Saved in:
Bibliographic Details
Main Authors: Kang, Dayoung, Kim, JongWon, Park, Jiho, Lee, Keonseock, Choi, Ji-Woong, So, Jinhyun
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2601.19561
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912852366852096
author Kang, Dayoung
Kim, JongWon
Park, Jiho
Lee, Keonseock
Choi, Ji-Woong
So, Jinhyun
author_facet Kang, Dayoung
Kim, JongWon
Park, Jiho
Lee, Keonseock
Choi, Ji-Woong
So, Jinhyun
contents Public olfaction datasets are small and fragmented across single molecules and mixtures, limiting learning of generalizable odor representations. Recent works either learn single-molecule embeddings or address mixtures via similarity or pairwise label prediction, leaving representations separate and unaligned. In this work, we propose AROMMA, a framework that learns a unified embedding space for single molecules and two-molecule mixtures. Each molecule is encoded by a chemical foundation model and the mixtures are composed by an attention-based aggregator, ensuring both permutation invariance and asymmetric molecular interactions. We further align odor descriptor sets using knowledge distillation and class-aware pseudo-labeling to enrich missing mixture annotations. AROMMA achieves state-of-the-art performance in both single-molecule and molecule-pair datasets, with up to 19.1% AUROC improvement, demonstrating a robust generalization in two domains.
format Preprint
id arxiv_https___arxiv_org_abs_2601_19561
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle AROMMA: Unifying Olfactory Embeddings for Single Molecules and Mixtures
Kang, Dayoung
Kim, JongWon
Park, Jiho
Lee, Keonseock
Choi, Ji-Woong
So, Jinhyun
Machine Learning
Artificial Intelligence
Public olfaction datasets are small and fragmented across single molecules and mixtures, limiting learning of generalizable odor representations. Recent works either learn single-molecule embeddings or address mixtures via similarity or pairwise label prediction, leaving representations separate and unaligned. In this work, we propose AROMMA, a framework that learns a unified embedding space for single molecules and two-molecule mixtures. Each molecule is encoded by a chemical foundation model and the mixtures are composed by an attention-based aggregator, ensuring both permutation invariance and asymmetric molecular interactions. We further align odor descriptor sets using knowledge distillation and class-aware pseudo-labeling to enrich missing mixture annotations. AROMMA achieves state-of-the-art performance in both single-molecule and molecule-pair datasets, with up to 19.1% AUROC improvement, demonstrating a robust generalization in two domains.
title AROMMA: Unifying Olfactory Embeddings for Single Molecules and Mixtures
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2601.19561