CIME4R: Exploring iterative, AI-guided chemical reaction optimization campaigns in their parameter space DOI Creative Commons
Christina Humer, Rachel Nicholls, Henry Heberle

et al.

Published: Dec. 22, 2023

Chemical reaction optimization (RO) is an iterative process that results in large and high-dimensional datasets. Current tools only allow for limited analysis understanding of parameter spaces, making it hard scientists to review or follow changes throughout the process. With recent emergence using artificial intelligence (AI) models aid RO, another level complexity was added. It critical assess quality a model’s prediction understand its decision human-AI collaboration trust calibration. To regard, we propose CIME4R—an open-source interactive web application analyzing RO data AI predictions. CIME4R supports users (i) comprehending space, (ii) investigating how developed over iterations, (iii) identifying factors reaction, (iv) model This aids informed decisions during helps them retrospect, especially realm AI-guided RO. decision-making through interaction between humans by combining strengths expert experience high computational precision. We tested together with domain experts verified usefulness three case studies. were able produce valuable insights from past campaigns make on which experiments perform next. believe beginning community project improves workflow working domain.

Language: Английский

Role of the human-in-the-loop in emerging self-driving laboratories for heterogeneous catalysis DOI
Christoph Scheurer, Karsten Reuter

Nature Catalysis, Journal Year: 2025, Volume and Issue: 8(1), P. 13 - 19

Published: Jan. 29, 2025

Language: Английский

Citations

4
Bayesian Optimization as a Sustainable Strategy for Early-Stage Process Development? A Case Study of Cu-Catalyzed C–N Coupling of Sterically Hindered Pyrazines DOI
Elena Braconi, Edouard Godineau

ACS Sustainable Chemistry & Engineering, Journal Year: 2023, Volume and Issue: 11(28), P. 10545 - 10554

Published: July 7, 2023

Bayesian optimization is a powerful machine learning technique that particularly well-suited for optimizing chemical reactions in the early stages of process development. It can efficiently explore vast reaction spaces and predict high-yielding conditions by evaluating only small number experiments. In this report, we investigated potential as tool to enhance sustainability synthesis. Specifically, focused on real-world early-stage development example: C–N coupling sterically encumbered bromo-pyrazines with amines. Our objective was identify sustainable utilize Earth-abundant copper catalysts non-hazardous solvents. We used optimizers various acquisition functions. assessed their performance identified key features affecting results. The optimized enabled synthesis range pyrazines pyridines moderate excellent yields.

Language: Английский

Citations

23
Accelerated exploration of heterogeneous CO2 hydrogenation catalysts by Bayesian-optimized high-throughput and automated experimentation DOI Creative Commons
Adrián Ramírez, Erwin Lam,

Daniel Pacheco Gutiérrez

et al.

Chem Catalysis, Journal Year: 2024, Volume and Issue: 4(2), P. 100888 - 100888

Published: Jan. 21, 2024

Language: Английский

Citations

16
Research Trend Analysis in the Field of Self-Driving Labs Using Network Analysis and Topic Modeling DOI Creative Commons

Woojun Jung,

Insung Hwang,

Keuntae Cho

et al.

Systems, Journal Year: 2025, Volume and Issue: 13(4), P. 253 - 253

Published: April 3, 2025

A self-driving lab (SDL) system that automates experimental design, data collection, and analysis using robotics artificial intelligence (AI) technologies. Its significance has grown substantially in recent years. This study analyzes the overall SDL research trends, examines changes specific topics, visualizes relational structure between authors to identify key contributors, extracts major themes from extensive texts highlight essential content. To achieve these objectives, trend analysis, network topic modeling were conducted on 352 papers collected Web of Science 2004 2023. ensure validity results, a correlation matrix was also performed. The results revealed three findings. First, surged since 2019, driven by advancements AI technologies, reflecting heightened activity this field. Second, modern scientific is advancing with focus data-driven approaches, applications, optimization through utilization SDLs. Third, exhibits interdisciplinary convergence, encompassing material optimization, biological processes, predictive algorithms. underscores growing importance SDLs as tool across diverse academic disciplines provides practical insights into sustainable future directions strategic approaches.

Language: Английский

Citations

0
BayBE: a Bayesian Back End for experimental planning in the low-to-no-data regime DOI Creative Commons
Martin Fitzner, Adrian Šošić, Alexander V. Hopp

et al.

Digital Discovery, Journal Year: 2025, Volume and Issue: unknown

Published: Jan. 1, 2025

The Bayesian Back End (BayBE) has a range of advanced features enabling scientists to go beyond the basic optimization loop and readily tackle real world experimental campaigns.

Language: Английский

Citations

0
Combining Bayesian optimization and automation to simultaneously optimize reaction conditions and routes DOI Creative Commons
Oliver Schilter,

Daniel Pacheco Gutiérrez,

Linnea M. Folkmann

et al.

Chemical Science, Journal Year: 2024, Volume and Issue: 15(20), P. 7732 - 7741

Published: Jan. 1, 2024

Combining a cloud-based Bayesian optimization platform with robotic synthesis accelerated the discovery of high conversion iodination terminal alkyne reactions in large search space over 12 000 possible 23 experiments.

Language: Английский

Citations

3
The Lab of the Future: Self-Driving Labs for Molecule Discovery DOI
Sean Ekins

GEN Biotechnology, Journal Year: 2024, Volume and Issue: 3(2), P. 83 - 86

Published: April 1, 2024

Self-driving laboratories (SDLs) are at the intersection of robotics, artificial intelligence, and laboratory automation. From materials design, small molecule discovery, to synthetic biology, SDLs have infiltrated a diverse range applications across academia industry. The following Perspective (inspired by Future Labs, workshop held NC State University in early 2024) describes how might become an integrated component future research processes generally applicable many development areas for increased innovation discovery.

Language: Английский

Citations

0
CIME4R: Exploring iterative, AI-guided chemical reaction optimization campaigns in their parameter space DOI Creative Commons

Christina Humer,

Rachel Nicholls, Henry Heberle

et al.

Journal of Cheminformatics, Journal Year: 2024, Volume and Issue: 16(1)

Published: May 10, 2024

Abstract Chemical reaction optimization (RO) is an iterative process that results in large, high-dimensional datasets. Current tools allow for only limited analysis and understanding of parameter spaces, making it hard scientists to review or follow changes throughout the process. With recent emergence using artificial intelligence (AI) models aid RO, another level complexity has been added. Helping assess quality a model’s prediction understand its decision critical supporting human-AI collaboration trust calibration. To address this, we propose CIME4R—an open-source interactive web application analyzing RO data AI predictions. CIME4R supports users ( i ) comprehending space, ii investigating how developed over iterations, iii identifying factors reaction, iv model This facilitates informed decisions during helps completed process, especially AI-guided RO. aids decision-making through interaction between humans by combining strengths expert experience high computational precision. We tested with domain experts verified usefulness three case studies. Using were able produce valuable insights from past campaigns make on which experiments perform next. believe beginning community project potential improve workflow working domain. Scientific contribution best our knowledge, first tailored peculiar requirements campaigns. Due growing use special focus facilitating models. evaluated verify practical usefulness.

Language: Английский

Citations

0
Highlights from the 56th Bürgenstock Conference on Stereochemistry 2023 DOI Creative Commons
Marc Reid, Christopher J. Teskey

Chemical Science, Journal Year: 2023, Volume and Issue: 14(35), P. 9244 - 9247

Published: Jan. 1, 2023

Herein, we share an overview of the scientific highlights from speakers at latest edition longstanding Bürgenstock Conference.

Language: Английский

Citations

0
CIME4R: Exploring iterative, AI-guided chemical reaction optimization campaigns in their parameter space DOI Creative Commons
Christina Humer, Rachel Nicholls, Henry Heberle

et al.

Published: Dec. 22, 2023

Chemical reaction optimization (RO) is an iterative process that results in large and high-dimensional datasets. Current tools only allow for limited analysis understanding of parameter spaces, making it hard scientists to review or follow changes throughout the process. With recent emergence using artificial intelligence (AI) models aid RO, another level complexity was added. It critical assess quality a model’s prediction understand its decision human-AI collaboration trust calibration. To regard, we propose CIME4R—an open-source interactive web application analyzing RO data AI predictions. CIME4R supports users (i) comprehending space, (ii) investigating how developed over iterations, (iii) identifying factors reaction, (iv) model This aids informed decisions during helps them retrospect, especially realm AI-guided RO. decision-making through interaction between humans by combining strengths expert experience high computational precision. We tested together with domain experts verified usefulness three case studies. were able produce valuable insights from past campaigns make on which experiments perform next. believe beginning community project improves workflow working domain.

Language: Английский

Citations

0