[Caice-csse] Spring 2025 AI at AU Forum Presentation Friday (3/21/2025) at 10:00am CT: LLM-Based Optimizers: Reasoning via Elitism

[Gerry Dozier]

Dear CAICE:

There will be a Spring 2025 AI at AU Forum Presentation Tomorrow (3/21/2025) at 10:00am CT in Lowder 127 and on Zoom at https://auburn.zoom.us/j/81706339239<https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Fauburn.zoom.us%2Fj%2F81706339239&data=05%7C02%7Ccaice-csse%40eng.auburn.edu%7C500daf569f9c44d7671f08dd67a39fd1%7Cccb6deedbd294b388979d72780f62d3b%7C0%7C0%7C638780674644234270%7CUnknown%7CTWFpbGZsb3d8eyJFbXB0eU1hcGkiOnRydWUsIlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMiIsIkFOIjoiTWFpbCIsIldUIjoyfQ%3D%3D%7C0%7C%7C%7C&sdata=bPN2X60iANuZfONU21RYAOScS8WMwkpPO3cnTzuFbKE%3D&reserved=0>. This presentation will be recorded and posted at https://eng.auburn.edu/ai-au/forum#spring-2025-forum<https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Feng.auburn.edu%2Fai-au%2Fforum%23spring-2025-forum&data=05%7C02%7Ccaice-csse%40eng.auburn.edu%7C500daf569f9c44d7671f08dd67a39fd1%7Cccb6deedbd294b388979d72780f62d3b%7C0%7C0%7C638780674644254566%7CUnknown%7CTWFpbGZsb3d8eyJFbXB0eU1hcGkiOnRydWUsIlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMiIsIkFOIjoiTWFpbCIsIldUIjoyfQ%3D%3D%7C0%7C%7C%7C&sdata=j%2Bib5HJrLiJbfboB6Xmny0BMwr%2BAJGB6BfEkpOHGUGQ%3D&reserved=0>.

Title: LLM-Based Optimizers: Reasoning via Elitism
Dr. Shuvayan Brahmachary, Products & Technology, Shell Technology Center Bengaluru

Abstract: Large Language Models (LLMs) have shown impressive capabilities in reasoning and problem-solving, making them intriguing candidates for optimization tasks. This talk presents the potential of LLMs as zero-shot optimizers across various complex scenarios, including multi-objective and high-dimensional problems. A novel approach, termed the Language-Model-Based Evolutionary Optimizer (LEO), is introduced, leveraging the unique strengths of LLMs for numerical optimization. Through a series of examples, ranging from benchmark tests to real-world engineering challenges like supersonic nozzle design and windfarm layout, the effectiveness of LEO is demonstrated. Comparisons with traditional optimization methods highlight the competitive performance of LLMs, while also addressing the challenges posed by their creative tendencies. Practical insights and future research directions are discussed to harness the full potential of LLMs in optimization.

Bio: Shuvayan Brahmachary received his PhD in Fluid and Thermal Science stream of Mechanical Engineering Department from Indian Institute of Technology Guwahati in the year 2019. Following his PhD, he worked as a postdoctoral researcher in the Department of Aeronautics and Astronautics, Kyushu University, Japan until 2021. In Kyushu University, he worked on aerodynamic shape optimization for high-speed vehicles and data-driven neural surrogates for faster optimization approaches. This was followed by his postdoctoral fellowship in the Department of Informatics, Technical University of Munich, Germany. At TUM, he worked on differentiable physics-based neural surrogates for spatio-temporal predictions in low-speed flow regimes. Presently, he is working as an AI researcher in Shell, Bangalore. His research interest includes scientific machine learning, computational fluid dynamics and optimization. For more details about the speaker, please see https://shuvayanb.github.io/about/<https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Fshuvayanb.github.io%2Fabout%2F&data=05%7C02%7Ccaice-csse%40eng.auburn.edu%7C500daf569f9c44d7671f08dd67a39fd1%7Cccb6deedbd294b388979d72780f62d3b%7C0%7C0%7C638780674644268567%7CUnknown%7CTWFpbGZsb3d8eyJFbXB0eU1hcGkiOnRydWUsIlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMiIsIkFOIjoiTWFpbCIsIldUIjoyfQ%3D%3D%7C0%7C%7C%7C&sdata=%2FUSOJgtch8zHC2wM3uH%2BNALcC1TAtC3NOoTTruHg3O8%3D&reserved=0>.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.eng.auburn.edu/pipermail/caice-csse/attachments/20250320/d104f28e/attachment.htm>