---
title: "通用分子基础模型成功迁移至多种嗅觉任务"
source: "arXiv AI/CL/LG（论文）"
source_tier: "t2"
url: "https://arxiv.org/abs/2608.25893v1"
permalink: "https://feed.astralor.com/items/itm_327adb4fda874578fe88740e"
published: "2026-08-26T15:11:15.000Z"
score: 75
channel: "ai"
tags: ["GS-LF","Uni-Mol2","分子表示学习","基础模型","机器嗅觉"]
---

# 通用分子基础模型成功迁移至多种嗅觉任务

## AI 点评

> 这篇论文让我眼前一亮的地方在于，它没有堆砌新架构，而是把一个通用分子模型微调后直接拿去打多个嗅觉任务，这种“一次训练、多处复用”的思路对实际应用特别有价值。对做分子性质预测、香精香料或气味传感器开发的人来说，这可能是降低数据成本的一个实用方向。不过立体化学感知的预测仍是难点，后续值得关注。

## 速览

该论文探索了分子基础模型在多任务嗅觉预测中的迁移能力。研究者在GS-LF基准上对Uni-Mol2进行微调，使其掌握多标签气味描述符预测，并在四个下游任务（跨数据集气味描述符、气味有无分类、对映体评估、混合物区分性）中不做额外训练直接评估。结果显示，微调后的模型在主要基准上媲美或超越专用嗅觉模型，并稳定迁移至各下游任务。对映体分析表明三维分子表示能区分镜像分子，但精确预测立体化学感知仍是挑战。研究支持“一次训练、多任务迁移”的机器嗅觉范式。

## 关键点

- 在GS-LF基准上微调Uni-Mol2，模型在四个下游嗅觉任务上无需额外训练即可迁移。
- 微调后的模型在主要基准上达到或超越专用嗅觉模型性能。
- 三维分子表示能区分对映异构体，而二维图模型无法做到。
- 研究支持机器嗅觉中“训练一次、跨任务迁移”的范式。

## 全文

Foundation models have transformed molecular property prediction, yet it remains unclear whether a molecular foundation model, fine-tuned on a single canonical olfactory prediction task, can learn representations that transfer across diverse machine olfaction problems. We investigate this question by fine-tuning Uni-Mol2 on the GS-LF benchmark for multi-label odor descriptor prediction and evaluating the resulting model, without additional deep-learning training, on four complementary downstream settings: cross-dataset odor descriptor prediction, odorous-versus-odorless classification, enantiomer evaluation, and odor mixture discriminability. The fine-tuned model matches or exceeds the performance of the state-of-the-art olfaction-specific baseline on the primary GS-LF benchmark and consistently transfers across these downstream evaluations. The enantiomer analysis further shows that three-dimensional molecular representations distinguish mirror-image molecules in a way that two-dimensional graph models fundamentally cannot, although accurately predicting the perceptual consequences of stereochemistry remains an open challenge. Together, these results support a train-once, transfer-across-tasks paradigm for machine olfaction and suggest that chemically pretrained molecular representations provide a strong foundation for transferable olfactory prediction.

---

引用前请核对原文;本文由星际流动整理:https://feed.astralor.com/items/itm_327adb4fda874578fe88740e