About Me
I’m a third year Ph.D. student at Purdue University under the supervision of Prof. Ananth Grama. Prior to that, I completed my undergraduate studies in Electronic Information Engineering with a minor in Artificial intelligence at the University of Science and Technology of China. [CV]
Research interests:
- Machine Learning, especially Trustworthy ML and Efficient ML.
NEWS
- May, 2024. I’m thrilled to start my internship at Texas Instrument!
- May, 2024. “A Theory of Fault-Tolerant Learning” is accepted by ICML 2024 (spotlight 3.5%) !
- Apr, 2023. “Learning Functional Distributions with Private Labels” is accepted by ICML 2023!
Projects
Cascade reward sampling for efficient decoding-time alignment
Aligning large language models (LLMs) with human preferences is critical for their deployment. Recently, decoding-time alignment has emerged as an effective plug-and-play technique that requires no fine-tuning of model parameters. However, generating text that achieves both high reward and high likelihood remains a significant challenge. Existing methods often fail to generate high-reward text or incur substantial computational costs. In this paper, we propose Cascade Reward Sampling (CARDS) to address both issues, guaranteeing the generation of high-reward and high-likelihood text with significantly low costs.
2024-08-02
1 min read
A Theory of Fault-Tolerant Learning
Developing machine learning models that account for potential faults encountered in real-world environments presents a fundamental challenge for mission-critical applications. In this paper, we introduce a novel theoretical framework grounded in learning theory for dealing with faults. In particular, we propose a framework called fault-tolerant PAC learning, aimed at identifying the most fault-tolerant models from a given hypothesis class (such as neural networks). We show that if faults occur randomly, fault-tolerant learning is equivalent to regular PAC learning.
2024-05-25
1 min read
Experience
Machine Learning Research Intern
Texas Instruments
2024.5 - 2024.8
- contribute to build TI LLM for code generation
Try everything