Project SCALE: HI/AP Machine learning-based optimization of materials for microelectronics POCs Program Engineering First Time Researcher (FTR) Fellowship Term Spring 2024 Status Closed Research Area Microelectronics, heterogeneous integration/ advanced packaging, machine learning, active learning, generative AI, computational MSE Description This project is one of several SCALE FTR 2024 research projects. SCALE is a Purdue microelectronics program that is sponsored by the Department of Defense and open to US Citizens. In SCALE, you will have opportunities for continuing research (paid or for credit), networking and skill building, as well as industry and government internships throughout your time at Purdue. By applying to this project, you can be considered for any of the SCALE FTR projects with one application. See https://nanohub.org/groups/scale/research_opportunities/research_purdue/ftr2024 to view all of the SCALE FTR research projects for Spring 2024. New materials with desired properties are critical for microelectronics. For example, new alloys with stringent thermo-mechanical performance are needed to replace lead-based solders. Computational methods have become an integral part of material discovery and optimization. These tools can accelerate the introduction of new materials and enable the co-design of materials and devices. Various methods such as predictive modeling, high-throughput simulations, and material informatics have accelerated the discovery of optimal materials reducing the number of costly and time-consuming experiments. Recent advancements in artificial intelligence (AI) have especially garnered attention from the scientific community as they provide solutions beyond traditional modeling approaches. In this project, generative AI and active learning will be utilized to optimize material composition for microelectronics with a focus on properties of interest for heterogeneous integration and advanced packaging. Generative AI models such as diffusion models will be trained on existing experimental data to generate composition-dependent solder microstructure. In addition, active learning methods along with molecular simulations will be utilized to discover the optimal composition of alloys for packaging and other applications. Students will be trained for the entire process of developing, fine-tuning, and applying machine learning for computational material science. Supervisor Gabriella Maria Schr Torres Mentor Alejandro H Strachan Type of work required Students will be trained for the entire process of developing, fine-tuning, and applying machine learning for computational material science. Websites https://nanohub.org/groups/scale/research_opportunities/research_purdue/ftr2024 Qualifications US Citizenship, GPA>2.80 required We are looking for motivated and hard-working undergraduates. All applicants should be capable of working independently while effectively communicating within a team setting. Preferred Majors: Computer science, Mechanical Engineering, Materials Science and Engineering, Electrical Engineering, Physics. Required Experience and Skills: Basic computer coding experience with any language, python preferred Desired skills: Experience with Machine Learning Credit Hours 0 Weekly Hours 10 (estimated)