Tzu-Han Lin

Engineering VI, Room 368

404 Westwood Plaza

Los Angeles, CA 90095

Ph.D. student @ UCLA

I am a first-year Ph.D. student in Computer Science at UCLA, where I am fortunate to be advised by Prof. Kai-Wei Chang.

Research Focus: My research centers on Large Language Models (LLMs). I am interested in developing LLMs and agents that are resource-efficient and reliable.

Currently, I’m thinking about:

Language Agents:
- Trustworthiness: How can we make agents more interpretable, reliable, and faithful to users?
- Development: When should we optimize the agent harness versus the model weights?
Model Architecture:
- Looped LLM: How should looped LLMs allocate and utilize test-time compute?
Self-Improving AI:
- Alignment: How can we maintain alignment when human supervision is difficult to provide?
- Evaluation at Scale: How can we efficiently evaluate progress on long-horizon, open-ended tasks, such as scientific research?
- Efficient Discovery: How can we improve the efficiency of test-time training for discovery?

Collaboration & Mentoring: I am open to research collaborations and happy to mentor students interested in LLM research. If you are interested in working together or discussing research ideas, feel free to reach out via email!

Previously, I received my M.S. degree from the Graduate Institute of Networking and Multimedia at National Taiwan University, where I was advised by Prof. Yun-Nung (Vivian) Chen. My thesis focused on the efficeincy of search agents and received the ACLCLP Master’s Thesis Award of Excellence. I was also a visiting research intern at the University of Virginia, where I worked with Prof. Yu Meng.

I received my bachelor’s degree from National Taiwan University. During my undergraduate studies, I worked with Prof. Hung-yi Lee and Prof. Yun-Nung (Vivian) Chen on speech processing and natural language processing.

Misc: When I am not doing research, I like to listen to different genre of musics, especially Blues, Rock, and Grunge. I’m a huge fan of John Mayer and I like to play eletric guitars!

news

Oct 21, 2026	My master’s thesis, AdaSearch, received the ACLCLP Master’s Thesis Award of Excellence!
Aug 21, 2026	AdaSearch is accepted to EMNLP’26 main conference. See you in Budapest 🇭🇺!
Apr 08, 2026	I’m thrilled to announce that I will join UCLA as a Computer Science Ph.D. student in Fall 2026, where I will be advised by Prof. Kai-Wei Chang!
Dec 18, 2025	Our new preprint AdaSearch: Balancing Parametric Knowledge and Search in Large Language Models via Reinforcement Learning is out! We investigate how LLM search agents should balance parametric knowledge and search, and propose AdaSearch, a framework that teaches agents to explicitly decide when to search via RL. It achieves superior self-knowledge awareness without complex reward engineering, and provides transparent decision-making rationales.
May 27, 2025	Excited to share that I have joined Prof. Yu Meng’s group as a visiting research intern, and I’m grateful for the opportunity to work with the team!

selected publications

AdaSearch: Balancing Parametric Knowledge and Search in Large Language Models via Reinforcement Learning

Tzu-Han Lin, Wei-Lin Chen, Chen-An Li, Hung-yi Lee, Yun-Nung Chen, and Yu Meng

In The 2026 Conference on Empirical Methods in Natural Language Processing, Oct 2026

To appear

Abs PDF Code

Equipping large language models (LLMs) with search engines via reinforcement learning (RL) promises effective search agents. However, adaptively balancing internal parametric knowledge with external search remains a challenge, as overreliance on search introduces unnecessary cost and risks exposure to noisy or malicious content, while relying solely on parametric knowledge risks hallucination. Prior efforts mitigate search overuse through tool-call reward shaping, which requires heavy reward engineering and conflates necessary and unnecessary search. To address these limitations, we revisit the evaluation of search agents through an F1-based decision metric, revealing that prior methods often overlook readily available parametric knowledge. Motivated by this, we propose AdaSearch, a simple two-stage, outcome-driven RL framework that disentangles problem-solving from the decision to search, making the decision process explicit and interpretable. Extensive experiments demonstrate that AdaSearch significantly improves search-decision quality and reduces unnecessary search calls, with only a small trade-off in QA accuracy relative to always-search.
Transferring Textual Preferences to Vision-language Understanding Through Model Merging

Chen-An Li, Tzu-Han Lin, Yun-Nung Chen, and Hung-yi Lee

In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers), Jul 2025

Abs PDF Code

Large vision-language models (LVLMs) perform outstandingly across various multimodal tasks. However, their ability to evaluate generated content remains limited, and training vision-language reward models (VLRMs) with preference data is computationally expensive. This paper explores a training-free alternative by merging text-based reward models (RMs) with LVLMs to create VLRMs. Our approach shows that integrating these models leads to improved performance over LVLMs’ scoring and text-based RMs, offering an efficient method for incorporating textual preferences into LVLMs.
DogeRM: Equipping Reward Models with Domain Knowledge through Model Merging

Tzu-Han Lin, Chen-An Li, Hung-Yi Lee, and Yun-Nung Chen

In Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, Nov 2024

Abs PDF Code

Reinforcement learning from human feedback (RLHF) is a popular strategy for aligning large language models (LLMs) with desired behaviors. Reward modeling is a crucial step in RLHF. However, collecting paired preference data for training reward models is often costly and time-consuming, especially for domain-specific preferences requiring expert annotation. To address this challenge, we propose the **Do**main knowled**ge** merged **R**eward **M**odel (**DogeRM**), a novel framework that integrates domain-specific knowledge into a general reward model by model merging. The experiments demonstrate that DogeRM enhances performance across different benchmarks and provide a detailed analysis showcasing the effects of model merging, showing the great potential of facilitating model alignment."
PEFT for Speech: Unveiling Optimal Placement, Merging Strategies, and Ensemble Techniques

Tzu-Han Lin^*, How-Shing Wang^*, Hao-Yung Weng^**, Kuang-Chen Peng^**, Zih-Ching Chen, and Hung-yi Lee

In 2024 IEEE International Conference on Acoustics, Speech, and Signal Processing Workshops (ICASSPW), Apr 2024

Abs PDF

Parameter-Efficient Fine-Tuning (PEFT) is increasingly recognized as an effective method in speech processing. However, the optimal approach and the placement of PEFT methods remain inconclusive. Our study conducts extensive experiments to compare different PEFT methods and their layer-wise placement adapting Differentiable Architecture Search (DARTS). We also explore the use of ensemble learning to leverage diverse PEFT strategies. The results reveal that DARTS does not outperform the baseline approach, which involves inserting the same PEFT method into all layers of a Self-Supervised Learning (SSL) model. In contrast, an ensemble learning approach, particularly one employing majority voting, demonstrates superior performance. Our statistical evidence indicates that different PEFT methods learn in varied ways. This variation might explain why the synergistic integration of various PEFT methods through ensemble learning can harness their unique learning capabilities more effectively compared to individual layer-wise optimization.