I am an assistant professor in the ECE department at UC Riverside.

My research focuses on machine learning, reinforcement learning, and foundation models, with an emphasis on developing efficient and reliable learning algorithms and systems for large-scale, multimodal problems. I am particularly interested in the following research directions:

• Test-time training and scaling: enhancing pretrained model performance and reliability during inference, especially on new or unseen tasks.

• Interactive agents: leveraging foundation models to develop efficient and adaptive decision-making agents in interactive environments.

Previously, I received my Ph.D. in Computer Sciences from the University of Wisconsin–Madison, advised by Robert Nowak. During Ph.D., I worked on theoretical foundations of interactive machine learning, including active learning and contextual bandits. I now aim to incorporate these algorithmic insights into the design of practical AI systems — see examples below.

Selected Recent Papers

• Test-Time Matching: Unlocking Compositional Reasoning in Multimodal Models
  Yinglun Zhu, Jiancheng Zhang, and Fuzhi Tang
  arXiv preprint 2025

• Online Finetuning Decision Transformers with Pure RL Gradients
  Junkai Luo and Yinglun Zhu
  arXiv preprint 2025

• Strategic Scaling of Test-Time Compute: A Bandit Learning Approach
  Bowen Zuo and Yinglun Zhu
  arXiv preprint 2025

• Towards Multimodal Active Learning: Efficient Learning with Limited Data Pairs
  Jiancheng Zhang and Yinglun Zhu
  arXiv preprint 2025

• Efficient Sequential Decision Making with Large Language Models
  Dingyang Chen, Qi Zhang, and Yinglun Zhu
  Conference on Empirical Methods in Natural Language Processing (EMNLP) 2024, [Code]

Publications

• Chain-of-Region: Visual Language Models Need Details for Diagram Analysis
  Xue Li, Yiyou Sun, Wei Cheng, Yinglun Zhu, and Haifeng Chen
  International Conference on Learning Representations (ICLR) 2025

• Efficient Sparse PCA via Block-Diagonalization
  Alberto Del Pia, Dekun Zhou, and Yinglun Zhu
  International Conference on Learning Representations (ICLR) 2025, [Code]

• LeMix: Unified Scheduling for LLM Training and Inference on Multi-GPU Systems
  Yufei Li, Zexin Li, Yinglun Zhu, Cong Liu
  Real-Time Systems Symposium (RTSS) 2025

• Efficient Sequential Decision Making with Large Language Models
  Dingyang Chen, Qi Zhang, and Yinglun Zhu
  Conference on Empirical Methods in Natural Language Processing (EMNLP) 2024, [Code]

• An Experimental Design Framework for Label-Efficient Supervised Finetuning of Large Language Models
  Gantavya Bhatt, Yifang Chen, Arnav M. Das, Jifan Zhang, Sang T. Truong, Stephen Mussmann, Yinglun Zhu, Jeffrey Bilmes, Simon S. Du, Kevin Jamieson, Jordan T. Ash, and Robert Nowak
  Findings of the Association for Computational Linguistics: ACL (ACL Findings) 2024

• LabelBench: A Comprehensive Framework for Benchmarking Adaptive Label-Efficient Learning
  Jifan Zhang, Yifang Chen, Gregory Canal, Arnav Das, Gantavya Bhatt, Stephen Mussmann, Yinglun Zhu, Jeffrey Bilmes, Simon Du, Kevin Jamieson, and Robert Nowak
  Journal of Data-Centric Machine Learning Research 2024, [Code]
  \(\bigstar\) Also selected for a Poster Award at Midwest ML Symposium 2024

• Interactive Machine Learning: From Theory to Scale
  Yinglun Zhu
  Ph.D. Dissertation, University of Wisconsin-Madison, 2023

• Infinite Action Contextual Bandits with Reusable Data Exhaust
  Mark Rucker, Yinglun Zhu, and Paul Mineiro
  International Conference on Machine Learning (ICML) 2023

• Active Learning with Neural Networks: Insights from Nonparametric Statistics
  Yinglun Zhu and Robert Nowak
  Conference on Neural Information Processing Systems (NeurIPS) 2022

• Efficient Active Learning with Abstention
  Yinglun Zhu and Robert Nowak
  Conference on Neural Information Processing Systems (NeurIPS) 2022

• Contextual Bandits with Large Action Spaces: Made Practical
  Yinglun Zhu, Dylan Foster, John Langford, and Paul Mineiro
  International Conference on Machine Learning (ICML) 2022, [Code]
  \(\bigstar\) Incorporated into the leading machine learning library Vowpal Wabbit (see here for detailed instructions).

• Contextual Bandits with Smooth Regret: Computational Efficiency in Continuous Action Spaces
  Yinglun Zhu and Paul Mineiro
  International Conference on Machine Learning (ICML) 2022, [Code]
  \(\bigstar\) Selected for a full oral presentation (top 2.1%)

• Near Instance Optimal Model Selection for Pure Exploration Linear Bandits
  Yinglun Zhu, Julian Katz-Samuels, and Robert Nowak
  International Conference on Artificial Intelligence and Statistics (AISTATS) 2022, [Code]

• Pareto Optimal Model Selection in Linear Bandits
  Yinglun Zhu and Robert Nowak
  International Conference on Artificial Intelligence and Statistics (AISTATS) 2022, [Code]

• Pure Exploration in Kernel and Neural Bandits
  Yinglun Zhu\(^\star\), Dongruo Zhou\(^\star\), Ruoxi Jiang\(^\star\), Quanquan Gu, Rebecca Willett, and Robert Nowak
  Conference on Neural Information Processing Systems (NeurIPS) 2021, [Code]

• On Regret with Multiple Best Arms
  Yinglun Zhu and Robert Nowak
  Conference on Neural Information Processing Systems (NeurIPS) 2020, [Code]

• Robust Outlier Arm Identification
  Yinglun Zhu, Sumeet Katariya, and Robert Nowak
  International Conference on Machine Learning (ICML) 2020, [Code]

Group

Ph.D. Students: Jiancheng Zhang (Fall 2024 - present), Bowen Zuo (Fall 2024 - present)

Prospective Students: If you are interested in joining my lab, please apply to the ECE or CSE Ph.D. program at UC Riverside and list me as a potential advisor.

Teaching

• Spring 2025: EE 260 Large Models and Advances in AI

• Winter 2025: EE 114 Probability, Random Variables, and Random Processes in Electrical Engineering

• Fall 2024: EE/CS 228 Introduction to Deep Learning

• Spring 2024: EE 260 Large Models and Advances in AI

• Fall 2023: EE/CS 228 Introduction to Deep Learning