ISSSR 2024 Keynote 1

When Software Models Meet AI: Early Experience on Intelligent Model-Driven Engineering


Abstract


Model-Driven Engineering (MDE) is widely used in safety critical software development. The software model serves as the core of MDE, used for expressing software requirements, architectural design, and behavioral design. Formal verification conducted on the software model to ensure correctness, following which code is generated based on the model. Traditional MDE faces challenges such as low efficiency in model construction, high cost of secure code generation, and limited performance in model verification.

Recently, Large Language Models (LLMs) have emerged with remarkable language proficiency and code understanding & generating capabilities. When software models meet AI, we anticipate that AI will be able to address the bottleneck issues encountered in MDE.

This talk will explore the convergence of intelligent methods with software modeling, code generation, and formal verification. It will also offer insights into the prospects of model-driven development environments leveraging novel intelligent systems, and report some of our recent exploration of these areas.

Speaker


Chunming Hu's avatar
Professor Chunming Hu China

Dean, School of Software

Beihang University


Prof. Chunming Hu, Dean of the School of Software at Beihang University, China, has a B.Eng (CS, 2000) and Ph.D. (CS, 2006) from the School of Computer Science & Engineering at Beihang University. He has been working at Beihang University as a faculty member since 2006. His main research interests include large-scale distributed systems, software engineering, system virtualization, and resource scheduling in cloud/big data systems. He is a Senior Member of the China Computer Federation (CCF) and the System Software Technical Committee at CCF. He has published over 80 academic papers in related research fields. For years, he has been collaborating closely with the industry to develop resource management mechanisms and scheduling methods for internet-scale cloud data centers to deal with the bulk of online/offline jobs. He also works with aeronautic and astronautic industries on developing methods and tools for model-driven safety-critical software systems.