Xinyue Wang

Xinyue Wang

PhD Student of Data Science

University of California San Diego

About Me

Hi! I am a PhD student in the Halıcıoğlu Data Science Institute (HDSI) at University of California San Diego. My research interest lays in the intersection of Causality and Machine Learning, include Causal Discovery, Causal Representation Learning and Causal Machine Learning. My ultimate question is, can we discover world model with the self-consistent loop of autonomously Causal Discovery and Causal Inference.

Currently, I conduct causality-related research in HDSI, advised by Dr. Biwei Huang. Before joining in UCSD, I mainly focused on developing automated causal discovery algorithm for large complex system, advised by Dr. Konrad Kording. I am also interested in Brain-Computer Interface and Computational Neuroscience, and conducted research related to Neurofeedback system advised by Dr. Gan Huang. Meanwhile, I am a data science enthusiast, have won multiple medals on Kaggle.

Interests
  • Causal Discovery
  • Causal Representation Learning
  • Causal Machine Learning
Education
  • PhD in Data Science, present

    University of California San Diego

  • MSE in Bioengineering, 2023

    University of Pennsylvania

  • BEng in Biomedical Engineering, 2021

    Shenzhen University

  • Exchange Student, 2020

    University of Pennsylvania

Experience

 
 
 
 
 
Kording Lab
Student Researcher
Nov 2021 – May 2023 University of Pennsylvania, Pennsylvania, United States

Responsibilities include:

  • Conduct large complex system simulations and perturbations studies on NMOS6502 microprocessor.
  • Design meta-learning algorithms to infer the causal relationships between system parameters.
  • Collaborate with the lab to do research on Theoretical Deep Learning and Causal Inference.
 
 
 
 
 
UPenn CIS 522 Deep Learning
Teaching Assistant
Jan 2023 – May 2023 University of Pennsylvania, Pennsylvania, United States

Responsibilities include:

  • Led and mentored a pod of 15 students in the twelve-week Deep Learning study.
  • Helped three groups of students design and develop their Deep Learning projects.
  • Designed and organized homework and tutorials.
 
 
 
 
 
Neuromatch Academy Deep Learning Program
Teaching Assistant
Jul 2022 – Jul 2022 California, United States

Responsibilities include:

  • Led and mentored students in the three-week Deep Learning tutorials study.
  • Helped students design and develop their Deep Learning projects based on Computer Vision.
  • Managed and organized tutorials learning and communications with mentors and students.
 
 
 
 
 
Kerry Rehab
Engineer Intern
Nov 2020 – Dec 2020 Guangdong, China

Responsibilities include:

  • Conduct data process pipeline development to analyze and model the EEG data.
  • Assisted recruit subjects and collected EEG data for following neural feedback analysis.
 
 
 
 
 
Laboratory of Medical Informatics & Neural Dynamics
Student Researcher
Sep 2019 – May 2021 Shenzhen University, Guangdong, China

Responsibilities include:

  • Developed a Python based real-time brain wave visualization tool for host computer
  • Developed a multi-module C++ based real-time Neural Feedback System on OpenBCI
  • Designed and conducted tests to quantify the intensity and depth of the system’s modulation of brain wave

Activity&Accomplish­ments

Peer Reviewer
Served as a peer reviewer for prestigious conferences and journals including UAI (2024), CLeaR (2024), RLC (2024), and TMLR.
Shenzhen University
Fellowships and Awards

Fellowships and Awards during acquisition of Bachelor’s Degree:

  • Twice Learning Star
  • Twice Innovation and Invention Star
  • Outstanding Bachelor Thesis
  • Honor Prize in the 2019 International Mathematical Contest in Modeling
  • Third Place in the 2019 National Biomedical Engineering Design Competition

Projects

Causal-Copilot: An Autonomous Causal Analysis Agent
A toolkit that integrates LLM capabilities and domain expertise to enable automated causal analysis for scientific researchers and data scientists.
Causal-Copilot: An Autonomous Causal Analysis Agent
Learning Causal Discovery
Learn to discover causality inside a large complex system without human prior
Learning Causal Discovery
Sartorius Neuronal Cells Segmentation
A robust deep learning pipeline for segmenting neuronal cells in microscopy images.
Sartorius Neuronal Cells Segmentation

Recent Publications

Quickly discover relevant content by filtering publications.
(2024). Causal-Copilot: An Autonomous Causal Analysis Agent.

Cite

(2024). Towards Generalizable Reinforcement Learning via Causality-Guided Self-Adaptive Representations. arXiv preprint arXiv:2407.20651.

Cite

(2023). Deep Networks as Paths on the Manifold of Neural Representations. ICML-TAGML Workshop.

Cite URL

(2023). Learning domain-specific causal discovery from time series. Transactions on Machine Learning Research.

Cite URL

(2020). The Real Time EEG Phase Locked Feedback Control for Alpha Amplitude and Frequency Regulation: An OpenBCI Implementation. 2020 9th International Conference on Bioinformatics and Biomedical Science.

Cite

Contact