Training Text-to-Molecule Models with Context-Aware Tokenization

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Kim, Seojin, Song, Hyeontae, Nam, Jaehyun, Shin, Jinwoo
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866915498862575616
author Kim, Seojin
Song, Hyeontae
Nam, Jaehyun
Shin, Jinwoo
author_facet Kim, Seojin
Song, Hyeontae
Nam, Jaehyun
Shin, Jinwoo
contents Recently, text-to-molecule models have shown great potential across various chemical applications, e.g., drug-discovery. These models adapt language models to molecular data by representing molecules as sequences of atoms. However, they rely on atom-level tokenizations, which primarily focus on modeling local connectivity, thereby limiting the ability of models to capture the global structural context within molecules. To tackle this issue, we propose a novel text-to-molecule model, coined Context-Aware Molecular T5 (CAMT5). Inspired by the significance of the substructure-level contexts in understanding molecule structures, e.g., ring systems, we introduce substructure-level tokenization for text-to-molecule models. Building on our tokenization scheme, we develop an importance-based training strategy that prioritizes key substructures, enabling CAMT5 to better capture the molecular semantics. Extensive experiments verify the superiority of CAMT5 in various text-to-molecule generation tasks. Intriguingly, we find that CAMT5 outperforms the state-of-the-art methods using only 2% of training tokens. In addition, we propose a simple yet effective ensemble strategy that aggregates the outputs of text-to-molecule models to further boost the generation performance. Code is available at https://github.com/Songhyeontae/CAMT5.git.
format Preprint
id arxiv_https___arxiv_org_abs_2509_04476
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Training Text-to-Molecule Models with Context-Aware Tokenization
Kim, Seojin
Song, Hyeontae
Nam, Jaehyun
Shin, Jinwoo
Computation and Language
Artificial Intelligence
Recently, text-to-molecule models have shown great potential across various chemical applications, e.g., drug-discovery. These models adapt language models to molecular data by representing molecules as sequences of atoms. However, they rely on atom-level tokenizations, which primarily focus on modeling local connectivity, thereby limiting the ability of models to capture the global structural context within molecules. To tackle this issue, we propose a novel text-to-molecule model, coined Context-Aware Molecular T5 (CAMT5). Inspired by the significance of the substructure-level contexts in understanding molecule structures, e.g., ring systems, we introduce substructure-level tokenization for text-to-molecule models. Building on our tokenization scheme, we develop an importance-based training strategy that prioritizes key substructures, enabling CAMT5 to better capture the molecular semantics. Extensive experiments verify the superiority of CAMT5 in various text-to-molecule generation tasks. Intriguingly, we find that CAMT5 outperforms the state-of-the-art methods using only 2% of training tokens. In addition, we propose a simple yet effective ensemble strategy that aggregates the outputs of text-to-molecule models to further boost the generation performance. Code is available at https://github.com/Songhyeontae/CAMT5.git.
title Training Text-to-Molecule Models with Context-Aware Tokenization
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2509.04476