Portrait of Haoyu Liu

Computer Science · Algorithms & Efficiency

Haoyu Liu

Ph.D. candidate in Computer Science at NTU CCDS, working with Prof. Siqiang Luo in Singapore.

I design scalable graph algorithms and efficient learning methods for large, dynamic graphs. My research spans Personalized PageRank, graph kernels, approximation algorithms, and data structures, alongside lightweight and heterophilous Graph Neural Networks and spectral benchmarking. More recently, I have worked on Graph-LLM integration and temporal retrieval, with a consistent focus on reducing computation and memory while preserving theoretical guarantees and practical effectiveness.

01 / Background

From statistics to graph algorithms.

Now

Ph.D. in Computer Science

Nanyang Technological University, Singapore · advised by Prof. Siqiang Luo.

Earlier

Research Intern

AI Lab, Samsung Research China - Beijing, under Dr. Yang Liu. Worked on embodied AI agents and household instruction following.

Foundation

B.Sc. in Statistics and Big Data

Renmin University of China.

02 / Recognition

Research that travels.

Best Newcomer Research Paper Award

SIGMOD/PODS 2026 · for Near-Optimality for Single-Source Personalized PageRank

PREMIA Best Student Paper Award

2025 in Singapore · for the spectral GNN benchmark study

3rd place, CVPR 2022 ALFRED Challenge

For embodied vision-and-language interaction agents.

03 / Publications

Recent work

Selected papers from my publication record.

View all publications

alphabetical author order · * equal contribution

Open to conversations

Have a graph problem worth scaling?

I am happy to discuss algorithms, efficiency, and collaboration at the intersection of graph data and machine learning.