Undergraduate Interns Chaeyoung and Chaeri's Paper Accepted at DSN 2026

We are incredibly proud to announce that our undergraduate interns, Chaeyoung Lee and Chaeri Jung, have had their paper accepted to the 56th Annual IEEE/IFIP International Conference on Dependable Systems and Networks (DSN 2026) as co-first authors (equal contribution). DSN is widely recognized as a premier, top-tier conference in the fields of system dependability and cybersecurity. Achieving an acceptance at DSN as undergraduate students is truly an extraordinary accomplishment and a testament to their exceptional capabilities.

A Journey of Dedication and Resilience

Chaeyoung and Chaeri began their internship in our lab right at the start of their 3rd year of undergraduate studies at the Division of Artificial Intelligence Engineering, Sookmyung Women’s University (SMU). From day one, they demonstrated remarkable commitment and enthusiasm, dedicating themselves to their research and frequently working overtime to achieve their goals.

They have been incredibly self-motivated throughout their research journey. Even when their experiments faced challenges and produced unexpected results, they never gave up. They pushed their work forward with resilience and exhibited outstanding dedication during the demanding rebuttal phase, doing exactly what was necessary to secure this acceptance. Another proof of their relentless commitment can be found in our December 2025 post, which beautifully captures the intense final hours leading up to their very first paper submission.

A glimpse of the rebuttal phase---even during the lab's Membership Training (MT).

Despite being only 3rd-year students, they have proven their immense potential. Notably, Chaeyoung belongs to the very first pioneering cohort of the Division of Artificial Intelligence Engineering, which was newly established in 2023. She is also the first student from her department to achieve such a milestone. Equally impressive is Chaeri, who transferred into the Division of Artificial Intelligence Engineering from the Division of Advanced Engineering (첨단공학부). Since her transfer, Chaeri has consistently demonstrated overwhelmingly exceptional performance in her academic coursework, distinguishing herself remarkably even among the division’s original cohort.

During their research, they took the initiative to directly communicate with the Cyber Analysis & Defense department at Fraunhofer FKIE in Germany. Their proactive efforts enabled them to successfully obtain crucial malicious domain samples from DGArchive, significantly contributing to the depth and rigor of their work.

About Their Research

The accepted paper is titled “DRIFT: Drift-Resilient Invariant-Feature Transformer for DGA Detection.”

Domain Generation Algorithms (DGAs) continuously evolve, posing a persistent challenge to reliable network defense. Existing deep-learning-based detectors often suffer from severe performance degradation over time due to temporal concept drift. To address this, Chaeyoung and Chaeri introduced an innovative drift-resilient detection framework that combines a hybrid character- and subword-level tokenization strategy with multi-task self-supervised pre-training.

Notably, they rigorously evaluated their model on a massive longitudinal dataset spanning a nine-year period—comprising over 260 million domain samples (including nearly 50 million unique benign domains and 150 million unique DGA domains). Their extensive findings demonstrate that this proposed method significantly mitigates temporal degradation and outperforms state-of-the-art baselines. This work offers a highly dependable foundation for long-term defense against ever-evolving DGA threats.

This photo, taken at 23:47 on October 22, 2025, stands as another piece of evidence of their incredible dedication and late-night commitment to their research.

I want to express my deepest respect for them and extend my warmest congratulations on this incredible achievement!