Conference: NSF Workshop on Hardware-Software Co-design for Neuro-Symbolic Computation

Duke University, Durham NC

Investigators

Abstract

Despite the impressive advancements in artificial intelligence (AI) systems, there still exist significant challenges, such as enabling high-level reasoning in real-world contexts and dynamic compositionality within resource-constrained environments. Neuro-symbolic computation represents the convergence of two prominent AI methodologies: traditional symbolic and neural networks. By integrating symbolic and statistical methods, neuro-symbolic computation unleashes essential features of modularity for explainability and compositionality. However, the approach introduces significant computational complexity. The purpose of this project is to organize a National Science Foundation (NSF) sponsored workshop named "Hardware-Software Co-design for Neuro-Symbolic Computation." The primary objective of this workshop is to gather leading experts in relevant research domains to explore and identify the scientific and technological challenges associated with high-performance, integrated neuro-symbolic AI computing systems, and provide valuable guidance to the broader research community. To ensure comprehensive coverage of the topic, the team will extend invitations to prominent researchers from academia, industry, government laboratories, and funding agencies with diverse backgrounds, including circuit design, computer architecture, design automation, system development, algorithms, and applications. Underrepresented groups will be particularly encouraged to attend the workshop and receive priority in the considerations. This project aims to identify the critical strategies for semiconductor and AI workforce development and technology transfer. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.