Research Experience

iDEA-iSAIL Lab

Research Intern, Feb 2026 - present. Advised by Prof. Hanghang Tong.

• Conducted research in theoretical reinforcement learning, studying sample and computation efficient algorithms for online learning in finite-horizon Markov Decision Processes under function approximation and stochastic dynamics.
• Used concentrability to develop the first sample and computation efficient algorithm for online reinforcement learning in stochastic-transition MDPs under the linear Q^π realizability assumption, resolving an open problem left by prior work restricted to deterministic dynamics.
• Established a PAC sample complexity guarantee of Õ(d²H⁷C^*/ε³) by combining techniques from both offline and online reinforcement learning theory, dynamic programming, concentration analysis, and linear function approximation.
• Manuscript in preparation: Frozen Policy Iteration for MDPs with Stochastic Transitions under Linear Q^π Realizability.

Illinois Combinatorics Lab for Undergraduate Experience (ICLUE)

Research Intern, Aug 2024 - present. Advised by Prof. Alexander Yong.

• Proved a new asymptotic result that the probability that the Schensted insertion algorithm for a uniformly random permutation in S_n exhibits a special "bumping" interaction converges to 1 as n approaches infinity.
• Derived the result using rigorous probability bounds on the RSK linear operator by leveraging the Vershik-Kerov-Logan-Shepp limit shape curve as the central analytic tool.
• Developed proofs requiring techniques spanning real analysis (uniform convergence, limit arguments), probability theory (concentration inequalities, convergence in probability), and asymptotic analysis (bounding combinatorial quantities in the large-n regime).
• Paper in submission to the Electronic Journal of Combinatorics.

University of California San Diego

Research Intern, Jun 2023 - Dec 2024. Advised by Prof. Xinyu Zhang.

• Developed EveGuard, a software-based defense framework to protect voice privacy from vibrometry-based side channels using adversarial audio by using a perturbation generator model (PGM) to effectively suppress sensor-based eavesdropping while preserving high audio quality.
• Implemented Eve-GAN, a novel domain translation task for inferring eavesdropped signals, enabling end-to-end training of PGM, utilizing few-shot learning techniques to reduce data collection overhead.
• Achieved a protection rate of over 97% against audio classifiers, hindering eavesdropped reconstruction.
• Paper published in the IEEE Symposium on Security and Privacy 2025.

Illinois Mathematics Lab (formerly Illinois Geometry Lab)

Research Intern, Jan 2024 - present. Advised by Prof. AJ Hildebrand.

• Researched the data science perspective of sports matches, analyzing the predictability of outcomes by comparing prediction accuracies across polls, the betting market, Elo ratings, and mathematical models like Bradley-Terry.
• Presented findings at Rose-Hulman Undergraduate Math Conference 2024, UIUC Undergraduate Research Symposium 2024, and Joint Mathematics Meeting 2025.
• Paper in submission to the Mathematics and Sports (MAS) Journal.

Carnegie Mellon University Human-Computer Interaction Institute (HCII)

Research Intern, May 2025 - Feb 2026. Advised by Prof. Jeff Bigham.

• Designed and evaluated an automated journalism pipeline using SOTA LLMs to replicate end-to-end editorial workflows, from transcript segmentation to headline generation and topic prioritization.
• Orchestrated large-scale crowd-sourced study and demonstrated that LLM-based headline quality and topic prioritization can meet and exceed professional standards.
• Paper in Human-centered Evaluation and Auditing of Language Models (HEAL) Workshop at CHI 2026.

National University of Singapore SEEDER Group

Research Intern, May 2024 - Aug 2024. Advised by Prof. Xuanyao Fong.

• Developed machine learning pipeline to detect, classify, and analyze diseases in plant leaves using U-Net segmentation and CNN classification models, achieving 90% segmented classification rate.
• Stacked deep learning models with traditional ML models and augmented training data with cGAN-generated synthetic images, achieving over 95% detection accuracy.

Presentations

• [2025] Carnegie Mellon University Human-Computer Interaction Institute REU Symposium

  Title: LLMs for Extracting Agenda Items from City Council Meetings and Generating Newsworthy Headlines
  Authors: David Xia
  Program: CMU HCII REU
  Presentation: Poster

• [2025] Algebra, Geometry and Combinatorics Day (joint work presented by Duy Phan)

  Title: RSK linear operators and the Vershik-Kerov-Logan-Shepp curve
  Authors: Duy Phan, David Xia
  Program: Algecom
  Presentation: Poster

• [2025] Joint Mathematics Meeting

  Title: Predictability in Major Sports Leagues: Betting Market vs Model Based Predictions
  Authors: David Xia
  Program: Joint Mathematics Meeting
  Presentation: Poster

• [2024] Illinois Mathematics Lab Research Symposium

  Title: Predictability in College Sports
  Authors: Yihan Gao, Samuel Lam, and David Xia
  Presentation: Poster

• [2024] Rose-Hulman Undergraduate Math Conference

  Title: Predictability in College Sports: Comparing the Accuracies of Prediction Models for College Football and Basketball Games
  Authors: Yihan Gao, Samuel Lam, and David Xia
  Program: Rose-Hulman Undergraduate Math Conference
  Presentation: Oral Talk

• [2024] University of Illinois Urbana-Champaign Undergraduate Research Symposium

  Title: Predictability and Competitiveness in Sports Leagues
  Authors: Yihan Gao, Samuel Lam, and David Xia
  Program: UIUC Undergraduate Research Symposium
  Presentation: Poster

Grants and Awards

• [2025] CMU HCII REU: NSF 2349558

• [2025] ICLUE RTG: NSF 1937241

Relevant Coursework

• CS: Machine Learning, NLP, Deep Learning, Algorithms, Numerical Methods, Numerical Analysis
• Math: Honors Real Analysis, Graph Theory, Combinatorics, Abstract Linear Algebra, Abstract Algebra I--II, Algebraic Combinatorics, Linear Programming
• Stats: Stochastic Processes, Applied Bayesian Analysis, Survival Analysis

Languages

• English: Native
• Mandarin Chinese: Fluent, HSK6

Links

Email

GitHub

LinkedIn

Google Scholar