Avatar

Yanna Ding

Email: dingy6@rpi.edu

About Me

I am a PhD student in Computer Science at Rensselaer Polytechnic Institute advised by Professor Jianxi Gao. I earned my B.S. in Computer Science and Mathematics from the University of Toronto, where I had the privilege of being mentored by Professor Ishtiaque Ahmed.

I’m interested in the dynamics and learning behaviors of complex systems, spanning networked dynamical processes and modern language models. My work focuses on networked dynamical systems, including reverse engineering, network inference, and time series prediction, with applications to understanding neural network training dynamics. I am also broadly interested in exploring the behaviors of language models, such as in-context learning and task-level inference.

Publications

  • Less is More: Efficient Weight “Farcasting” with 1-Layer Neural Network
    Xiao Shou, Debarun Bhattacharjya, Yanna Ding, Chen Zhao, Rui Li, Jianxi Gao Accepted by International Conference on Database Systems for Advanced Applications (DASFAA) 2025

  • Predicting Time Series of Networked Dynamical Systems without Knowing Topology
    Yanna Ding, Zijie Huang, Malik Magdon-Ismail, Jianxi Gao
    arXiv
    [PDF]

  • EMO: Epigraph Based Multilevel Optimization For Enhancing Chain Of Thought Reasoning Capabilities
    Songtao Lu, Yanna Ding, Lior Horesh, Jianxi Gao, Malik Magdon-Ismail
    Accepted by IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) 2025

  • Architecture-Aware Learning Curve Extrapolation via Graph Ordinary Differential Equation
    Yanna Ding, Zijie Huang, Xiao Shou, Yihang Guo, Yizhou Sun, Jianxi Gao
    Accepted by AAAI Conference on Artificial Intelligence (AAAI) 2025
    [PDF]

  • Inferring from Logits: Exploring Best Practices for Decoding-Free Generative Candidate Selection
    Mingyu Derek Ma, Yanna Ding, Zijie Huang, Jianxi Gao, Yizhou Sun, Wei Wang
    Accepted by the 4th Efficient Natural Language and Speech Processing workshop (NeurIPS ENLSP) 2024

  • Efficient parameter inference in networked dynamical systems via steady states: A surrogate objective function approach integrating mean-field and nonlinear least squares
    Yanna Ding, Jianxi Gao, Malik Magdon-Ismail
    Published in Phys. Rev. E 109, 034301 (2024)
    [PDF] [DOI]

  • Learning Network Dynamics from Noisy Steady States
    Yanna Ding, Jianxi Gao, Malik Magdon-Ismail
    Published in IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining (ASONAM) 2023
    [DOI]

Earlier Projects

  • Inter-coder Reliability

    University of Toronto Supervisor: Prof. Priyank Chandra Summer 2020

    During research analysis, such as categorizing social media posts, different researchers often have to label a shared dataset. It is essential to determine inter-coder reliability in order to make the research results more convincing in this case. This project implements a standalone application to calculate different inter-rater reliability metrics (e.g., Scott’s Pi).

    Your GIF 1

  • COVID-19 Related Stigma On Social Media

    University of Toronto Supervisor: Prof. Syed Ishtiaque Ahmed Summer 2020 - Present

    The COVID-19 pandemic has led to increased stigma, prejudice, and hate against people of East Asian descent living in North America and around the world. In particular, those individuals perceived to be from China or to a Chinese-heritage are frequently targeted on social media platforms such as Twitter, Facebook, and Instagram. This project studies the stigma, fear, and misinformation on social media related to COVID-19. It aims to increase public awareness of the harm and impact of stigma and reduce the spread of stigma. During the project, I collected and maintained a dataset of 650+ million tweets, including 220k potentially stigmatizing tweets with geo-location data; visualized and analyzed various statistical metrics on the previous data, including distribution of sentiment (labeled by the BERT model) around the globe.

    Your GIF

Awards