Enregistré dans:
Détails bibliographiques
Auteurs principaux: Soares, Eduardo, Shirasuna, Victor, Brazil, Emilio Vital, Cerqueira, Renato, Zubarev, Dmitry, Schmidt, Kristin
Format: Preprint
Publié: 2024
Sujets:
Accès en ligne:https://arxiv.org/abs/2407.20267
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866916848791977984
author Soares, Eduardo
Shirasuna, Victor
Brazil, Emilio Vital
Cerqueira, Renato
Zubarev, Dmitry
Schmidt, Kristin
author_facet Soares, Eduardo
Shirasuna, Victor
Brazil, Emilio Vital
Cerqueira, Renato
Zubarev, Dmitry
Schmidt, Kristin
contents Large-scale pre-training methodologies for chemical language models represent a breakthrough in cheminformatics. These methods excel in tasks such as property prediction and molecule generation by learning contextualized representations of input tokens through self-supervised learning on large unlabeled corpora. Typically, this involves pre-training on unlabeled data followed by fine-tuning on specific tasks, reducing dependence on annotated datasets and broadening chemical language representation understanding. This paper introduces a large encoder-decoder chemical foundation models pre-trained on a curated dataset of 91 million SMILES samples sourced from PubChem, which is equivalent to 4 billion of molecular tokens. The proposed foundation model supports different complex tasks, including quantum property prediction, and offer flexibility with two main variants (289M and $8\times289M$). Our experiments across multiple benchmark datasets validate the capacity of the proposed model in providing state-of-the-art results for different tasks. We also provide a preliminary assessment of the compositionality of the embedding space as a prerequisite for the reasoning tasks. We demonstrate that the produced latent space is separable compared to the state-of-the-art with few-shot learning capabilities.
format Preprint
id arxiv_https___arxiv_org_abs_2407_20267
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle A Large Encoder-Decoder Family of Foundation Models For Chemical Language
Soares, Eduardo
Shirasuna, Victor
Brazil, Emilio Vital
Cerqueira, Renato
Zubarev, Dmitry
Schmidt, Kristin
Machine Learning
Artificial Intelligence
Chemical Physics
Large-scale pre-training methodologies for chemical language models represent a breakthrough in cheminformatics. These methods excel in tasks such as property prediction and molecule generation by learning contextualized representations of input tokens through self-supervised learning on large unlabeled corpora. Typically, this involves pre-training on unlabeled data followed by fine-tuning on specific tasks, reducing dependence on annotated datasets and broadening chemical language representation understanding. This paper introduces a large encoder-decoder chemical foundation models pre-trained on a curated dataset of 91 million SMILES samples sourced from PubChem, which is equivalent to 4 billion of molecular tokens. The proposed foundation model supports different complex tasks, including quantum property prediction, and offer flexibility with two main variants (289M and $8\times289M$). Our experiments across multiple benchmark datasets validate the capacity of the proposed model in providing state-of-the-art results for different tasks. We also provide a preliminary assessment of the compositionality of the embedding space as a prerequisite for the reasoning tasks. We demonstrate that the produced latent space is separable compared to the state-of-the-art with few-shot learning capabilities.
title A Large Encoder-Decoder Family of Foundation Models For Chemical Language
topic Machine Learning
Artificial Intelligence
Chemical Physics
url https://arxiv.org/abs/2407.20267