Life Cycle Sustainability Assessment (LCSA) Section at ISIE invite you to attend our webinar series on “Data, Algorithms and Tools for Life Cycle Sustainability Assessment”. The webinar series serve as a dedicated space to showcase the latest development in data availability and interoperability, new algorithms to improve the scalability and transparency of LCA, and tools for lowering the barriers for integrating LCA with other industrial ecology models.
In this upcoming webinar, we are pleased to host Dr. Bu Zhao (University at Albany, USA) and Dr. Gargeya Vunnava (Amazon, USA).
Time: Feb. 26, 2025 (Wednesday), 7-8 am PST
Zoom link: https://ubc.zoom.us/j/67141510128?pwd=kNSMogqvh717Kka2FmrPE4b8J2puab.1
Below, please find the title, abstract, and a short bio of each presenter:
-------------
A Data-Centric Investigation on the Challenges of Machine Learning Methods for Bridging Life Cycle Inventory Data Gaps.
Presenter: Dr. Bu Zhao
Abstract: Life cycle assessment (LCA) is a systematic approach to quantify the environmental impacts of a product system from its entire life cycle. Despite the great success, the inventory data gap has been a fundamental challenge that limits the application of LCA to emerging new processes. Machine learning (ML) methods are among the possible solutions that can mitigate these data gaps in an automated and scalable way. Nonetheless, the performance of existing ML methods is unstable which limits the trustworthiness and generalizability of the models. In this study, we conducted a data-centric investigation to delineate the causes of the unstable performance using a similarity-based ML framework based on Ecoinvent 3.1 unit process (UPR) database. We found that the pattern of imbalance in the data for method development, manifest by the substantial differences in (1) flow & process availability and (2) the order of magnitude of their values, is a major cause of the unstable performance. We also identified the causes due to the challenges with ML method development workflow, particularly, the steps of data preprocessing, and ML model training (e.g., randomness in train-test data splits). In addition, we also tested the proposed ML method on the U.S. Life Cycle Inventory database, where we observed that the generalizability of the method was highly influenced by the database size of the application. To address these issues, we proposed that further research should focus on reducing the barriers in database integration such that both the size and balance of the data for ML method development can be improved.
Bio: Dr. Bu Zhao is an Assistant Professor in the Department of Environmental and Sustainable Engineering at the University at Albany. Prior to this role, he was an Eric and Wendy Schmidt Postdoctoral Fellow at Cornell University and earned his Ph.D. in Environment and Sustainability and Scientific Computing from the University of Michigan, Ann Arbor. Dr. Zhao’s research lies at the intersection of environmental engineering, sustainability, and artificial intelligence, where he employs cutting-edge AI techniques to analyze the complex interactions between environmental impacts and human activities from a systems perspective. His work aims to advance sustainable solutions through data-driven insights and innovative modeling approaches. Dr. Zhao is actively involved in the academic community, serving as the Managing Editor of Resources, Conservation & Recycling and as a Guest Editor for Energy and AI.
Emission Factor Recommendation for Carbon Footprinting with Generative AI
Presenter: Dr. Gargeya Vunnava
Abstract: Accurately quantifying greenhouse gas (GHG) emissions from products and business activities is crucial for organizations to measure their environmental impact and undertake mitigation actions. Life cycle assessment (LCA) is the scientific discipline for measuring GHG emissions associated with each stage of a product or activity, from raw material extraction to disposal. Measuring the emissions outside of a product owner's control is challenging, and practitioners rely on emission factors (EFs) – estimates of GHG emissions per unit of activity – to model and estimate indirect impacts. These EFs come from prior LCA studies and are collated into databases. The current practice of manually finding the appropriate EF to use from databases is time-consuming, error-prone, and requires domain expertise, hindering scalability and accuracy in emissions quantification. We present a novel AI-assisted method that leverages large language models to automatically recommend EFs. Our method parses business activity descriptions and recommends the appropriate EF with a human-interpretable justification. We benchmark our solution across multiple domains and find it achieves state-of-the-art performance in EF recommendation, with an average Precision@1 of 88.4%. By streamlining and automating the EF selection process, our AI-assisted method enables scalable and accurate quantification of GHG emissions, supporting organizations' sustainability initiatives and driving progress toward net-zero emissions targets across industries.
Bio: Dr. Gargeya Vunnava is a Research Scientist in Amazon's World Wide Sustainability organization, where he develops novel computational methods to enable more scalable and data-driven life cycle assessments. He earned his Ph.D. in Environmental Sustainability from Purdue University, specializing in mechanistic process modeling and physical input-output analysis to quantify the environmental and economic impacts of products and industrial systems. His research has been published in leading journals including Energy & Environmental Science, Applied Energy, Journal of Industrial Ecology, and Journal of Cleaner Production. He has also presented his work at prominent computer science conferences such as ACM COMPASS and the NeurIPS Climate Change AI workshop. Dr. Vunnava is passionate about leveraging the latest advances in data science and modeling to drive sustainability solutions.