I am a third-year PhD student in the Machine Learning program at Georgia Tech’s School of Electrical and Computer Engineering. I’m working on machine learning on graphs under the supervision of Prof. Pan Li. Before coming to Georgia Tech, I was a PhD student at Purdue University in the Department of Computer Science. Prior to that, I worked at Tencent Big Data Lab for a while after receiving my master’s degree from the University of Illinois Urbana-Champaign. Currently, I focus on geometric & graph deep learning, epsecially on interpretable, generalizable, and scalable learning methods to advance scientific discovery.

My Chinese given name is 思奇(siqi), pronounced as 思/si/ 奇/tʃiː/. My Chinese surname is 苗(miao), pronounced as 苗/mɪaʊ/. It’s also fine to call me Aki.

I will be seeking internship opportunities for the summer of 2025. If you know of any exciting openings, please feel free to contact me!

🔥 News

  • 2024.06: Our paper on graph state space model (GSSC) is released on arXiv.
  • 2024.05: Our paper HEPT is accepted to ICML 2024 and selected for oral presentation! See you in Vienna!
  • 2024.03: Gave a talk at A3D3 institute HEP group on HEPT.
  • 2024.02: Our paper on efficient point transformer (HEPT) is released on arXiv.
Old news in 2023
  • 2023.10: Our paper on benchmarking OOD performance in geometric deep learning is released on arXiv.
  • 2023.07: One paper accepted to CTD’23.
  • 2023.03: Received ICLR’23 Travel Award! Thanks ICLR and see you in Kigali in the coming May!
  • 2023.01: Our paper LRI is accepted to ICLR 2023 and NeurIPS 2022 AI4Science!
Old news in 2022
  • 2022.11: Gave a talk at IML Machine Learning Working Group on LRI with slides here.
  • 2022.11: I will be moving to Georgia Tech in December as a PhD student in Machine Learning (ECE) with Prof. Pan Li.
  • 2022.11: Our latest paper on model interpretability (LRI) will show up at NeurIPS 2022 AI for Science Workshop!
  • 2022.10: Gave two lectures at CS593-CLG(code) and Phys570-AI(code).
  • 2022.07: Gave two talks at FastML and AITime on our ICML 2022 paper GSAT(slides).
  • 2022.07: Awarded free registration and graduate travel grant for ICML! Thanks!
  • 2022.05: One paper accepted to ICML 2022!

📝 Publications

ICML 2024
sym

LSH-Based Efficient Point Transformer with Applications in High-Energy Physics
Siqi Miao, Zhiyuan Lu, Mia Liu, Javier Duarte, Pan Li. ICML 2024 (Oral, Top 1.5%)
Paper Github License

ICLR 2023
sym

Interpretable Geometric Deep Learning via Learnable Randomness Injection
Siqi Miao, Yunan Luo, Mia Liu, Pan Li. ICLR 2023
Paper Github License

Preprint
sym

What Can We Learn from State Space Models for Machine Learning on Graphs?
Yinan Huang*, Siqi Miao*, Pan Li. Preprint
(*Equal contribution, listed in alphabetical order)
Paper Github

🎖 Honors and Awards

  • 2023.03 ICLR’23 Travel Award
  • 2022.07 ICML’22 Travel Award
  • 2022.07 Purdue Graduate Travel Grant
  • 2020.03 Tencent Innovation Award
  • 2020.01 Tencent Outstanding Employee (Top 10%)

💬 Invited Talks

  • 2024.03, A3D3 Institute HEP Group
  • 2022.11, Inter-Experimental LHC Machine Learning Working Group, CERN
  • 2022.10, Department of Physics, Purdue University
  • 2022.08, AI Time
  • 2022.07, Fast Machine Learning Lab

💻 Experience

  • 2023.01 - Present, Georgia Tech, Research Assistant, Atlanta, USA
  • 2021.08 - 2022.12, Purdue University, Research Assistant, West Lafayette, USA
  • 2019.07 - 2020.05, Tencent, Machine Learning Engineer, Guangzhou, China
  • 2018.05 - 2018.08, Tencent, Machine Learning Intern, Guangzhou, China

📖 Academic Services

  • Conference Reviewer: ICML, ICLR, NeurIPS, LOG, WWW
  • Journal Reviewer: IEEE Transactions on Signal Processing

🎮 Misc

  • The Legend of Zelda: Breath of the Wild and The Witcher 3: Wild Hunt are the best games ever. Genshin Impact is also awesome!
  • I’m a big fan of Lego Technic. I once combined them with RC cars to build a few cool Lego cars that could drift and climb!
  • I have 4.5k+ followers on Zhihu, a Quora-like platform in China.
  • I built my own PC and home automation system. I also liked to play with Jetson TX2 and Raspberry Pi.