Minghao Sun

PhD Candidate · Department of Computer Science
National University of Singapore

Hi! I'm Minghao, a PhD candidate at the Department of Computer Science, National University of Singapore. I'm advised by Prof. Yang Zhang and supported by the President's Graduate Fellowship (PGF).

Before NUS, I earned my Bachelor's from Zhejiang University in Veterinary Medicine and Statistics, with a minor in Computer Science. This interdisciplinary path first motivated my interest in adaptation and learning, which I now study in AI systems.

Email: minghao [dot] sun [at] u [dot] nus [dot] edu

Google Scholar · LinkedIn

Taken at Hong Kong Museum of Art 香港藝術館

Research

Learning from experience. I study how AI systems can improve through interaction and accumulate useful experience over time. I'm interested in how agents acquire informative experience, learn from sparse or imperfect feedback, consolidate experience into reusable knowledge and memories, and continue adapting without catastrophic forgetting. More broadly, I am interested in general learning mechanisms that can improve as experience and computation scale.

AI4Science, with science as a testbed. Scientific discovery provides a challenging environment for studying these questions: feedback is often costly, noisy, delayed, and only partially aligned with the true objective. My work spans biomolecular design, particularly RNA and proteins, AI agents, and automated scientific discovery. I am especially interested in moving beyond static prediction and generation toward systems that choose actions or experiments, learn from their outcomes, and improve through repeated interaction.

Knowledge across timescales. In the longer term, I am interested in how intelligent systems should acquire, retain, update, and transfer knowledge across different timescales, and whether general principles of learning from experience can transfer across digital, scientific, and even embodied environments.

I welcome collaborations on experience-driven agent learning, automated scientific discovery, and biological experimental validation for RNA and protein design.

News

[2026/05] One paper accepted at KDD 2026 AI4Science Track.
[2026/05] One paper published in eLife.

Selected Publications (Google Scholar)

* equal contribution · † corresponding author

KDD’26AI4Science

The Forgetting-Learning Trade-off: Making Reinforcement Learning Work for Protein Language Models. [Paper]

Hanqun Cao^*, Hongrui Zhang^*, Junde Xu, Zhou Zhang, Lingdong Shen, Minghao Sun, Ge Liu, Jinbo Xu, Wu-Jun Li, Jinren Ni, Cesar de la Fuente-Nunez, Tianfan Fu, Shuting Jin, Pheng-Ann Heng, Fang Wu^†

eLife

ProteinConformers: large-scale and energetically profiled descriptions of protein conformational landscapes. [Paper]

Yihang Zhou^*, Chen Wei^*, Minghao Sun, Lin Wang, Jin Song, Fanding Xu, Yang Li, Wei Zheng, Yang Zhang^†

Preprintunder review

Multi-Oracle Agreement Reveals the Limits of Self-Consistency Evaluation in RNA Design. [Paper]

Minghao Sun, Hanqun Cao, Fang Wu, Zhou Zhang, Zhiyuan Liu, Tianfan Fu, Pheng-Ann Heng, Yang Zhang^†

arXiv2026

Ligand-Conditioned Discrete Diffusion for Protein Sequence–Structure Co-Design. [Paper]

Chen Wei, Fanding Xu, Minghao Sun, Zhiyuan Liu, Lin Wang, Tianrui Jia, Yihang Zhou^†, Yang Zhang^†

arXiv2025

RiboPO: Preference Optimization for Structure- and Stability-Aware RNA Design. [Paper]

Minghao Sun^*, Hanqun Cao^*, Zhou Zhang^*, Chen Wei, Liang Wang, Tianrui Jia, Zhiyuan Liu, Tianfan Fu, Xiangru Tang, Yejin Choi, Pheng-Ann Heng, Fang Wu^†, Yang Zhang^†

NeurIPS’25D&B Track

ProteinConformers: Benchmark Dataset for Simulating Protein Conformational Landscape Diversity and Plausibility. [Paper]

Yihang Zhou^*, Chen Wei^*, Minghao Sun, Jin Song, Yang Li, Lin Wang^†, Yang Zhang^†

Genes2022

Bioinformatics and Experimental Analyses Reveal NFIC as an Upstream Transcriptional Regulator for Ischemic Cardiomyopathy. [Paper]

Yang Ye^*, Qiao Jin^*, Qian Gong, Aoqi Li, Minghao Sun, Sibo Jiang, Yulan Jin, Zhe Zhang, Jin He, Lenan Zhuang^†

Genes2021

Molecular Phylogenetic Analysis of the AIG Family in Vertebrates. [Paper]

Yuqi Huang^*, Minghao Sun^*, Lenan Zhuang^†, Jin He^†

Academic Services

Reviewer. ICML 2026 (Gold Reviewer), NeurIPS 2026.

Co-Reviewer. Nature Communications (Early Career Researcher, ECR).

Mentoring. Chelsea Yep (NUS CS Master's Student).