Zhi-Wei Li (李枝蔚)

PhD student in NYU Center for Neural Science studying computatinal modeling of human cognition in Gureckis lab, New York University.

Here's my email.

Let me introduce myself in these three ways:

  • As CogSci Researcher
  • As Knowledge Researcher
  • As a Human
  • As an explorer
  • As a cognitive science researcher: Understanding human with probabilistic modeling

    Attracted by: Human curiosity, including the emotional drives (what emotion drives more interest?) and informational drives (how to quantify the informational state that motivates more learning?).

    Tricks: Designing experiments, programming, coming up mathematical models to explain humans.

    Thesis: Probabilistic models of subjective judgments

    After 5 years I finally got my PhD from New York University, center for neural science. My thesis studied three topics related with the challenging topic of subjective judgments -- yes, it's quite a journey to extract some common mechanism among people's idiosyncratic judgments and feelings. The first two topics in my thesis respectively addressed the emotional drive and informational drive for self-motivated learning.

    The common methodology of these three projects including behavioral paradigms and probabilistic modeling. Even if you are not an expert in probability theories, your brain may be one. I agree it seems crazy that some neuroscientists claim our brains could perform probabilistic calculations with Bayesian formalism...But throughout the years of my PhD research, I gradually came to realize that probabilistic modeling is a useful starting point, especially when investigating relatively new topics.

    You can check out my full thesis: here .

    Other cognitive science researches

    Topics and Trends in Cognitive Science

    This project is more NLP flavored and purely data-driven. We characterized trends in the field of cognitive science with the dynamic topic model.


    As a human knowledge researcher for better human-computer collaboration

    Internship at Autodesk: HCI Researcher for Tribal Knowledge Authoring (2020/06)

    Designing a structured grammar to capture designer knowledge; Developing a software interface for recording design intents which feeds into automated designing programs.

    Presentation slides here.

    Scientific Knowledge Organized Through Question Network

    This is my long-term passion project started when I was in my physics major in undergrad, when I found out that rather than trying to figure out the reasoning behind every formula on the textbook, the better way to learn is to focus on the question those formula are trying to solve, then explore the solution myself. All the questions can be organized into an organic network which is much easier to understand and follow. This method could be applied to almost any discipline of learning. Throughout years, I have tried different ways to progress this idea:


    As a human

    Being a human is much more than cognition and knowledge. The two major areas of my personal growth in recent years: about body/mind-body connection and about inter-human relationship. I'd like to share some helpful links here.

    Somatic education and Feldenkrais

    After taking Feldenkrais classes ("awareness through movement") for 2 years, I joined Kelowna Feldenkrais Teacher Training in 2021 to become a Feldenkrais practitioner. I hope to learn more about my body with concrete practice, and will one day be able to help others feeling more ease, comfort and joy using their body.

    If you cannot find a local practitioner, there are many online video classes. Alfons is one of my favorite Youtube teachers -- I would just find a mat, lay on the floor, and spend some relaxing time with my body.

    If you would like to read original materials from Moshe Feldenkrais and getting a better understanding of the worldview behind this method, here's two books I've read so far and like a lot:

    Besides Feldenkrais, other body training I love include Parkour (if you are in NYC, search for NY Parkour), Brazilian Jiu-Jitsu and rock climbing. Or just walk and feel the body's flow.

    Authentic relating

    I completed level 1 and level 2 Authentic Relating Training at ART International (2020-2021). It is a transformative framework that helped me to be more authentic yet accepting when relating with others. It freed me from a nerd mostly living in my own head to be more perceptive about emotion and other people's deeper being beyond language. The very loving and supporting community also helped me to gain more confidence of being my real self yet still able to connect with others.

    Another resource in the same vein I found very helpful is Non-Violent Communication. It's simple yet highly insightful. Practicing it helped a lot with my relationship with family and partner.

    As an explorer

    I'll keep adding things I spend time learning about in this section.

    Blockchain and decentralized organization

    Yes I was attracted by the buzzword. Yes, based on my background, it's not surprising that I'm more interested in the human part than the money part. Currently I'm researching the reward structure design of DAOs (Decentralized Autonomous Organization), working on the data analysis project with Token Engineering Academy.

    (I've also completed RAI workshop from the Token Engineering Academy learning about the details of this protocol model)

    This is a puzzling world with so many things to learn. One great resource I liked is Kernel which goes much deeper into the idealistic side of the blockchain world (also people call it "Web3"?). Scientists may be especially interested in the projects that Opscientia is doing.

    But really, what I found most useful is just talking to people in this field! People are so open in sharing their knowledge and passion.