Speech · Audio · Multimodal Learning

Runqiu Xu 徐润秋

I am a Visiting Student Researcher at the Speech, Audio, and Language Technologies (SALT) Lab at the University of Texas at Austin, working with Professor David Harwath. I recently completed my M.S. in Data Science at the University of Chicago. Before that, I earned a B.S. in Applied Mathematics and a B.A. in Finance at Wuhan University.

Portrait of Runqiu Xu

Research

I study how machines can reason with who is speaking, what is being said, and how multimodal interaction unfolds over time.

Preprints

Voices as Handles: Reasoning about Speaker Identity with Frozen Text LLMs

Runqiu Xu, Zhisheng Zheng, and David Harwath

Preprint, 2026 arXiv GitHub (to be released once accepted)

Publications

Silence-the-Mimic: Accelerating Imperceptible Perturbation Generation Against Voice Cloning

Runqiu Xu

IEEE Spoken Language Technology (SLT), 2026 · To appear arXiv GitHub

Regional Heterogeneity Modeling for Spatiotemporal Crop Yield Forecasting

Heng Xiong, Runqiu Xu, and Wenjun Jiang

Communications in Statistics – Theory and Methods, 2026 Paper

Presentations

Evaluation of Driving Behavior and Risk Rating with Machine Learning Techniques Based on Telematics Data

Runqiu Xu, Wenjun Jiang, Rogemar Mamon, Heng Xiong

International Conference on Econometrics and Statistics (EcoSta 2023)

An Improved Neural Network Approach for Agricultural Index Insurance Design

Runqiu Xu, Heng Xiong, Wenjun Jiang

26th International Congress on Insurance: Mathematics and Economics (IME 2023)

Experience

2026 — Present

The University of Texas at Austin

Visiting Student Researcher · SALT Lab

Education

2024 — 2026

The University of Chicago

M.S. in Data Science

2020 — 2024

Wuhan University

B.S. in Applied Mathematics and B.A. in Finance

Fun Corner

High-school me made a Steam game, Oneness. Email me for a free key :)