Yeonghun Kang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Yeonghun Kang has a total of 167 co-authors. Here are the top co-authors by number of shared publications:
- Jihan Kim
Department of Chemical and Biomolecular Engineering Korea Advanced Institute of Science and Technology (KAIST) Daejeon Republic of Korea
10 shared publications - Hyunsoo Park
Imperial College London
4 shared publications - Berend Smit
École Polytechnique Fédérale de Lausanne (EPFL) Rue de l’Industrie 17
3 shared publications Alán Aspuru-Guzik
Vector Institute for Artificial Intelligence
2 shared publicationsEric S. Isbrandt
Acceleration Consortium, 700 University Ave., Toronto, Canada, M7A 2S4
2 shared publicationsFangshi Du
Department of Computer Science, University of Toronto, 40 St George St., Toronto, Canada, M5S 2E4
2 shared publications
Latest Publications
- Open Access
- 10 June 2026
- Published by American Chemical Society (ACS) in ChemRxiv
El Agente Seguro: An Agent for Chemical Safety
Laboratory safety is becoming increasingly complex as chemical and materials research integrates automation, artificial intelligence, and data-driven experimentation. However, existing safety management approaches rely heavily on manual interpretation or on unconstrained language-model reasoning, which can lead to inconsistent and unreliable outcomes. In this work, we present El Agente Seguro, a graph-based multi-agent framework for safety planning and evaluation in chemical and materials laboratories. The system separates coordination and execution through a central safety director and a dependency-aware safety-solving graph, in which specialized subagents perform structured tasks with schema-constrained data exchange. This design enables controlled and data-driven safety reasoning across multiple layers, including chemical hazards, reaction risks, regulatory constraints, and opera-tional considerations. Benchmark evaluation across seven representative safety tasks demonstrates that El Agente Seguro achieves consistently strong performance relative to frontier general-purpose language models, while also providing more comprehensive and practically useful safety reasoning according to ranking-based evaluations. Beyond benchmark assessment, El Agente Seguro further demonstrates real-world safety capabilities including protocol-level hazard analysis, autonomous procedure modification, dual-use and prohibited chemical enforcement, multimodal SOP generation from laboratory images and experimental procedures, and computationally guided reaction safety prediction through integration with external scientific agents. These results show that structured, dependency-aware multi-agent workflows can improve context-sensitive safety reasoning beyond unconstrained language-model inference alone. Overall, El Agente Seguro establishes a scalable, extensible and practically deployable framework for intelligent laboratory safety into both human-operated and autonomous research environments.
- 10 June 2026
- Published by American Chemical Society (ACS) in ChemRxiv
Open AccessEl Agente Seguro: An Agent for Chemical Safety
Laboratory safety is becoming increasingly complex as chemical and materials research integrates automation, artificial intelligence, and data-driven experimentation. However, existing safety management approaches rely heavily on manual interpretation or on unconstrained language-model reasoning, which can lead to inconsistent and unreliable outcomes. In this work, we present El Agente Seguro, a graph-based multi-agent framework for safety planning and evaluation in chemical and materials laboratories. The system separates coordination and execution through a central safety director and a dependency-aware safety-solving graph, in which specialized subagents perform structured tasks with schema-constrained data exchange. This design enables controlled and data-driven safety reasoning across multiple layers, including chemical hazards, reaction risks, regulatory constraints, and opera-tional considerations. Benchmark evaluation across seven representative safety tasks demonstrates that El Agente Seguro achieves consistently strong performance relative to frontier general-purpose language models, while also providing more comprehensive and practically useful safety reasoning according to ranking-based evaluations. Beyond benchmark assessment, El Agente Seguro further demonstrates real-world safety capabilities including protocol-level hazard analysis, autonomous procedure modification, dual-use and prohibited chemical enforcement, multimodal SOP generation from laboratory images and experimental procedures, and computationally guided reaction safety prediction through integration with external scientific agents. These results show that structured, dependency-aware multi-agent workflows can improve context-sensitive safety reasoning beyond unconstrained language-model inference alone. Overall, El Agente Seguro establishes a scalable, extensible and practically deployable framework for intelligent laboratory safety into both human-operated and autonomous research environments.
- Open Access
- 5 June 2026
- Published by American Chemical Society (ACS) in ChemRxiv
El Agente Seguro: An Agent for Chemical Safety
Laboratory safety is becoming increasingly complex as chemical and materials research integrates automation, artificial intelligence, and data-driven experimentation. However, existing safety management approaches rely heavily on manual interpretation or on unconstrained language-model reasoning, which can lead to inconsistent and unreliable outcomes. In this work, we present El Agente Seguro, a graph-based multi-agent framework for safety planning and evaluation in chemical and materials laboratories. The system separates coordination and execution through a central safety director and a dependency-aware safety-solving graph, in which specialized subagents perform structured tasks with schema-constrained data exchange. This design enables controlled and data-driven safety reasoning across multiple layers, including chemical hazards, reaction risks, regulatory constraints, and opera-tional considerations. Benchmark evaluation across seven representative safety tasks demonstrates that El Agente Seguro achieves consistently strong performance relative to frontier general-purpose language models, while also providing more comprehensive and practically useful safety reasoning according to ranking-based evaluations. Beyond benchmark assessment, El Agente Seguro further demonstrates real-world safety capabilities including protocol-level hazard analysis, autonomous procedure modification, dual-use and prohibited chemical enforcement, multimodal SOP generation from laboratory images and experimental procedures, and computationally guided reaction safety prediction through integration with external scientific agents. These results show that structured, dependency-aware multi-agent workflows can improve context-sensitive safety reasoning beyond unconstrained language-model inference alone. Overall, El Agente Seguro establishes a scalable, extensible and practically deployable framework for intelligent laboratory safety into both human-operated and autonomous research environments.
- 5 June 2026
- Published by American Chemical Society (ACS) in ChemRxiv
Open AccessEl Agente Seguro: An Agent for Chemical Safety
Laboratory safety is becoming increasingly complex as chemical and materials research integrates automation, artificial intelligence, and data-driven experimentation. However, existing safety management approaches rely heavily on manual interpretation or on unconstrained language-model reasoning, which can lead to inconsistent and unreliable outcomes. In this work, we present El Agente Seguro, a graph-based multi-agent framework for safety planning and evaluation in chemical and materials laboratories. The system separates coordination and execution through a central safety director and a dependency-aware safety-solving graph, in which specialized subagents perform structured tasks with schema-constrained data exchange. This design enables controlled and data-driven safety reasoning across multiple layers, including chemical hazards, reaction risks, regulatory constraints, and opera-tional considerations. Benchmark evaluation across seven representative safety tasks demonstrates that El Agente Seguro achieves consistently strong performance relative to frontier general-purpose language models, while also providing more comprehensive and practically useful safety reasoning according to ranking-based evaluations. Beyond benchmark assessment, El Agente Seguro further demonstrates real-world safety capabilities including protocol-level hazard analysis, autonomous procedure modification, dual-use and prohibited chemical enforcement, multimodal SOP generation from laboratory images and experimental procedures, and computationally guided reaction safety prediction through integration with external scientific agents. These results show that structured, dependency-aware multi-agent workflows can improve context-sensitive safety reasoning beyond unconstrained language-model inference alone. Overall, El Agente Seguro establishes a scalable, extensible and practically deployable framework for intelligent laboratory safety into both human-operated and autonomous research environments.
- Open Access
- 22 June 2025
- Published by American Chemical Society (ACS) in ChemRxiv
Bayesian Optimization Hackathon for Chemistry and Materials
The Acceleration Consortium and Merck KGaA hosted a 2-day virtual hackathon on March 27- 28, 2024, bringing together scientists to explore, collaborate, and innovate in the field of Bayesian optimization for the physical sciences. Participants were encouraged to select or develop Bayesian optimization algorithms, apply them to benchmarking tasks, design new benchmarks, create instructional tutorials, and describe real-world applications. With over 100 participants across 69 academic, industry, and government organizations located in 59 cities, 19 countries, and 4 continents, this was a global event. The outputs from this event, including developed algorithms, benchmarks, and tutorials, will serve as valuable resources for the research community, in addition to the new skills learned and connections formed. Released projects and general information are available at https://ac-bo-hackathon.github.io/ and other locations linked from individual project pages. This event demonstrates the potential of community-driven research efforts to accelerate advances in Bayesian optimization in chemistry and materials science.
- 22 June 2025
- Published by American Chemical Society (ACS) in ChemRxiv
Open AccessBayesian Optimization Hackathon for Chemistry and Materials
The Acceleration Consortium and Merck KGaA hosted a 2-day virtual hackathon on March 27- 28, 2024, bringing together scientists to explore, collaborate, and innovate in the field of Bayesian optimization for the physical sciences. Participants were encouraged to select or develop Bayesian optimization algorithms, apply them to benchmarking tasks, design new benchmarks, create instructional tutorials, and describe real-world applications. With over 100 participants across 69 academic, industry, and government organizations located in 59 cities, 19 countries, and 4 continents, this was a global event. The outputs from this event, including developed algorithms, benchmarks, and tutorials, will serve as valuable resources for the research community, in addition to the new skills learned and connections formed. Released projects and general information are available at https://ac-bo-hackathon.github.io/ and other locations linked from individual project pages. This event demonstrates the potential of community-driven research efforts to accelerate advances in Bayesian optimization in chemistry and materials science.
- 16 April 2025
- Published by American Chemical Society (ACS) in ACS Nano
- Vol. 19 (16) , 16175-16187
Molecular Sensitizer-Loaded Monolayer Organic Semiconductors for High-Performance H2S Sensors
Monolayer organic semiconductors offer high gas sensitivity due to their exposed active channels and absence of grain boundaries, allowing them to interact directly with the analytes. However, such single-component organic sensors are typically lacking in selective adsorption sites, leading to practical hindrances such as limited chemical selectivity and slow recovery. In this study, we demonstrate a highly crystalline monolayer C10-DNTT semiconductor film sensitized with F4-TCNQ (F4-TCNQ/C10-DNTT), fabricated using a one-step solution-shearing codeposition method, which overcomes the aforementioned challenges to serve as a high-performance chemical gas sensor. Compared to pristine C10-DNTT, the monolayer F4-TCNQ/C10-DNTT films showed an enhanced response to hydrogen sulfide (H2S), an industrial pollutant and an important biomarker for respiratory disorders, as the sensitizer molecules promoted electronic interactions at the gas-organic semiconductor interface. Specifically, this resulted in a 5.3-fold increase in sensitivity to 5 ppm of H2S, with high selectivity and improved recovery compared to pristine C10-DNTT sensors. The precisely defined two-dimensional (2D) structure of the F4-TCNQ/C10-DNTT films enabled the investigation of layer-dependent sensing characteristics, reinforcing the significance of the monolayer configuration for achieving highly sensitive and selective H2S sensing. This research provides an effective strategy for designing high-performance organic sensing devices and highlights the importance of layer precision in sensor response, contributing to the development of more efficient and reliable organic chemical sensors in the future.
- 16 April 2025
- Published by American Chemical Society (ACS) in ACS Nano
- Vol. 19 (16) , 16175-16187
Molecular Sensitizer-Loaded Monolayer Organic Semiconductors for High-Performance H2S Sensors
Monolayer organic semiconductors offer high gas sensitivity due to their exposed active channels and absence of grain boundaries, allowing them to interact directly with the analytes. However, such single-component organic sensors are typically lacking in selective adsorption sites, leading to practical hindrances such as limited chemical selectivity and slow recovery. In this study, we demonstrate a highly crystalline monolayer C10-DNTT semiconductor film sensitized with F4-TCNQ (F4-TCNQ/C10-DNTT), fabricated using a one-step solution-shearing codeposition method, which overcomes the aforementioned challenges to serve as a high-performance chemical gas sensor. Compared to pristine C10-DNTT, the monolayer F4-TCNQ/C10-DNTT films showed an enhanced response to hydrogen sulfide (H2S), an industrial pollutant and an important biomarker for respiratory disorders, as the sensitizer molecules promoted electronic interactions at the gas-organic semiconductor interface. Specifically, this resulted in a 5.3-fold increase in sensitivity to 5 ppm of H2S, with high selectivity and improved recovery compared to pristine C10-DNTT sensors. The precisely defined two-dimensional (2D) structure of the F4-TCNQ/C10-DNTT films enabled the investigation of layer-dependent sensing characteristics, reinforcing the significance of the monolayer configuration for achieving highly sensitive and selective H2S sensing. This research provides an effective strategy for designing high-performance organic sensing devices and highlights the importance of layer precision in sensor response, contributing to the development of more efficient and reliable organic chemical sensors in the future.
- Top Cited Papers
- 21 January 2025
- Published by American Chemical Society (ACS) in Journal of the American Chemical Society
- Vol. 147 (5) , 3943-3958
Harnessing Large Language Models to Collect and Analyze Metal–Organic Framework Property Data Set
This research focused on the efficient collection of experimental metal-organic framework (MOF) data from scientific literature to address the challenges of accessing hard-to-find data and improving the quality of information available for machine learning studies in materials science. Utilizing a chain of advanced large language models (LLMs), we developed a systematic approach to extract and organize MOF data into a structured format. Our methodology successfully compiled information from more than 40,000 research articles, creating a comprehensive and ready-to-use data set. Specifically, data regarding MOF synthesis conditions and properties were extracted from both tables and text and then analyzed. Subsequently, we utilized the curated database to analyze the relationships between synthesis conditions, properties, and structure. Through machine learning, we identified the existence of a gap between simulation data and experimental data, and further analysis revealed the factors contributing to this discrepancy. Additionally, we leveraged the extracted synthesis condition data to develop a synthesis condition recommender system. This system suggests optimal synthesis conditions based on the provided precursors, offering a practical tool to refine synthesis strategies. This underscores the importance of experimental datasets in advancing MOF research.
- 21 January 2025
- Published by American Chemical Society (ACS) in Journal of the American Chemical Society
- Vol. 147 (5) , 3943-3958
Top Cited PapersHarnessing Large Language Models to Collect and Analyze Metal–Organic Framework Property Data Set
This research focused on the efficient collection of experimental metal-organic framework (MOF) data from scientific literature to address the challenges of accessing hard-to-find data and improving the quality of information available for machine learning studies in materials science. Utilizing a chain of advanced large language models (LLMs), we developed a systematic approach to extract and organize MOF data into a structured format. Our methodology successfully compiled information from more than 40,000 research articles, creating a comprehensive and ready-to-use data set. Specifically, data regarding MOF synthesis conditions and properties were extracted from both tables and text and then analyzed. Subsequently, we utilized the curated database to analyze the relationships between synthesis conditions, properties, and structure. Through machine learning, we identified the existence of a gap between simulation data and experimental data, and further analysis revealed the factors contributing to this discrepancy. Additionally, we leveraged the extracted synthesis condition data to develop a synthesis condition recommender system. This system suggests optimal synthesis conditions based on the provided precursors, offering a practical tool to refine synthesis strategies. This underscores the importance of experimental datasets in advancing MOF research.
- Open Access
- 12 September 2024
- Published by American Chemical Society (ACS) in JACS Au
- Vol. 4 (10) , 3727-3743
From Data to Discovery: Recent Trends of Machine Learning in Metal–Organic Frameworks
Renowned for their high porosity and structural diversity, metal-organic frameworks (MOFs) are a promising class of materials for a wide range of applications. In recent decades, with the development of large-scale databases, the MOF community has witnessed innovations brought by data-driven machine learning methods, which have enabled a deeper understanding of the chemical nature of MOFs and led to the development of novel structures. Notably, machine learning is continuously and rapidly advancing as new methodologies, architectures, and data representations are actively being investigated, and their implementation in materials discovery is vigorously pursued. Under these circumstances, it is important to closely monitor recent research trends and identify the technologies that are being introduced. In this Perspective, we focus on emerging trends of machine learning within the field of MOFs, the challenges they face, and the future directions of their development.
- 12 September 2024
- Published by American Chemical Society (ACS) in JACS Au
- Vol. 4 (10) , 3727-3743
Open AccessFrom Data to Discovery: Recent Trends of Machine Learning in Metal–Organic Frameworks
Renowned for their high porosity and structural diversity, metal-organic frameworks (MOFs) are a promising class of materials for a wide range of applications. In recent decades, with the development of large-scale databases, the MOF community has witnessed innovations brought by data-driven machine learning methods, which have enabled a deeper understanding of the chemical nature of MOFs and led to the development of novel structures. Notably, machine learning is continuously and rapidly advancing as new methodologies, architectures, and data representations are actively being investigated, and their implementation in materials discovery is vigorously pursued. Under these circumstances, it is important to closely monitor recent research trends and identify the technologies that are being introduced. In this Perspective, we focus on emerging trends of machine learning within the field of MOFs, the challenges they face, and the future directions of their development.
- Top Cited PapersOpen Access
- 3 June 2024
- Published by Springer Nature in Nature Communications
- Vol. 15 (1) , 1-13
ChatMOF: an artificial intelligence system for predicting and generating metal-organic frameworks using large language models
ChatMOF is an artificial intelligence (AI) system that is built to predict and generate metal-organic frameworks (MOFs). By leveraging a large-scale language model (GPT-4, GPT-3.5-turbo, and GPT-3.5-turbo-16k), ChatMOF extracts key details from textual inputs and delivers appropriate responses, thus eliminating the necessity for rigid and formal structured queries. The system is comprised of three core components (i.e., an agent, a toolkit, and an evaluator) and it forms a robust pipeline that manages a variety of tasks, including data retrieval, property prediction, and structure generations. ChatMOF shows high accuracy rates of 96.9% for searching, 95.7% for predicting, and 87.5% for generating tasks with GPT-4. Additionally, it successfully creates materials with user-desired properties from natural language. The study further explores the merits and constraints of utilizing large language models (LLMs) in combination with database and machine learning in material sciences and showcases its transformative potential for future advancements.
- 3 June 2024
- Published by Springer Nature in Nature Communications
- Vol. 15 (1) , 1-13
Top Cited PapersOpen AccessChatMOF: an artificial intelligence system for predicting and generating metal-organic frameworks using large language models
ChatMOF is an artificial intelligence (AI) system that is built to predict and generate metal-organic frameworks (MOFs). By leveraging a large-scale language model (GPT-4, GPT-3.5-turbo, and GPT-3.5-turbo-16k), ChatMOF extracts key details from textual inputs and delivers appropriate responses, thus eliminating the necessity for rigid and formal structured queries. The system is comprised of three core components (i.e., an agent, a toolkit, and an evaluator) and it forms a robust pipeline that manages a variety of tasks, including data retrieval, property prediction, and structure generations. ChatMOF shows high accuracy rates of 96.9% for searching, 95.7% for predicting, and 87.5% for generating tasks with GPT-4. Additionally, it successfully creates materials with user-desired properties from natural language. The study further explores the merits and constraints of utilizing large language models (LLMs) in combination with database and machine learning in material sciences and showcases its transformative potential for future advancements.
- 19 March 2024
- Published by American Chemical Society (ACS) in ACS Applied Materials & Interfaces
- Vol. 16 (13) , 16853-16860
Multimodal Transformer for Property Prediction in Polymers
In this work, we designed a multimodal transformer that combines both the Simplified Molecular Input Line Entry System (SMILES) and molecular graph representations to enhance the prediction of polymer properties. Three models with different embeddings (SMILES, SMILES + monomer, and SMILES + dimer) were employed to assess the performance of incorporating multimodal features into transformer architectures. Fine-tuning results across five properties (i.e., density, glass-transition temperature (Tg), melting temperature (Tm), volume resistivity, and conductivity) demonstrated that the multimodal transformer with both the SMILES and the dimer configuration as inputs outperformed the transformer using only SMILES across all five properties. Furthermore, our model facilitates in-depth analysis by examining attention scores, providing deeper insights into the relationship between the deep learning model and the polymer attributes. We believe that our work, shedding light on the potential of multimodal transformers in predicting polymer properties, paves a new direction for understanding and refining polymer properties.
- 19 March 2024
- Published by American Chemical Society (ACS) in ACS Applied Materials & Interfaces
- Vol. 16 (13) , 16853-16860
Multimodal Transformer for Property Prediction in Polymers
In this work, we designed a multimodal transformer that combines both the Simplified Molecular Input Line Entry System (SMILES) and molecular graph representations to enhance the prediction of polymer properties. Three models with different embeddings (SMILES, SMILES + monomer, and SMILES + dimer) were employed to assess the performance of incorporating multimodal features into transformer architectures. Fine-tuning results across five properties (i.e., density, glass-transition temperature (Tg), melting temperature (Tm), volume resistivity, and conductivity) demonstrated that the multimodal transformer with both the SMILES and the dimer configuration as inputs outperformed the transformer using only SMILES across all five properties. Furthermore, our model facilitates in-depth analysis by examining attention scores, providing deeper insights into the relationship between the deep learning model and the polymer attributes. We believe that our work, shedding light on the potential of multimodal transformers in predicting polymer properties, paves a new direction for understanding and refining polymer properties.
- 20 November 2023
- Published by American Chemical Society (ACS) in ACS Applied Materials & Interfaces
- Vol. 15 (48) , 56375-56385
Enhancing Structure–Property Relationships in Porous Materials through Transfer Learning and Cross-Material Few-Shot Learning
Porous materials have emerged as promising solutions for a wide range of energy and environmental applications. However, the asymmetric development in the field of metal-organic frameworks (MOFs) has led to a data imbalance when it comes to MOFs versus other porous materials such as covalent organic frameworks (COFs), porous polymer networks (PPNs), and zeolites. To address this issue, we introduce PMTransformer (Porous Material Transformer), a multimodal Transformer model pretrained on a vast data set of 1.9 million hypothetical porous materials, including metal-organic frameworks, covalent organic frameworks, porous polymer networks, and zeolites. PMTransformer showcases remarkable transfer learning capabilities, resulting in state-of-the-art performance in predicting various porous material properties. To address the challenge of asymmetric data aggregation, we propose cross-material few-shot learning, which leverages the synergistic effect among different porous material classes to enhance the fine-tuning performance with a limited number of examples. As a proof of concept, we demonstrate its effectiveness in predicting band gap values of COFs using the available MOF data in the training set. Moreover, we established cross-material relationships in porous materials by predicting the unseen properties of other classes of porous materials. Our approach presents a new pathway for understanding the underlying relationships among various classes of porous materials, paving the way toward a more comprehensive understanding and design of porous materials.
- 20 November 2023
- Published by American Chemical Society (ACS) in ACS Applied Materials & Interfaces
- Vol. 15 (48) , 56375-56385
Enhancing Structure–Property Relationships in Porous Materials through Transfer Learning and Cross-Material Few-Shot Learning
Porous materials have emerged as promising solutions for a wide range of energy and environmental applications. However, the asymmetric development in the field of metal-organic frameworks (MOFs) has led to a data imbalance when it comes to MOFs versus other porous materials such as covalent organic frameworks (COFs), porous polymer networks (PPNs), and zeolites. To address this issue, we introduce PMTransformer (Porous Material Transformer), a multimodal Transformer model pretrained on a vast data set of 1.9 million hypothetical porous materials, including metal-organic frameworks, covalent organic frameworks, porous polymer networks, and zeolites. PMTransformer showcases remarkable transfer learning capabilities, resulting in state-of-the-art performance in predicting various porous material properties. To address the challenge of asymmetric data aggregation, we propose cross-material few-shot learning, which leverages the synergistic effect among different porous material classes to enhance the fine-tuning performance with a limited number of examples. As a proof of concept, we demonstrate its effectiveness in predicting band gap values of COFs using the available MOF data in the training set. Moreover, we established cross-material relationships in porous materials by predicting the unseen properties of other classes of porous materials. Our approach presents a new pathway for understanding the underlying relationships among various classes of porous materials, paving the way toward a more comprehensive understanding and design of porous materials.
- Top Cited Papers
- 13 March 2023
- Published by Springer Nature in Nature Machine Intelligence
- Vol. 5 (3) , 309-318
A multi-modal pre-training transformer for universal transfer learning in metal–organic frameworks
- 13 March 2023
- Published by Springer Nature in Nature Machine Intelligence
- Vol. 5 (3) , 309-318
Top Cited PapersA multi-modal pre-training transformer for universal transfer learning in metal–organic frameworks