Self-Supervised Scaling of Terminal Environments for Scientific Domains
Self-supervised construction of scientific terminal-agent environments through software workflow reconstruction and verification on hidden inputs.
UNIVERSITY OF GEORGIA · COMPUTER SCIENCE
余骥超 Ph.D. Student
I am a Ph.D. student at the University of Georgia, advised by Prof. Ninghao Liu.
I received my master's degree in Electrical & Computer Engineering from The Ohio State University, where I worked with Prof. Jia Liu, and completed my undergraduate studies at Wuhan University of Science and Technology.
Machine learning systems, federated learning, and resource-efficient model training.
Distributed and multi-objective optimization for learning systems.
Language model evaluation, automated assessment, and fairness in education.
Self-supervised construction of scientific terminal-agent environments through software workflow reconstruction and verification on hidden inputs.
* Equal contribution · LNCS 15881, pp. 268–275.
Examining how automated scoring systems treat English language learners.
Assessing higher-order cognitive skills using educational and psychological benchmarks.
A multi-domain evaluation of reasoning capabilities and limitations in o1-preview.
35(12), 123–126 · In Chinese
Advances in Computer Science Research, vol. 93, pp. 30–33
Issue 12 · In Chinese
Issue 20 · In Chinese
Issue 9 · In Chinese
Ph.D. student in Computer Science
Advisor: Ninghao Liu
M.S. in Electrical & Computer Engineering
Advisor: Jia Liu
B.Eng. in Electronic Information Engineering
Joined as a research assistant
Teaching Assistant · Advanced Digital Design
Teaching Assistant · Signals & Systems
GET IN TOUCH