cv

Ph.D. (expected June 2028)

Georgia Institute of Technology

Computer Science

• Advisor: Dr. Cindy Xiong Bearfield

M.S.

Georgia Institute of Technology

Computer Science, Visual Analytics

B.Eng.

Shanghai Jiao Tong University

Electrical and Computer Engineering

• Advisor: Dr. Shuai Li

Course Instructor

Georgia Institute of Technology

Introduction to Information Visualization (CS 4460)

Teaching Assistant

Georgia Institute of Technology

Introduction to Information Visualization (CS 4460)

Embedded Software Engineer Intern

Bosch China

Deep Learning-based Gesture Recognition Algorithm Development

• Developed gesture recognition algorithms for Human-Vehicle Interaction using DNN.
• Applied the attention network to the neural network for dynamic gesture classification.
• Achieved 90% accuracy for 16 static gestures and 9 dynamic gestures on webcam with 720p@30fps.

CoCo Grant ($1,000)

Predict Reader Insights in Visualization by Modeling Cognitive Affordances. PI: Songwen Hu. Faculty advisors: Dr. Cindy Xiong Bearfield and Dr. Apurva Ratan Murty.

MathWorks Gift Fund ($2,000)

Gramm Extension for Exporting Function Compatibility. Working on MATLAB’s open-source visualization library Gramm to support web interactivity.

Interaction Modulates Visual Interpretation and Trust in Uncertainty Visualizations

Songwen Hu, Alex Endert, Cindy Xiong Bearfield

Vision Sciences Society (VSS) 2026

Oral presentation and poster.

Visual Embellishments are Potential Distractions in Double-Column Reading

Songwen Hu, Chase Stokes, Marti Hearst, Cindy Xiong Bearfield

IEEE VIS 2026, VISxVISION workshop

To be presented at VIS 2026. Oral presentation.

Interactive Visualization Recommendation with Hier-SUCB

Songwen Hu, Ryan A. Rossi, Tong Yu, Junda Wu, Handong Zhao, Sungchul Kim, Shuai Li

International World Wide Web Conference 2025 (WWW 25')

Accepted by International World Wide Web Conference 2025

VisChatter: Enhance Synchronous Collaboration on Visualization Dashboard through Visual Annotations

Songwen Hu, Tong Yu, Ryan A. Rossi, Sungchul Kim, Cindy Xiong Bearfield

Annual Meeting of the Cognitive Science Society 2025 (CogSci 25')

Presented at CogSci 2025. Oral presentation and poster.

Perceptual Benefits of Animation are Task-Dependent: Effects of Staging and Tracing in Dynamic Displays

Songwen Hu, Ouxun Jiang, Jeffrey Riedmiller, Cindy Xiong Bearfield

IEEE Visualization Conference 2024 (VIS 24')

Presented at IEEE VIS 2024; also presented as a poster at VSS 2024.

Hierarchical Conversational Preference Elicitation with Bandit Feedback

Jinhang Zuo, Songwen Hu, Shuai Li, Tong Yu, Handong Zhao, Carlee Joe-Wong

Conference on Information and Knowledge Management 2022 (CIKM 22')

Accepted by Conference on Information and Knowledge Management 2022

Reverse-Engineering Insight Formation from Visualizations using Brain and Eye Signals

Advisors: Cindy Xiong Bearfield (Georgia Tech), Ratan Murty (Georgia Tech)

• Investigating how low-level perceptual features interact with high-level reasoning processes in visualization comprehension.
• Combining fMRI and eye-tracking to quantify perceptual and cognitive load across visual cortex and frontal cortex.
• Developing a computational model to predict and interpret perceptual-cognitive demands from visualizations.

VisChatter – Visual Annotations for Collaborative Analytics with Multimodality

Advisors: Cindy Xiong Bearfield (Georgia Tech), Tong Yu (Adobe)

• Designed and implemented an interactive visualization dashboard with real-time annotation for synchronous collaborative analytics.
• Integrated speech recognition, LLM-based keyword extraction, and custom annotation APIs for seamless multimodal input.
• Conducted controlled A/B testing, measuring user engagement and task efficiency improvements.

Interactive Web Integration for Gramm Visualizations in MATLAB

Advisors: Cindy Xiong Bearfield (Georgia Tech), Vijay Iyer (MathWorks), Pierre Morel (ULCO)

• Extended Gramm, an open-source MATLAB visualization library, with a feature to export plots as interactive web visualizations.
• Integrated Vega and SVG-based interactivity to enable zooming, panning, and tooltips directly in the browser.
• Collaborated with the open-source community to design, implement, and document the new functionality.

Assessing Immersiveness of Visual Storytelling in VR with Bio-signals

Advisors: Jeffery Riedmiller (Dolby Research), Cindy Xiong Bearfield (Georgia Tech)

• Developed a pipeline to collect and analyze physiological signals during VR storytelling.
• Piloting VR scenes varying in narrative structure and visual intensity.
• Conducting data-driven evaluation of perceived immersion and emotional response.

Vision Rehabilitation through Multisensory Interaction

Advisors: Tina Liu (Georgetown University), Cindy Xiong Bearfield (Georgia Tech)

• Built interactive prototypes for training spatial attention and visual reconstruction.
• Integrated visual and auditory cues to enhance sensory compensation and engagement.
• Piloting test sessions with clinical collaborators to assess user feedback and accessibility.

Assessing Cognitive Load of Word-scale Visualizations

Advisors: Marti Hearst (UC Berkeley), Cindy Xiong Bearfield (Georgia Tech)

• Designed and executed a controlled eye-tracking study to assess how embedded visual embellishments affect reading comprehension across varied data densities.
• Quantified cognitive disruption by analyzing psychophysiological metrics (eye saccades).
• Explored word-scale visualization design guidelines for fluent reading flow.