Yichen Xu

Yichen Xu is a Doctoral Assistant at the Programming Methods Laboratory (LAMP1) within the School of Computer and Communication Sciences at École Polytechnique Fédérale de Lausanne (EPFL). He is currently pursuing his Ph.D. in Computer Science under the supervision of Professor Martin Odersky, the creator of Scala.

• Bachelor of Engineering in Computer Science, Beijing University of Posts and Telecommunications (2018–2022)
• Doctor of Philosophy in Computer Science, EPFL (2022–present)

Yichen's research bridges programming language theory and machine learning. His primary focus lies in advancing Scala's type system, particularly through work on Capture Calculus and Generalized Algebraic Data Types (GADTs), aiming to achieve safe and efficient effect tracking. He has also made significant contributions to graph representation learning, especially in contrastive learning frameworks for graphs. His work spans both theoretical foundations and practical implementations in compilers and open-source libraries.

His recent publications reveal a strong trend in formal methods for programming languages and self-supervised learning on graphs. Key themes include data race prevention via flexible type systems, formalization of inference rules in type calculi, and empirical studies on contrastive learning with adaptive augmentation and structure-aware mining. These works have been published in venues such as IWACO, NeurIPS, WWW, and Scala Symposium.

• First Prize in NSCSCC 2020 competition (EasterMIPS CPU project)

Yichen has been actively involved in mentoring and collaborative research, contributing to major projects like CodeGeeX and the Scala 3 compiler (dotty). He has worked closely with researchers including Prof. Shu Wu, Prof. Zhilin Yang, Dr. Aleksander Boruch-Gruszecki, and Yanqiao Zhu. His open-source contributions include tools like PyGCL, fscala2c, and meow, reflecting a strong commitment to practical software development and community engagement.

He is part of the Programming Methods Laboratory at EPFL and has previously been affiliated with research groups at CRIPAC (CASIA) and IIIS (Tsinghua University), forming a broad collaborative network in programming languages and machine learning.