Research Opportunity Number: ACEE-03

Project Title: Large Language Model Applications in Environmental Sustainability

Project Summary: The multidisciplinary nature of grand environmental problems and the ever-changing natural and anthropogenic conditions, demands extensive information retrieval and sophisticated investigations for resolution. However, challenges persist with limited data for hard(expensive)-to-measure variables, and remain significant hinder AI implementation in emerging research topics in environmental sustainability, such as decarbonization, resources recovery, and emerging contaminant (e.g., PFAS, microplastics) management. To resolve the problem of limited data, transfer learning, a type of ML technique that facilitates the transfer of knowledge acquired from one domain to another, presents a practical and effective solution. Large language models (LLMs) represent an ideal avenue to development, as environmental systems encompass diverse facets that cannot be adequately captured by a single or a few datasets. In contrast, LLMs possess the capability to process information from various sources, types, and formats. In this internship opportunity, the student(s) will have the opportunity to engage the cutting-edge LLM research that is applied to environment, sustainability, and energy through data collection, model development, and fine-tuning. The overall objective of this project is to develop domain-specific LLM for question-answering tests and implementation. The research will be co-mentored by Professor and Research Scholar.

Student Roles and Responsibilities: Software development, hardware maintenance, control

Additional Considerations: Program, Research, Literature Review, Report
Start and end dates are flexible, but the student should be available essentially full-time (at least 35 hours/week)

Department/Institute: Andlinger Center for Energy and the Environment

Faculty Sponsor: Jason Ren

Participation Dates: 6/17/2024 to 7/28/2024

Stipend Offered: $0

Number of Internships Available: 0-1

Application Deadline: March 15, 2024, midnight eastern daylight time