MolTC, which can efficiently
integrate rich graphical information of molecular pairs. For achieving a
unified MRL, MolTC innovatively develops a dynamic parameter-sharing strategy
for cross-dataset information exchange, and introduces a Multi-hierarchical CoT
principle to refine training paradigm. Our experiments, conducted across twelve
varied datasets involving over 4,000,000 molecular pairs, demonstrate the
superiority of our method over current GNN and LLM-based baselines. On the top
of that, a comprehensive Molecular Interactive Instructions dataset is
constructed for the development of biochemical LLM, including our MolTC. Code
is available at this https URL

该研究提出了一个基于 LLMs 的多模态框架 MolTC，以图像化信息为基础，实现了统一的分子间相互作用预测，并进行了包括 400 万分子对的多数据集实验，证明了该方法在基于 GNN 和 LLM 的基准测试中的优越性。

MolTC: 语言模型中的分子关系建模

MolTC: Towards Molecular Relational Modeling In Language Models

Recently there has been increased interest in semi-supervised classification
in the presence of graphical information. A new class of learning models has
emerged that relies, at its most basic level, on classifying the data after
first applying a graph convolution. To understand the merits of this approach,
we study the classification of a mixture of Gaussians, where the data
corresponds to the node attributes of a stochastic block model. We show that
graph convolution extends the regime in which the data is linearly separable by
a factor of roughly $1/\sqrt{D}$, where $D$ is the expected degree of a node,
as compared to the mixture model data on its own. Furthermore, we find that the
linear classifier obtained by minimizing the cross-entropy loss after the graph
convolution generalizes to out-of-distribution data where the unseen data can
have different intra- and inter-class edge probabilities from the training
data.

该研究探讨了在图形信息存在的情况下，基于图卷积进行数据分类的方法，发现图卷积扩展了数据线性可分的范围，并且在最小化交叉熵损失后，得到的线性分类器具有很好的泛化能力。