Search for: All records

Abstract Irradiation increases the yield stress and embrittles light water reactor (LWR) pressure vessel steels. In this study, we demonstrate some of the potential benefits and risks of using machine learning models to predict irradiation hardening extrapolated to low flux, high fluence, extended life conditions. The machine learning training data included the Irradiation Variable for lower flux irradiations up to an intermediate fluence, plus the Belgian Reactor 2 and Advanced Test Reactor 1 for very high flux irradiations, up to very high fluence. Notably, the machine learning model predictions for the high fluence, intermediate flux Advanced Test Reactor 2 irradiations are superior to extrapolations of existing hardening models. The successful extrapolations showed that machine learning models are capable of capturing key intermediate flux effects at high fluence. Similar approaches, applied to expanded databases, could be used to predict hardening in LWRs under life-extension conditions. 

Accurate and comprehensive material databases extracted from research papers are crucial for ma- terials science and engineering, but their development requires significant human effort. With large language models (LLMs) transforming the way humans interact with text, LLMs provide an oppor- tunity to revolutionize data extraction. In this study, we demonstrate a simple and efficient method for extracting materials data from full-text research papers leveraging the capabilities of LLMs com- bined with human supervision. This approach is particularly suitable for mid-sized databases and requires minimal to no coding or prior knowledge about the extracted property. It offers high recall and nearly perfect precision in the resulting database. The method is easily adaptable to new and superior language models, ensuring continued utility. We show this by evaluating and comparing its performance on GPT-3 and GPT-3.5/4 (which underlie ChatGPT), as well as free alternatives such as BART and DeBERTaV3. We provide a detailed analysis of the method’s performance in extracting sentences containing bulk modulus data, achieving up to 90% precision at 96% recall, depending on the amount of human effort involved. We further demonstrate the method’s broader effectiveness by developing a database of critical cooling rates for metallic glasses over twice the size of previous human curated databases. 

The Informatics Skunkworks program provides a new framework for engaging undergraduates in research experiences, with a focus on the interface of data science and materials science. The program seeks to provide authentic research, engaged personal learning, and professional development while also being efficient, accessible, and scalable. Initially developed at the University of Wisconsin-Madison, participation continues to grow, with over 90 students engaged in research or training activities during the Fall 2021 semester from 4 institutions. The Skunkworks focuses on reducing barriers to engagement for mentors and students in undergraduate research by replacing bespoke and ad-hoc approaches with efforts and infrastructure that are reusable and scalable, including simplified standardized recruiting methods, online modular training resources, flexible undergraduate accessible software tools, long-term research projects with many similar but distinct components to engage large teams, and support from a learning community. For example, new students have the option to participate in a modular, self-paced, online onboarding curriculum that teaches students the basic skills needed for most data science projects, thereby dramatically reducing the mentor time needed to engage students with limited background in machine learning research. Projects are authentic research challenges that strive to allow for large flexible teams, thereby scaling up their impact from the typical engagement of just one or two students and allowing for extensive peer teaching. Throughout the program, professional development activities are efficiently delivered through standardized materials to teach critical research skills like record keeping, establishing group expectations and dynamics, and networking. These skills are also reinforced at workshop events hosted during the semester, which are effectively delivered online and yield growing impact for modest effort as the community grows. The program has been successfully implemented as evidenced by the last two semesters' evaluation findings through interviews, focus groups, and pre-post surveys. The students reported a positive attitude towards the program. Students' perception about machine learning knowledge and skills and their self-confidence improved after they got involved in the program. The instructors and mentors indicated positive teaching and mentoring experiences, and shared ideas on the further improvement of the program. Building on its early successes the team is continuing to implement evaluation data-driven improvements to the program with the goal of continuing to grow through new collaborations.