Search

Because sustainability remains a wicked problem, more sophisticated tools need to be applied to identify better solutions in a more efficient manner and align with the 11th, 12th, and 13th sustainable development goals: sustainable cities and communities, responsible consumption and production, and climate action. To ease the burdens of conducting sustainability studies, especially life cycle assessments (LCA), practitioners may consider integrating large language models (LLM) into LCAs. This emerging application may offer some advantages due to the capability of these models to generate and process text quickly and efficiently, decreasing the time it takes to complete an LCA and increasing the accessibility of LCAs. In this perspective, we assess the ability of LLMs to complete LCA tasks and encourage the LCA community to study the potential strategies for enhancing the integration of LLMs in LCA methodologies and collaborate to develop standards for responsible use. Because of these advantages, LLMs show promise for life cycle inventory data collection and interpreting the life cycle impact assessment. Challenges arise primarily from the inclusion of hallucinations in the content generated by the LLM, which can be mitigated if the LCA practitioner uses prompt engineering techniques. Moreover, the risk that models cannot take responsibility for generated content can be ameliorated by having the LCA practitioner carefully review the LLM output and take responsibility for decisions made based on the generated content. So long as appropriate steps are taken to overcome the challenges and risks of using of LLMs for LCA, the opportunities presented by integrating the generative AI models can streamline the LCA process and result in significant benefits for the LCA practitioner. [Display omitted] • Large language models have yet to be integrated into life cycle assessment tasks. • Large language models may offer significant time savings for life cycle assessments. • Life cycle inventory data can be acquired and processed by large language models. • Hallucinations in model output require careful reviewing of generated content. • Large language models cannot be held accountable for generated content. [ABSTRACT FROM AUTHOR]

• Review of emerging generative artificial intelligence models like large language models. • Trends of GenAI applications in systems design, optimization, and control. • Potential improvements of PSE methods using emerging GenAI. • Challenges like data needs, and ensuring trust in GenAI applications within PSE. This review article explores how emerging generative artificial intelligence (GenAI) models, such as large language models (LLMs), can enhance solution methodologies within process systems engineering (PSE). These cutting-edge GenAI models, particularly foundation models (FMs), which are pre-trained on extensive, general-purpose datasets, offer versatile adaptability for a broad range of tasks, including responding to queries, image generation, and complex decision-making. Given the close relationship between advancements in PSE and developments in computing and systems technologies, exploring the synergy between GenAI and PSE is essential. We begin our discussion with a compact overview of both classic and emerging GenAI models, including FMs, and then dive into their applications within key PSE domains: synthesis and design, optimization and integration, and process monitoring and control. In each domain, we explore how GenAI models could potentially advance PSE methodologies, providing insights and prospects for each area. Furthermore, the article identifies and discusses potential challenges in fully leveraging GenAI within PSE, including multiscale modeling, data requirements, evaluation metrics and benchmarks, and trust and safety, thereby deepening the discourse on effective GenAI integration into systems analysis, design, optimization, operations, monitoring, and control. This paper provides a guide for future research focused on the applications of emerging GenAI in PSE. [ABSTRACT FROM AUTHOR]

[Display omitted] The chemical industry is facing ever-increasing challenges for developing novel products and processes capable of reducing environmental impacts and curbing resource depletion. Yet, the interplay between molecular phenomena and the design of products and processes are often oversimplified. Machine learning stands uniquely positioned to disentangle the complexity of multiscale modeling by leveraging data to navigate the design spaces of multifaceted molecular systems. Herein, we limit our survey of machine learning applications in computational molecular design (CMD) to four elements: property estimation, catalysis, synthesis planning, and design methods. Through this perspective, we aim to offer a roadmap for future work on multiscale modeling that better explores the interplay between nanoscale features and macroscale decisions in product and process design. [ABSTRACT FROM AUTHOR]

• CAMD knowledge-based property models and solution techniques are reviewed. • Recent deep generative models for molecular design are surveyed. • The survey spotlights molecular representations, architectures, and benchmarks. • The direction toward hybrid methods is spotlighted to sidestep current limitations. • Perspective and future directions toward a unified approach are highlighted. The optimal design of compounds through manipulating properties at the molecular level is often the key to considerable scientific advances and improved process systems performance. This paper highlights key trends, challenges, and opportunities underpinning the Computer-Aided Molecular Design (CAMD) problems. A brief review of knowledge-driven property estimation methods and solution techniques, as well as corresponding CAMD tools and applications, are first presented. In view of the computational challenges plaguing knowledge-based methods and techniques, we survey the current state-of-the-art applications of deep learning to molecular design as a fertile approach towards overcoming computational limitations and navigating uncharted territories of the chemical space. The main focus of the survey is given to deep generative modeling of molecules under various deep learning architectures and different molecular representations. Further, the importance of benchmarking and empirical rigor in building deep learning models is spotlighted. The review article also presents a detailed discussion of the current perspectives and challenges of knowledge-based and data-driven CAMD and identifies key areas for future research directions. Special emphasis is on the fertile avenue of hybrid modeling paradigm, in which deep learning approaches are exploited while leveraging the accumulated wealth of knowledge-driven CAMD methods and tools. [ABSTRACT FROM AUTHOR]

• Pretraining on abundant 2D molecular graphs to enhance 3D tasks is explored. • The limitations of 3D molecular conformer generation are addressed. • Enhancements are proposed by expanding and pretraining molecular embeddings. • Chemical information is used to anchor molecular embeddings in chemistry principles. • Our approach yields advancements across evaluation metrics, setting new benchmarks. Predicting stable 3D molecular conformations from 2D molecular graphs is a challenging and resource-intensive task, yet it is critical for various applications, particularly drug design. Density functional theory (DFT) calculations set the standard for molecular conformation generation, yet they are computationally intensive. Deep learning offers more computationally efficient approaches, but struggles to match DFT accuracy, particularly on complex drug-like structures. Additionally, the steep computational demands of assembling 3D molecular datasets constrain the broader adoption of deep learning. This work aims to utilize the abundant 2D molecular graph datasets for pretraining a machine learning model, a step that involves initially training the model on a different task with a wealth of data before fine-tuning it for the target task of 3D conformation generation. We build on GeoMol, an end-to-end graph neural network (GNN) method for predicting atomic 3D structures and torsion angles. We examine the limitations of the GeoMol method and introduce new baselines to enhance molecular graph embeddings. Our computational results show that 2D molecular graph pretraining enhances the quality of generated 3D conformers, yielding a 7.7 % average improvement over state-of-the-art sequential methods. These advancements not only facilitate superior 3D conformation generation but also emphasize the potential of leveraging pretrained graph embeddings to boost performance in 3D chemical tasks with GNNs. [ABSTRACT FROM AUTHOR]