📄 Paper accepted to ICSE 2026, the premier conference in software engineering!
SmartC2Rust: Iterative, Feedback-Driven C-to-Rust Translation via Large Language Models for Safety and Equivalence

A paper by Momoko Shiraishi (4th-year Ph.D. student) from the Shinagawa Laboratory has been accepted to the IEEE/ACM International Conference on Software Engineering (ICSE 2026), which will be held in April 2026 in Rio de Janeiro, Brazil.
The paper proposes SmartC2Rust, a novel approach that automatically translates large-scale C programs (thousands of lines) into Rust using Large Language Models (LLMs). Existing methods often suffer from compilation failures, residual unsafe blocks, and loss of semantic equivalence with the original C code due to the limited context window and output inconsistency of LLMs. SmartC2Rust overcomes these issues by dividing the original C source into manageable units and iteratively refining the translation with LLM-based feedback, enabling the generation of semantically equivalent and safe Rust code.
By enabling automatic translation from legacy C programs to Rust, this research aims to fundamentally solve the memory safety issues arising by using the C language.
This research was conducted as an international collaboration with Associate Professor Yinzhi Cao from Johns Hopkins University. Professor Cao’s group conducts leading research in program analysis and software security, and the collaboration focused on ensuring safety and automation in language translation.
ICSE is the premier conference in software engineering, ranked A* by the CORE Conference Rankings and holding an h5-index of 98, making it the No.1 venue in Software Systems according to Google Scholar Metrics. In Cycle 1 of ICSE 2026, 660 papers were submitted. Among them, 60 were directly accepted, and 107 were invited for major revisions, of which 100 were eventually accepted. The final acceptance rate was 24.2% (160/660).
ICSE Data
- Google Scholar Metrics: h5-index: 98 (#1 Software Systems)
- CORE Rank: A* (CORE 2023)
- ACM Digital Library:
- Conference Website: ICSE 2026