Synergistic Benefits of Joint Molecule Generation and Property Prediction

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Izdebski, Adam, Olszewski, Jan, Gawade, Pankhil, Koras, Krzysztof, Korkmaz, Serra, Rauscher, Valentin, Tomczak, Jakub M., Szczurek, Ewa
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910201036144640
author Izdebski, Adam
Olszewski, Jan
Gawade, Pankhil
Koras, Krzysztof
Korkmaz, Serra
Rauscher, Valentin
Tomczak, Jakub M.
Szczurek, Ewa
author_facet Izdebski, Adam
Olszewski, Jan
Gawade, Pankhil
Koras, Krzysztof
Korkmaz, Serra
Rauscher, Valentin
Tomczak, Jakub M.
Szczurek, Ewa
contents Modeling the joint distribution of data samples and their properties allows to construct a single model for both data generation and property prediction, with synergistic benefits reaching beyond purely generative or predictive models. However, training joint models presents daunting architectural and optimization challenges. Here, we propose Hyformer, a transformer-based joint model that successfully blends the generative and predictive functionalities, using an alternating attention mechanism and a joint pre-training scheme. We show that Hyformer is simultaneously optimized for molecule generation and property prediction, while exhibiting synergistic benefits in conditional sampling, out-of-distribution property prediction and representation learning. Finally, we demonstrate the benefits of joint learning in a drug design use case of discovering novel antimicrobial~peptides.
format Preprint
id arxiv_https___arxiv_org_abs_2504_16559
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Synergistic Benefits of Joint Molecule Generation and Property Prediction
Izdebski, Adam
Olszewski, Jan
Gawade, Pankhil
Koras, Krzysztof
Korkmaz, Serra
Rauscher, Valentin
Tomczak, Jakub M.
Szczurek, Ewa
Machine Learning
Quantitative Methods
68T01
I.2.1
Modeling the joint distribution of data samples and their properties allows to construct a single model for both data generation and property prediction, with synergistic benefits reaching beyond purely generative or predictive models. However, training joint models presents daunting architectural and optimization challenges. Here, we propose Hyformer, a transformer-based joint model that successfully blends the generative and predictive functionalities, using an alternating attention mechanism and a joint pre-training scheme. We show that Hyformer is simultaneously optimized for molecule generation and property prediction, while exhibiting synergistic benefits in conditional sampling, out-of-distribution property prediction and representation learning. Finally, we demonstrate the benefits of joint learning in a drug design use case of discovering novel antimicrobial~peptides.
title Synergistic Benefits of Joint Molecule Generation and Property Prediction
topic Machine Learning
Quantitative Methods
68T01
I.2.1
url https://arxiv.org/abs/2504.16559