Qi Zhang
Generative AI
Srinivas Rangarajan and Qi Zhang
Organizers and Session Co-Chairs
Session Synopsis
The integration of Artificial Intelligence (AI) into chemical sciences is revolutionizing research, development, and industrial processes, significantly enhancing our capabilities in various areas, from material discovery and reaction optimization to process operations and business decision-making. Recently, advances in generative AI have shown great potential in driving substantial breakthroughs in innovation and productivity, transforming workflows and solution strategies for complex problems. However, developing generative AI solutions in chemical engineering and related fields presents unique challenges, including the complexity of modeling chemical systems and the necessity for high-quality, domain-specific data. This session will explore the challenges and opportunities of generative AI, providing insights into how to effectively and responsibly harness its transformative power. In particular, we consider the use of generative modeling, large language models, and AI assistants for property prediction, reaction optimization, molecule/material discovery, process design and optimization, supply chain management, scientific discovery in general (e.g., an AI co-scientist), and enhancing educational outcomes in core chemical engineering courses.
Srinivas Rangarajan
Invited Talks
Innovative Applications of Generative AI in Chemical Process and Manufacturing
The Generative Turn in Chemical Engineering: From Surrogate Models to Scientific Autonomy
Swee-Teng Chin
Senior Data Scientist
The Dow Chemical Company
USA
John Kitchin
Department of Chemical Engineering
Carnegie Mellon University
USA
Abstract
Generative AI is quickly becoming a game-changer in chemical processing and manufacturing, offering optimization and predictive tools that go beyond traditional modeling methods. This presentation explores how large language models (LLMs) are being integrated with specialized workflows to tackle complex process data, interpret images for diagnostics, and perform reliability analytics. We will cover methods that use AI-driven representations to increase efficiency in separation processes, speed up root cause analysis, and improve asset reliability through probabilistic modeling, illustrated by real-world case studies from Dow.
Abstract
Generative modeling is enabling a shift from fitting surrogate models to data to creating models that can design, reason, and explore autonomously. This talk will trace that transformation, beginning with surrogate models and deep learning frameworks that emulate first-principles equations, and moving toward generative models that invert them. I will discuss how approaches such as differentiable programming and generative models enable inverse design, process optimization, and discovery. I will illustrate how generative frameworks can couple with remote laboratories and structured data infrastructures to achieve scientific autonomy, systems that can propose, execute, and learn from experiments. The talk will conclude by considering what this “generative turn” could mean for the future of engineering, where design, simulation, and experimentation increasingly merge into continuous, closed-loop discovery.
Short Oral Presentations
Why Molecule-only Generative Models are not Enough: Toward Context-aware AI for Chemical Designs
Jana M. Weber, Delft University of Technology, The Netherlands
Process Systems Engineering Perspectives on Alignment of AI Models and Agents
Andrew Allman, University of Michigan, USA
Why Process Systems Engineering Challenges Current Paradigms of Monolithic Generative AI
Artur M. Schweidtmann, Delft University of Technology, The Netherlands
From Accuracy to Trust: GenAI Reasoning for Trustworthiness and Adaptation in Process Systems Engineering
Nikolaos Passalis and Michael C. Georgiadis, Aristotle University of Thessaloniki, Greece
Challenges and Opportunities for Vision–Language Models in Optimization and Control of Water Treatment
Minghao Han and Xunyuan Yin, Nanyang Technological University, Singapore
Poster Presentations
Making AI/ML an Everyone’s Game: Overcoming the Hype
Iiro Harjunkoski, Aalto University, Finland
Modeling and Control of Nonlinear Systems with Time-varying Characteristics
Ali Cinar, Mudassir Rashid, Mohammad Ahmadasas, Mate Siket, and Mustafa Bilgic
Illinois Institute of Technology, USA and Obuda University, Hungary
Generative AI in Biopharmaceutical Development and Manufacturing
Seongkyu Yoon, University of Massachusetts Lowell, USA
Feature-Space Engineering for Autonomous Plants: Rethinking Data Topology in Physics-Informed Machine Learning
Isaac Severinsen, Bryan Li, and Brent Young, University of Auckland, New Zealand
Challenges and Opportunities in Direct Lithium Extraction from Emerging Feedstocks: A Process Systems Engineering Perspective
Burcu Beykal, University of Connecticut, USA
Valorizing Historical Data Toward Pharmaceutical Process Development Intelligence
Ulderico Di Caprio and Artur M. Schweidtmann, Delft University of Technology, The Netherlands
Hard Constraints in Machine Learning for PSE: Paradigms and Open Questions
Giacomo Lastrucci and Artur M. Schweidtmann, Delft University of Technology, The Netherlands
Physics-informed Machine Learning on Steam-Turbine Vibrations
Charilaos Kazakos and Ioannis Georgiou, HELLENiQ Energy, Greece
Reflections on Translating PSE Research into Commercial Impact in the Age of AI
Michael Short, University of Surrey, UK
Physics-Informed Neural Networks (PINNs) for Chemical Process Simulation, Parameter Estimation, and Real-Time Operation
Sepehr Aarabi Dahej and Christopher L.E. Swartz, McMaster University, Canada
Towards an Autonomous Multi-Agent LLM Framework to Harmonize Techno-Economic and Life Cycle Data for Supply Chain Optimization
Styliani Avraamidou, Meng-Lin Tsai, Parth Brahmbhatt, University of Wisconsin-Madison, USA
