Scholar
Roy Xie
Google Scholar ID: iB-5P0EAAAAJ
Duke University
Citations & Impact
Citations
70
H-index
5
i10-index
2
Publications
12
Co-authors
0
Publications
Resume
Academic Achievements
- Interleaved Reasoning for Large Language Models via Reinforcement Learning Preprint, 2025
- Language Models (Mostly) Know When to Stop Reading NeurIPS, 2025
- Improving Model Alignment Through Collective Intelligence of Open-Source Models ICML, 2025
- ReCaLL: Membership Inference via Relative Conditional Log-Likelihoods EMNLP, 2024
- Adversarial Math Word Problem Generation EMNLP Findings, 2024
- Extracting Lexical Features from Dialects via Interpretable Dialect Classifiers NAACL, 2024
Research Experience
- Before Duke, worked with Antonios Anastasopoulos on low-resource NLP.
Education
- Second-year Ph.D. student in Computer Science at Duke University, advised by Bhuwan Dhingra.
Background
- Research interests: Post-training methods for LLMs, particularly in developing efficient algorithms to improve LLMs' reasoning ability, especially for complex multi-step agentic tasks.
Miscellany
- Supported by the 2025 Apple Scholars in AI/ML Ph.D. Fellowship and the 2024 National Science Foundation Graduate Research Fellowship.