Xiaolei (Alex) Chu

Hi, I’m Xiaolei (Alex), a fifth year Ph.D. candidate at UC Berkeley, advised by Prof. Ziqi Wang.

My research focuses on LLM post training, generative retrieval, diffusion models, reinforcement learning, and 3D vision. I also have a background in mechanics, uncertainty quantification, and statistical physics. I am always happy to connect and discuss potential collaborations.

Email  /  CV  /  Scholar  /  Github  /  LinkedIn

profile photo

News

May 2026Joining Google DeepMind as a research intern, working on generative retrieval.
Mar 2026New preprint “Social Amplification Dominates Collective Hazard Response” is now on arXiv. The paper uses mean-field theory from statistical physics to explain how social networks amplify collective panic when communities face hazards, revealing the conditions under which individual fear cascades into a system-wide phase transition.

Selected Publications
SteerGR architecture comparing joint and modular language conditioning for semantic ID generation and steerable retrieval
Shortcut or Semantics? Understanding Conditioning Interfaces for Steerable Generative Retrieval
Xiaolei Chu, Shao-Chuan Wang, Yu-Neng Chuang, Zhaoheng Zheng, Sanjay S. Girija, Ruining He, Aniruddh Nath, Li Wei, Lichan Hong, Xinyang Yi
Google DeepMind & YouTube, Aug. 2026
paper (access upon request)

SteerGR studies how natural-language preferences steer Gemma 3 to generate semantic IDs for retrieval. Comparing joint conditioning with modular language-state injection reveals how conditioning interfaces affect retrieval accuracy, phrasing robustness, and retention of general reasoning ability.
GeoScout camera trajectories for active reconstruction of ShapeNet objects
GeoScout: A Generalizable Geometry-Aware Policy for Active 3D Reconstruction
Xiaolei Chu, Guanren Zhou, Ziqi Wang, Khalid M. Mosalam
coming soon, 2026
code

GeoScout conditions a reinforcement-learning next-best-view policy on refined geometry captions, using language as a prior over unobserved shape to reconstruct unseen objects with fewer views than a matched no-caption policy.
Expert-grounded structural diagnosis linked to segmentation masks
Entity-Guided Expert-Grounded Visual Diagnosis for Structures in the Wild
Guanren Zhou, Xiaolei Chu, Khalid M. Mosalam
coming soon, 2026

Hierarchical SAM 3 retrieval and DINOv3 specialist classifiers ground a frozen vision-language model’s fine-grained diagnoses in segmentation masks, enabling traceability without new task-specific annotations or model fine-tuning.
CFM4EQ architecture with a conditional flow-matching prior, continuous condition encoder, latent diffusion U-Net, and spectrogram decoder
Hierarchical Conditional Generative Modeling for Site-based Ground Motion Synthesis
Xiaolei Chu, Maijia Su, Ziqi Wang
coming soon, 2026
code

Ground-motion energy and duration remain stochastic even under a fixed earthquake scenario. This project learns their joint distribution conditioned on magnitude, distance, and site conditions, then propagates this variability into three-component spectrogram synthesis. This hierarchy captures energy–duration dependence and record-to-record diversity while recovering expected magnitude scaling and distance attenuation.
Social amplification
Social Amplification Dominates Collective Hazard Response
Xiaolei Chu, Guanren Zhou, Marco Broccardo, Didier Sornette, Khalid M. Mosalam, Ziqi Wang
arXiv preprint, Mar. 2026
arXiv

Fine-tuned BERTweet estimates state-level stress prevalence from social-media posts to calibrate an interpretable network model, revealing that social influence outweighed direct exposure in over 80% of U.S. states.

Awards & Honors

May 20252nd Place, Objective Resilience Student Paper/Presentation Competition at EMI 2025, Anaheim, CA
Aug 2024Professor Alex & Georgia Scordelis Fellowship, UC Berkeley
Oct 2022Excellent Master's Thesis, Tongji University
Jun 2022Excellent Graduate, Tongji University
Oct 2021China National Scholarship
Aug 2019Citation for Outstanding College Graduate (Summa Cum Laude), China Civil Engineering Society
Jun 2019Excellent Undergraduate of Sichuan Province
Jan 20191st International Engineering Mechanics Competition: 1st Prize (Individual) & 2nd Prize (Team)
Apr 2018ASCE Steel Bridge Competition: 5th Overall & Best Display Medal (Group Leader), Sacramento, CA
Oct 20171st Prize, 9th Chinese Mathematics Competition (CMC)
Jul 20171st Prize, 11th Zhou Peiyuan National Mechanics Competition
Oct 20161st Prize, 8th Chinese Mathematics Competition (CMC)
2016–18Railway Alumni Scholarship, Chengdu

Academic Services

Journal Reviewer

  • Journal of Structural Engineering, ASCE
  • Journal of Engineering
  • Advanced Engineering Informatics
  • Mechanical Systems and Signal Processing
  • Engineering Structures
  • Measurement
  • Journal of Building Engineering
  • Automation in Construction
  • Reliability Engineering & System Safety
  • Information Processing and Management
  • Structural Safety

Professional Membership

  • Student Member of ASCE (2016–present)

Website source from Jon Barron.