Yuqing Zhou

Fairfax, Virginia

TRUSTWORTHY & SELF-EVOLVING AI

I build learning systems that can reason efficiently, improve from feedback, and remain reliable when the data or environment changes.

I am a Ph.D. student in Computer Science at George Mason University, advised by Prof. Ziwei Zhu. My current work spans efficient LLM reasoning and post-training, self-evolving AI, and robust learning under spurious correlations.

Previously, I was an Applied Scientist Intern at Amazon Web Services, where I worked on efficient LLM reasoning and agentic conversational AI. I earned an M.S. in Electrical and Computer Engineering from the University of Michigan and a B.Eng. from Southeast University.

Self-evolving AI LLM reasoning LLM post-training Trustworthy AI

View CV Publications Email

Open to Summer 2027 research internships I would be happy to connect about LLM reasoning, post-training, and self-evolving AI.

Efficient reasoning

Improving credit assignment during RL post-training so language models preserve useful reasoning while avoiding redundant or dead-end computation.

Self-evolving AI

Developing AI systems that improve through iterative self-refinement, confidence-aware feedback, and reliable adaptation.

Robust learning

Using causal learning to reduce reliance on spurious correlations and improve worst-group performance under distribution shift.

News

All updates

2026	Our paper “Confidence-Orchestrated Self-Evolution against Uncertain LLM Feedback” was accepted to EMNLP 2026.
Summer 2026	I returned to AWS as an Applied Scientist Intern to work on efficient LLM reasoning and RL-based post-training.
May 27, 2025	Glad to join AWS as an Applied Scientist Intern.
Apr 29, 2025	One paper about mitigating spurious correlations in text classification was accepted to NAACL 2025.
Nov 12, 2024	One paper about Shortcut Learning in NLP was accepted to Findings of EMNLP 2024.

Selected Publications

View all

EMNLP

Confidence-Orchestrated Self-Evolution against Uncertain LLM Feedback

Bowen Wei, Nan Wang, Yuqing Zhou, and 2 more authors

arXiv preprint arXiv:2605.28010, 2026

Accepted to EMNLP 2026

arXiv Bib

@article{wei2026confidence,
  title = {Confidence-Orchestrated Self-Evolution against Uncertain LLM Feedback},
  author = {Wei, Bowen and Wang, Nan and Zhou, Yuqing and Pan, Jinhao and Zhu, Ziwei},
  journal = {arXiv preprint arXiv:2605.28010},
  year = {2026},
  note = {Accepted to EMNLP 2026},
  peer_reviewed = {true},
}

NAACL
Fighting Spurious Correlations in Text Classification via a Causal Learning Perspective

Yuqing Zhou and Ziwei Zhu

In Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers), Apr 2025

Abs DOI Bib Code

In text classification tasks, models often rely on spurious correlations for predictions, incorrectly associating irrelevant features with the target labels. This issue limits the robustness and generalization of models, especially when faced with out-of-distribution data where such spurious correlations no longer hold. To address this challenge, we propose the Causally Calibrated Robust Classifier (CCR), which aims to reduce models’ reliance on spurious correlations and improve model robustness. Our approach integrates a causal feature selection method based on counterfactual reasoning, along with an unbiased inverse propensity weighting (IPW) loss function. By focusing on selecting causal features, we ensure that the model relies less on spurious features during prediction. We theoretically justify our approach and empirically show that CCR achieves state-of-the-art performance among methods without group labels, and in some cases, it can compete with the models that utilize group labels. Our code can be found at: https://github.com/yuqing-zhou/Causal-Learning-For-Robust-Classifier.
@inproceedings{zhou-zhu-2025-fighting, title = {Fighting Spurious Correlations in Text Classification via a Causal Learning Perspective}, author = {Zhou, Yuqing and Zhu, Ziwei}, editor = {Chiruzzo, Luis and Ritter, Alan and Wang, Lu}, booktitle = {Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)}, month = apr, year = {2025}, address = {Albuquerque, New Mexico}, publisher = {Association for Computational Linguistics}, url = {https://aclanthology.org/2025.naacl-long.215/}, doi = {10.18653/v1/2025.naacl-long.215}, pages = {4264--4274}, isbn = {979-8-89176-189-6}, peer_reviewed = {true}, }
Findings of EMNLP
Navigating the Shortcut Maze: A Comprehensive Analysis of Shortcut Learning in Text Classification by Language Models

Yuqing Zhou, Ruixiang Tang, Ziyu Yao, and 1 more author

In Findings of the Association for Computational Linguistics: EMNLP 2024, Nov 2024

Abs DOI Bib Code

Language models (LMs), despite their advances, often depend on spurious correlations, undermining their accuracy and generalizability. This study addresses the overlooked impact of subtler, more complex shortcuts that compromise model reliability beyond oversimplified shortcuts. We introduce a comprehensive benchmark that categorizes shortcuts into occurrence, style, and concept, aiming to explore the nuanced ways in which these shortcuts influence the performance of LMs. Through extensive experiments across traditional LMs, large language models, and state-of-the-art robust models, our research systematically investigates models’ resilience and susceptibilities to sophisticated shortcuts. Our benchmark and code can be found at: https://github.com/yuqing-zhou/shortcut-learning-in-text-classification.
@inproceedings{zhou-etal-2024-navigating-shortcut, title = {Navigating the Shortcut Maze: A Comprehensive Analysis of Shortcut Learning in Text Classification by Language Models}, author = {Zhou, Yuqing and Tang, Ruixiang and Yao, Ziyu and Zhu, Ziwei}, editor = {Al-Onaizan, Yaser and Bansal, Mohit and Chen, Yun-Nung}, booktitle = {Findings of the Association for Computational Linguistics: EMNLP 2024}, month = nov, year = {2024}, address = {Miami, Florida, USA}, publisher = {Association for Computational Linguistics}, url = {https://aclanthology.org/2024.findings-emnlp.146/}, doi = {10.18653/v1/2024.findings-emnlp.146}, pages = {2586--2614}, peer_reviewed = {true}, }
CIKM
A generalized propensity learning framework for unbiased post-click conversion rate estimation

Yuqing Zhou, Tianshu Feng, Mingrui Liu, and 1 more author

In Proceedings of the 32nd ACM International Conference on Information and Knowledge Management, Birmingham, United Kingdom, Nov 2023

Abs DOI Bib Code

This paper addresses the critical gap in the unbiased estimation of post-click conversion rate (CVR) in recommender systems. Existing CVR prediction methods, such as Inverse Propensity Score (IPS) and various Doubly Robust (DR) based estimators, overlook the impact of propensity estimation on the model bias and variance, thus leading to a debiasing performance gap. We propose a Generalized Propensity Learning (GPL) framework to directly minimize the bias and variance in CVR prediction models. The proposed method works as a complement to existing methods like IPS, DR, MRDR, and DRMSE to improve prediction performance by reducing their bias and variance. Extensive experiments on real-world datasets and semi-synthetic datasets demonstrate the significant performance promotion brought by our proposed method. Data and code can be found at: https://github.com/yuqing-zhou/GPL.
@inproceedings{zhou2023generalized, title = {A generalized propensity learning framework for unbiased post-click conversion rate estimation}, author = {Zhou, Yuqing and Feng, Tianshu and Liu, Mingrui and Zhu, Ziwei}, booktitle = {Proceedings of the 32nd ACM International Conference on Information and Knowledge Management}, pages = {3554--3563}, year = {2023}, url = {https://doi.org/10.1145/3583780.3614760}, doi = {10.1145/3583780.3614760}, location = {Birmingham, United Kingdom}, peer_reviewed = {true}, }