arXiv subjects

Kim E. Jelfs

Publications and source records attributed to Kim E. Jelfs.

7 recordsLinked to original sources

arXiv2026

Autoregressive latent diffusion for 3D molecule generation

Three-dimensional (3D) molecule generation has been dominated by diffusion models, which achieve strong generation quality but typically require the molecular size to be specified a priori. Recent autoregressive approaches have substantially narrowed the performance gap while naturally supporting variable-length generation and conditioning on partial molecular context. However, balancing unconditional and context-conditioned generation remains challenging. We introduce KRONOS, a latent autoregressive diffusion framework that generates molecules in the latent space of a pre-trained autoencoder, jointly modeling molecular graph topology and geometry, while retaining the flexibility of autoregressive generation. We further introduce a mixed training strategy inspired by Fill-in-the Middle (FIM) paradigm, enabling both unconditional and fragment-conditioned molecular generation within a single left-to-right autoregressive model. Experiments on QM9 and GEOM-Drugs demonstrate that KRONOS achieves leading unconditional generation performance among autoregressive methods, while remaining competitive with diffusion models. Moreover, fragment-conditioned generation is achieved with negligible impact on unconditional generation performance, demonstrating that both generation paradigms can be supported within a single architecture.

cs.LG↗

arXiv2026

PolyMon: A Unified Framework for Polymer Property Prediction

Accurate prediction of polymer properties is essential for materials design, but remains challenging due to data scarcity, diverse polymer representations, and the lack of systematic evaluation across modelling choices. Here, we present PolyMon, a unified and accessible framework that integrates multiple polymer representations, machine learning methods, and training strategies within a single, accessible platform. PolyMon supports various descriptors and graph construction strategies for polymer representations, and includes a wide range of models, from tabular models to graph neural networks, along with flexible training strategies including multi-fidelity learning, Δ-learning, active learning, and ensemble learning. Using five key polymer properties as benchmarks, we perform systematic evaluations to assess how representations and models affect predictive performance. These case studies further illustrate how different training strategies can be applied within a consistent workflow to leverage limited data and incorporate physical model derived information. Overall, PolyMon provides a comprehensive and extensible foundation for benchmarking and advancing machine learning-based polymer property prediction. The code is available at github.com/fate1997/polymon.

cond-mat.soft↗

arXiv2026

A Computational Study of Organic Molecular Crystals for Photocatalytic Water Splitting

Organic crystalline materials are potential candidates for photocatalytic overall water splitting (OWS). Although organic crystals have been heavily investigated for application in organic electronics, such as organic light-emitting diodes (OLEDs) and solar cells, there have been comparatively fewer studies into OWS in these materials. A major challenge is the large number of electronic and structural criteria that must be met for a material to make a viable OWS photocatalyst. Optical absorption, reduction and oxidation potentials and charge-transport properties are among the key considerations, and these are influenced both by molecular properties and the solid-state packing arrangement, making computational modelling challenging. Here, we investigate a series of known organic electronic materials that have published crystal structures using periodic density functional theory (DFT) and compare their calculated electronic properties of optical absorption and reduction and oxidation potentials with literature experimental data. Furthermore we perform a series of gas-phase molecular calculations which show a good agreement with literature data and periodic DFT for the optoelectronic properties of the organic molecular crystals studied, showing that gas-phase molecular calculations could be used to screen organic crystals for OWS at a reduced computational cost.

cond-mat.mtrl-sci↗

arXiv2025

The Catechol Benchmark: Time-series Solvent Selection Data for Few-shot Machine Learning

Machine learning has promised to change the landscape of laboratory chemistry, with impressive results in molecular property prediction and reaction retro-synthesis. However, chemical datasets are often inaccessible to the machine learning community as they tend to require cleaning, thorough understanding of the chemistry, or are simply not available. In this paper, we introduce a novel dataset for yield prediction, providing the first-ever transient flow dataset for machine learning benchmarking, covering over 1200 process conditions. While previous datasets focus on discrete parameters, our experimental set-up allow us to sample a large number of continuous process conditions, generating new challenges for machine learning models. We focus on solvent selection, a task that is particularly difficult to model theoretically and therefore ripe for machine learning applications. We showcase benchmarking for regression algorithms, transfer-learning approaches, feature engineering, and active learning, with important applications towards solvent replacement and sustainable manufacturing.

cs.LG↗

arXiv2025

Comparative Analysis of search Approaches to Discover Donor Molecules for Organic Solar Cells

Identifying organic molecules with desirable properties from the extensive chemical space can be challenging, particularly when property evaluation methods are time-consuming and resource intensive. In this study, we illustrate this challenge by exploring the chemical space of large oligomers, constructed from monomeric building blocks, for potential use in organic photovoltaics (OPV). For this purpose, we developed a python package to search the chemical space using a building block approach: stk-search. We use stk-search to compare a variety of search algorithms, including those based upon Bayesian optimization and evolutionary approaches. Initially, we evaluated and compared the performance of different search algorithms within a precomputed search space. We then extended our investigation to the vast chemical space of molecules formed of 6 building blocks (6-mers), comprising over $10^{14}$ molecules. Notably, while some algorithms show only marginal improvements over a random search approach in a relatively small, precomputed, search space, their performance in the larger chemical space is orders of magnitude better. Specifically, Bayesian optimization identified a thousand times more promising molecules with the desired properties compared to random search, using the same computational resources.

physics.chem-ph↗

arXiv2024

Applying Multi-Fidelity Bayesian Optimization in Chemistry: Open Challenges and Major Considerations

Multi fidelity Bayesian optimization (MFBO) leverages experimental and or computational data of varying quality and resource cost to optimize towards desired maxima cost effectively. This approach is particularly attractive for chemical discovery due to MFBO's ability to integrate diverse data sources. Here, we investigate the application of MFBO to accelerate the identification of promising molecules or materials. We specifically analyze the conditions under which lower fidelity data can enhance performance compared to single-fidelity problem formulations. We address two key challenges, selecting the optimal acquisition function, understanding the impact of cost, and data fidelity correlation. We then discuss how to assess the effectiveness of MFBO for chemical discovery.

cs.LG↗

arXiv2024

PolyCL: Contrastive Learning for Polymer Representation Learning via Explicit and Implicit Augmentations

Polymers play a crucial role in a wide array of applications due to their diverse and tunable properties. Establishing the relationship between polymer representations and their properties is crucial to the computational design and screening of potential polymers via machine learning. The quality of the representation significantly influences the effectiveness of these computational methods. Here, we present a self-supervised contrastive learning paradigm, PolyCL, for learning high-quality polymer representation without the need for labels. Our model combines explicit and implicit augmentation strategies for improved learning performance. The results demonstrate that our model achieves either better, or highly competitive, performances on transfer learning tasks as a feature extractor without an overcomplicated training strategy or hyperparameter optimisation. Further enhancing the efficacy of our model, we conducted extensive analyses on various augmentation combinations used in contrastive learning. This led to identifying the most effective combination to maximise PolyCL's performance.

cs.LG↗