Digital discovery最新文献_第3页

Exploring the expertise of large language models in materials science and metallurgical engineering†

IF 6.2

Digital discovery Pub Date : 2025-01-20 DOI: 10.1039/D4DD00319E

Christophe Bajan and Guillaume Lambard

{"title":"Exploring the expertise of large language models in materials science and metallurgical engineering†","authors":"Christophe Bajan and Guillaume Lambard","doi":"10.1039/D4DD00319E","DOIUrl":"https://doi.org/10.1039/D4DD00319E","url":null,"abstract":"The integration of artificial intelligence into various domains is rapidly increasing, with Large Language Models (LLMs) becoming more prevalent in numerous applications. This work is included in an overall project which aims to train an LLM specifically in the field of materials science. To assess the impact of this specialized training, it is essential to establish the baseline performance of existing LLMs in materials science. In this study, we evaluated 15 different LLMs using the MaScQA question answering (Q&A) benchmark. This benchmark comprises questions from the Graduate Aptitude Test in Engineering (GATE), tailored to test models' capabilities in answering questions related to materials science and metallurgical engineering. Our results indicate that closed-source LLMs, such as Claude-3.5-Sonnet and GPT-4o, perform the best with an overall accuracy of ∼84%, while open-source models, such as Llama3-70b and Phi3-14b, top at ∼56% and ∼43%, respectively. These findings provide a baseline for the raw capabilities of LLMs on Q&A tasks applied to materials science, and emphasise the substantial improvement that could be brought to open-source models via prompt engineering and fine-tuning strategies. We anticipate that this work could push the adoption of LLMs as valuable assistants in materials science, demonstrating their utilities in this specialised domain and related sub-domains.","PeriodicalId":72816,"journal":{"name":"Digital discovery","volume":" 2","pages":" 500-512"},"PeriodicalIF":6.2,"publicationDate":"2025-01-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://pubs.rsc.org/en/content/articlepdf/2025/dd/d4dd00319e?page=search","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143396440","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Composition and structure analyzer/featurizer for explainable machine-learning models to predict solid state structures†

IF 6.2

Digital discovery Pub Date : 2025-01-17 DOI: 10.1039/D4DD00332B

Emil I. Jaffal, Sangjoon Lee, Danila Shiryaev, Alex Vtorov, Nikhil Kumar Barua, Holger Kleinke and Anton O. Oliynyk

{"title":"Composition and structure analyzer/featurizer for explainable machine-learning models to predict solid state structures†","authors":"Emil I. Jaffal, Sangjoon Lee, Danila Shiryaev, Alex Vtorov, Nikhil Kumar Barua, Holger Kleinke and Anton O. Oliynyk","doi":"10.1039/D4DD00332B","DOIUrl":"https://doi.org/10.1039/D4DD00332B","url":null,"abstract":"Traditional and non-classical machine learning models for solid-state structure prediction have predominantly relied on compositional features (derived from properties of constituent elements) to predict the existence of a structure and its properties. However, the lack of structural information can be a source of suboptimal property mapping and increased predictive uncertainty. To address this challenge, we have introduced a strategy that generates and combines both compositional and structural features with minimal programming expertise required. Our approach utilizes open-source, interactive Python programs named Composition Analyzer Featurizer (CAF) and Structure Analyzer Featurizer (SAF). CAF generates numerical compositional features from a list of formulae provided in an Excel file, while SAF extracts numerical structural features from a .cif file by generating a supercell. 133 features from CAF and 94 features from SAF are used either individually or in combination to cluster nine structure types in equiatomic AB intermetallics. The performance is comparable to those with features from JARVIS, MAGPIE, mat2vec, and OLED datasets in PLS-DA, SVM, and XGBoost models. Our SAF + CAF features provide a cost-efficient and reliable solution, even with the PLS-DA method, where a significant fraction of the most contributing features is the same as those identified in the more computationally intensive XGBoost models.","PeriodicalId":72816,"journal":{"name":"Digital discovery","volume":" 2","pages":" 548-560"},"PeriodicalIF":6.2,"publicationDate":"2025-01-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://pubs.rsc.org/en/content/articlepdf/2025/dd/d4dd00332b?page=search","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143396443","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Balancing molecular information and empirical data in the prediction of physico-chemical properties†

IF 6.2

Digital discovery Pub Date : 2025-01-15 DOI: 10.1039/D4DD00154K

Johannes Zenn, Dominik Gond, Fabian Jirasek and Robert Bamler

{"title":"Balancing molecular information and empirical data in the prediction of physico-chemical properties†","authors":"Johannes Zenn, Dominik Gond, Fabian Jirasek and Robert Bamler","doi":"10.1039/D4DD00154K","DOIUrl":"https://doi.org/10.1039/D4DD00154K","url":null,"abstract":"Predicting the physico-chemical properties of pure substances and mixtures is a central task in thermodynamics. Established prediction methods range from fully physics-based ab initio calculations, which are only feasible for very simple systems, over descriptor-based methods that use some information on the molecules to be modeled together with fitted model parameters (e.g., quantitative-structure–property relationship methods or classical group contribution methods), to representation-learning methods, which may, in extreme cases, completely ignore molecular descriptors and extrapolate only from existing data on the property to be modeled (e.g., matrix completion methods). In this work, we propose a general method for combining molecular descriptors with representation learning using the so-called expectation maximization algorithm from the probabilistic machine-learning literature, which uses uncertainty estimates to trade off between the two approaches. The proposed hybrid model exploits chemical structure information using graph neural networks, but it automatically detects cases where structure-based predictions are unreliable, in which case it corrects them by representation-learning based predictions that can better specialize to unusual cases. The effectiveness of the proposed method is demonstrated using the prediction of activity coefficients in binary mixtures as an example. The results are compelling, as the method significantly improves predictive accuracy over the current state of the art, showcasing its potential to advance the prediction of physico-chemical properties in general.","PeriodicalId":72816,"journal":{"name":"Digital discovery","volume":" 3","pages":" 683-693"},"PeriodicalIF":6.2,"publicationDate":"2025-01-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://pubs.rsc.org/en/content/articlepdf/2025/dd/d4dd00154k?page=search","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143602056","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A multi-robot–multi-task scheduling system for autonomous chemistry laboratories†

IF 6.2

Digital discovery Pub Date : 2025-01-14 DOI: 10.1039/D4DD00313F

Junyi Zhou, Man Luo, Linjiang Chen, Qing Zhu, Shan Jiang, Fei Zhang, Weiwei Shang and Jun Jiang

{"title":"A multi-robot–multi-task scheduling system for autonomous chemistry laboratories†","authors":"Junyi Zhou, Man Luo, Linjiang Chen, Qing Zhu, Shan Jiang, Fei Zhang, Weiwei Shang and Jun Jiang","doi":"10.1039/D4DD00313F","DOIUrl":"https://doi.org/10.1039/D4DD00313F","url":null,"abstract":"We present a multi-robot–multi-task scheduling system designed for autonomous chemistry laboratories to enhance the efficiency of executing complex chemical experiments. Building on the herein formulated and developed scheduling algorithms and employing a constraint programming approach, the scheduling system optimizes task allocation across three robots and 18 experimental stations, facilitating the coordinated and concurrent execution of experiments. The system allows for dynamic task insertion during ongoing operations without significant disruption, enhancing laboratory efficiency and flexibility while providing a scalable solution for high-throughput experimentation. In real-world applications involving four diverse chemical experiments with varied step counts, step durations, and sample throughputs, the system demonstrated its ability to reduce total execution time by nearly 40% compared to sequential execution of individual experiments, where in-experiment tasks were already optimized for concurrency. Our multi-robot–multi-task system represents a timely and significant advancement in autonomous chemistry, enabling automated laboratories to conduct experiments with greater efficiency and versatility. By reducing the time and resources required for experimentation, it accelerates the pace of scientific discovery and offers a robust framework for developing more sophisticated autonomous laboratories capable of handling increasingly complex and diverse scientific tasks.","PeriodicalId":72816,"journal":{"name":"Digital discovery","volume":" 3","pages":" 636-652"},"PeriodicalIF":6.2,"publicationDate":"2025-01-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://pubs.rsc.org/en/content/articlepdf/2025/dd/d4dd00313f?page=search","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143602064","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Does one need to polish electrodes in an eight pattern? Automation provides the answer†

IF 6.2

Digital discovery Pub Date : 2025-01-10 DOI: 10.1039/D4DD00323C

Naruki Yoshikawa, Gun Deniz Akkoc, Sergio Pablo-García, Yang Cao, Han Hao and Alán Aspuru-Guzik

引用次数: 0

Automated computational workflows for muon spin spectroscopy

IF 6.2

Digital discovery Pub Date : 2025-01-10 DOI: 10.1039/D4DD00314D

Ifeanyi J. Onuorah, Miki Bonacci, Muhammad M. Isah, Marcello Mazzani, Roberto De Renzi, Giovanni Pizzi and Pietro Bonfà

{"title":"Automated computational workflows for muon spin spectroscopy","authors":"Ifeanyi J. Onuorah, Miki Bonacci, Muhammad M. Isah, Marcello Mazzani, Roberto De Renzi, Giovanni Pizzi and Pietro Bonfà","doi":"10.1039/D4DD00314D","DOIUrl":"https://doi.org/10.1039/D4DD00314D","url":null,"abstract":"Positive muon spin rotation and relaxation spectroscopy is a well established experimental technique for studying materials. It provides a local probe that generally complements scattering techniques in the study of magnetic systems and represents a valuable alternative for materials that display strong incoherent scattering or neutron absorption. Computational methods can effectively quantify the microscopic interactions underlying the experimentally observed signal, thus substantially boosting the predictive power of this technique. Here, we present an efficient set of algorithms and workflows devoted to the automation of this task. In particular, we adopt the so-called DFT+μ procedure, where the system is characterized in the density functional theory (DFT) framework with the muon modeled as a hydrogen impurity. We devise an automated strategy to obtain candidate muon stopping sites, their dipolar interaction with the nuclei, and hyperfine interactions with the electronic ground state. We validate the implementation on well-studied compounds, showing the effectiveness of our protocol in terms of accuracy and simplicity of use.","PeriodicalId":72816,"journal":{"name":"Digital discovery","volume":" 2","pages":" 523-538"},"PeriodicalIF":6.2,"publicationDate":"2025-01-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://pubs.rsc.org/en/content/articlepdf/2025/dd/d4dd00314d?page=search","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143396441","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

General data management workflow to process tabular data in automated and high-throughput heterogeneous catalysis research†‡

IF 6.2

Digital discovery Pub Date : 2025-01-10 DOI: 10.1039/D4DD00350K

Erwin Lam, Tanguy Maury, Sebastian Preiss, Yuhui Hou, Hannes Frey, Caterina Barillari and Paco Laveille

{"title":"General data management workflow to process tabular data in automated and high-throughput heterogeneous catalysis research†‡","authors":"Erwin Lam, Tanguy Maury, Sebastian Preiss, Yuhui Hou, Hannes Frey, Caterina Barillari and Paco Laveille","doi":"10.1039/D4DD00350K","DOIUrl":"https://doi.org/10.1039/D4DD00350K","url":null,"abstract":"Data management and processing are crucial steps to implement streamlined and standardized data workflows for automated and high-throughput laboratories. Electronic laboratory notebooks (ELNs) have proven to be effective to manage data in combination with a laboratory information management system (LIMS) to connect data and inventory. However, streamlined data processing does still pose a challenge on an ELN especially with large data. Herein we present a Python library that allows streamlining and automating data management of tabular data generated within a data-driven, automated high-throughput laboratory with a focus on heterogeneous catalysis R&D. This approach speeds up data processing and avoids errors introduced by manual data processing. Through the Python library, raw data from individual instruments related to a project are downloaded from an ELN, merged in a relational database fashion, processed and re-uploaded back to the ELN. Straightforward data merging is especially important, since information stemming from multiple devices needs to be processed together. By providing a configuration file that contains all the data management information, data merging and processing of individual data sources is executed. Having established streamlined data management workflows allows standardization of data handling and contributes to the implementation and use of open research data following Findable, Accessible, Interoperable and Reusable (FAIR) principles in the field of heterogeneous catalysis.","PeriodicalId":72816,"journal":{"name":"Digital discovery","volume":" 2","pages":" 539-547"},"PeriodicalIF":6.2,"publicationDate":"2025-01-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://pubs.rsc.org/en/content/articlepdf/2025/dd/d4dd00350k?page=search","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143396442","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Predicting hydrogen atom transfer energy barriers using Gaussian process regression†

IF 6.2

Digital discovery Pub Date : 2025-01-10 DOI: 10.1039/D4DD00174E

Evgeni Ulanov, Ghulam A. Qadir, Kai Riedmiller, Pascal Friederich and Frauke Gräter

{"title":"Predicting hydrogen atom transfer energy barriers using Gaussian process regression†","authors":"Evgeni Ulanov, Ghulam A. Qadir, Kai Riedmiller, Pascal Friederich and Frauke Gräter","doi":"10.1039/D4DD00174E","DOIUrl":"10.1039/D4DD00174E","url":null,"abstract":"Predicting reaction barriers for arbitrary configurations based on only a limited set of density functional theory (DFT) calculations would render the design of catalysts or the simulation of reactions within complex materials highly efficient. We here propose Gaussian process regression (GPR) as a method of choice if DFT calculations are limited to hundreds or thousands of barrier calculations. For the case of hydrogen atom transfer in proteins, an important reaction in chemistry and biology, we obtain a mean absolute error of 3.23 kcal mol−1 for the range of barriers in the data set using SOAP descriptors and similar values using the marginalized graph kernel. Thus, the two GPR models can robustly estimate reaction barriers within the large chemical and conformational space of proteins. Their predictive power is comparable to a graph neural network-based model, and GPR even outcompetes the latter in the low data regime. We propose GPR as a valuable tool for an approximate but data-efficient model of chemical reactivity in a complex and highly variable environment.","PeriodicalId":72816,"journal":{"name":"Digital discovery","volume":" 2","pages":" 513-522"},"PeriodicalIF":6.2,"publicationDate":"2025-01-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11747964/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143030366","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

AI agents in chemical research: GVIM – an intelligent research assistant system†

IF 6.2

Digital discovery Pub Date : 2025-01-10 DOI: 10.1039/D4DD00398E

Kangyong Ma

{"title":"AI agents in chemical research: GVIM – an intelligent research assistant system†","authors":"Kangyong Ma","doi":"10.1039/D4DD00398E","DOIUrl":"https://doi.org/10.1039/D4DD00398E","url":null,"abstract":"This work utilizes collected and organized instructional data from the field of chemical science to fine-tune mainstream open-source large language models. To objectively evaluate the performance of the fine-tuned models, we have developed an automated scoring system specifically for the chemistry domain, ensuring the accuracy and reliability of the evaluation results. Building on this foundation, we have designed an innovative chemical intelligent assistant system. This system employs the fine-tuned Mistral NeMo model as one of its primary models and features a mechanism for flexibly invoking various advanced models. This design fully considers the rapid iteration characteristics of large language models, ensuring that the system can continuously leverage the latest and most powerful AI capabilities. A major highlight of this system is its deep integration of professional knowledge and requirements from the chemistry field. By incorporating specialized functions such as molecular visualization, SMILES string processing, and chemical literature retrieval, the system significantly enhances its practical value in chemical research and applications. More notably, through carefully designed mechanisms for knowledge accumulation, skill acquisition, performance evaluation, and group collaboration, the system can optimize its professional abilities and interaction quality to a certain extent.","PeriodicalId":72816,"journal":{"name":"Digital discovery","volume":" 2","pages":" 355-375"},"PeriodicalIF":6.2,"publicationDate":"2025-01-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://pubs.rsc.org/en/content/articlepdf/2025/dd/d4dd00398e?page=search","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143396423","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Active learning driven prioritisation of compounds from on-demand libraries targeting the SARS-CoV-2 main protease† 主动学习驱动了针对SARS-CoV-2主要蛋白酶的按需文库中化合物的优先顺序。

IF 6.2

Digital discovery Pub Date : 2025-01-08 DOI: 10.1039/D4DD00343H

Ben Cree, Mateusz K. Bieniek, Siddique Amin, Akane Kawamura and Daniel J. Cole

{"title":"Active learning driven prioritisation of compounds from on-demand libraries targeting the SARS-CoV-2 main protease†","authors":"Ben Cree, Mateusz K. Bieniek, Siddique Amin, Akane Kawamura and Daniel J. Cole","doi":"10.1039/D4DD00343H","DOIUrl":"10.1039/D4DD00343H","url":null,"abstract":"FEgrow is an open-source software package for building congeneric series of compounds in protein binding pockets. For a given ligand core and receptor structure, it employs hybrid machine learning/molecular mechanics potential energy functions to optimise the bioactive conformers of supplied linkers and functional groups. Here, we introduce significant new functionality to automate, parallelise and accelerate the building and scoring of compound suggestions, such that it can be used for automated de novo design. We interface the workflow with active learning to improve the efficiency of searching the combinatorial space of possible linkers and functional groups, make use of interactions formed by crystallographic fragments in scoring compound designs, and introduce the option to seed the chemical space with molecules available from on-demand chemical libraries. As a test case, we target the main protease (Mpro) of SARS-CoV-2, identifying several small molecules with high similarity to molecules discovered by the COVID moonshot effort, using only structural information from a fragment screen in a fully automated fashion. Finally, we order and test 19 compound designs, of which three show weak activity in a fluorescence-based Mpro assay, but work is needed to further optimise the prioritisation of compounds for purchase. The FEgrow package and full tutorials demonstrating the active learning workflow are available at https://github.com/cole-group/FEgrow.","PeriodicalId":72816,"journal":{"name":"Digital discovery","volume":" 2","pages":" 438-450"},"PeriodicalIF":6.2,"publicationDate":"2025-01-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11726688/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143017414","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0