LTVAE : Hybrid LSTM –Transformer VAE Architecture for Molecular SMILES Representation
Journal of Computational Chemistry · 05.11.2026 · 47 (29)
DOI: 10.1002/jcc.70521 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Computational Chemistry, online erschienen am 03.11.2026. Verfasst von Md Halim Mondol und 2 weiteren Autor:innen. Beteiligte Einrichtung: Electrical and Computer Engineering, North Carolina State University, Raleigh, North Carolina, USA. Die Arbeit nennt 25 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Science Foundation.
Abstract (Originalsprache)
A hybrid LSTM–Transformer Variational Autoencoder (LTVAE) framework is introduced in this paper for learning compact and chemically meaningful latent representations from SMILES strings, providing a SMILES‐to‐latent‐to‐SMILES representation model trained on ZINC15 and evaluated on ZINC15 and an out‐of‐distribution dye dataset. In the proposed model, sequential and structural dependencies in molecular strings are encoded using a bidirectional LSTM, while a Transformer‐based decoder is employed to reconstruct valid molecular structures in the evaluated settings. When trained on a curated ZINC15 dataset, the framework achieves reasonable reconstruction performance, with 98.58% token‐level accuracy, 92.19% string‐level accuracy, a 99.30% chemical validity ratio, and an average Levenshtein distance of 0.29 using a 64‐dimensional latent space. Reconstruction under an out‐of‐distribution evaluation setting is evaluated on a structurally distinct dye dataset, where 95.56% token‐level accuracy and 85.30% validity are achieved. The learned latent space preserves trends aligned with both general molecular descriptors and dye‐relevant structural descriptors, particularly conjugation length and aromatic surface fraction, with Fsp providing complementary information on molecular saturation. Controlled perturbations in the latent space suggest smooth and chemically consistent structural variations, illustrating the continuity and interpretability of the learned latent manifold.
Schlagwörter
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Electrical and Computer Engineering, North Carolina State University, Raleigh, North Carolina, USA.
- Chemical and Biomolecular Engineering, North Carolina State University, Raleigh, North Carolina, USA.
ORCID-Profile: Md Halim Mondol, Qingshan Wei, Edgar Lobaton
Förderung
- National Science Foundation
Zitieren
APA
Mondol, M. H., Wei, Q., & Lobaton, E. (2026). LTVAE : Hybrid LSTM –Transformer VAE Architecture for Molecular SMILES Representation. Journal of Computational Chemistry, 47(29). https://doi.org/10.1002/jcc.70521
Vancouver
Mondol MH, Wei Q, Lobaton E. LTVAE : Hybrid LSTM –Transformer VAE Architecture for Molecular SMILES Representation. Journal of Computational Chemistry. 2026;47(29). doi:10.1002/jcc.70521
BibTeX
@article{Mondol2026,
author = {Mondol, Md Halim and Wei, Qingshan and Lobaton, Edgar},
title = {LTVAE : Hybrid LSTM –Transformer VAE Architecture for Molecular SMILES Representation},
journal = {Journal of Computational Chemistry},
year = {2026},
volume = {47},
number = {29},
doi = {10.1002/jcc.70521}
}
EUVIMED