NPC tool accepted at ICAART 2026

NPC, a tool built by our students Pachanitha Saeheng, Napat Boongaree, and Chutweeraya Sriwilailak, was accepted for publication at ICAART 2026, the 18th International Conference on Agents and Artificial Intelligence. NPC detects code snippets generated by ChatGPT and explains each decision, using machine learning with a local neighborhood sampling strategy that the team proposed for post-hoc explanations. It runs as a web application and gives educators an easy-to-understand explanation for each classification. Professors who evaluated the tool found the explanations clear and helpful, which suggests it can support academic integrity in programming education. Dr. Chaiyong Ragkhitwetsagul, Dr. Teeradaj Racharak of Tohoku University, and Dr. Ekapol Chuangsuwanich of Chulalongkorn University co-supervised the project, which began during a 2024 research internship at JAIST.