Xuxin Tang

Human-AI Interaction Researcher

(How to pronouce my first name: “Sheshin”.)

Hello! 👋 I’m a Computer Science PhD student at Virginia Tech’s Sanghani Center for Artificial Intelligence & Data Analytics, advised by Prof. Chris North. Leveraging three years of industrial experience in full-stack and machine learning engineering, I design and implement human-AI collaborative systems using state-of-the-art LLM and agentic frameworks. I bridge the gap between AI reasoning and human oversight by architecting intuitive visual interfaces that prioritize system transparency, generalizability, and operational efficiency. My unique background in large-scale recommendation systems allows me to transition abstract cognitive frameworks into reliable, real-world applications.

Before joining Virginia Tech, I spent nearly three years in industry, starting as a full-stack development engineer before advancing to machine learning engineering where I worked on large-scale recommendation systems. I hold both B.Eng. and M.S. degrees from Wuhan University.

👉Explore my latest system, S-PRISM, to see human-AI collaboration in action. Simply click ‘Report Generation’ from the header, recognize the space, and re-run the generation to experience the system’s responsiveness and transparency firsthand.

Google Scholar/Twitter/LinkedIn

News

Aug 2, 2026	Our paper CriterionSI has been accepted by IEEE VIS 2026! Explore our work and code.
Mar 2, 2026	Our paper on Semantic Prompting framework and agentic S-PRISM system is accepted by ACM AVI 2026! You can explore the system here: Link.
Jun 17, 2025	Our paper on AI Fairness is officially accepted by IEEE JBHI 2025! Check out our work: Link
Mar 25, 2025	Our paper Imprinto is accepted by CHI 2025!
Oct 30, 2024	Exited to present my research on visual analytics and human-LLM collaboration to Bloomberg UX Research Group!

Selected Publications

2026

Drag, Infer, Reproject: Grounding LLMs through Spatial Interaction for Image Clustering

Liu Yang, Xuxin Tang, Xu Jiahao, and 1 more author

IEEE VIS, 2026

PDF
Semantic Prompting: Agentic Incremental Narrative Refinement through Spatial Semantic Interaction

Xuxin Tang, Ibrahim Tahmid, Eric Krokos, and 3 more authors

ACM AVI, 2026

HTML PDF

2025

Enhancing Fairness and Accuracy in Diagnosing Type 2 Diabetes in Young Population

Tanmoy Sarkar Pias, Yiqi Su, Xuxin Tang, and 3 more authors

IEEE JBHI, 2025

HTML
Imprinto: Enhancing Infrared Inkjet Watermarking for Human and Machine Perception.

Martin Feick, Xuxin Tang, Raul Garcia-Martin, and 5 more authors

ACM CHI, 2025

HTML PDF

2024

Steering LLM Summarization with Visual Workspaces for Sensemaking

Xuxin Tang, Eric Krokos, Can Liu, and 4 more authors

IEEE VIS NLVIZ Workshop, 2024

PDF

2019

Quantifying cloud elasticity with container-based autoscaling

Fan Zhang, Xuxin Tang, Xiu Li, and 2 more authors

Future Generation Computer Systems, 2019

Abs PDF

Containers have been a pervasive approach to help rapidly develop, test and update the Internet of Things applications (IoT). The autoscaling of containers can adaptively allocate computing resources for various data volumes over time. Therefore, elasticity, a critical feature of a cloud platform, is significant to measure the performance of lightweight containers. In this paper, we propose a framework with container auto- scaler. It monitors containers resource usage and accordingly scales in or scales out containers in need. Further, we define elasticity mathematically in order to quantify the cloud elasticity using the proposed framework. Extensive experiments are carried out with different workload modes, workload durations, and scaling cool-down period of times. Experiment results show that the framework captures the workload variation firmly with a very short delay. We also find out that the cloud platform shows the best elasticity in repeat workload mode due to its recurring and predictable feature. Finally, we discover the length of the cool-down period should be properly set up in order to balance system stability and good elasticity.

2017

Quantifying Cloud Elasticity with Container-Based Autoscaling

Xuxin Tang, Fan Zhang, Xiu Li, and 2 more authors

15th Intl Conf on Pervasive Intelligence and Computing (PICom 2017), 2017

Abs PDF

Containers have been a pervasive approach to help rapidly develop, test and update the Internet of Things applications (IoT). The autoscaling of containers can adaptively allocate computing resources for various data volumes over time. Therefore, elasticity, a critical feature of a cloud platform, is significant to measure the performance of lightweight containers. In this paper, we propose a framework with container auto-scaler. It monitors containers resource usage and accordingly scales in or scales out containers in need. Further, we define elasticity mathematically in order to quantify the cloud elasticity using the proposed framework. Extensive experiments are carried out with different workload modes, workload durations, and scaling cool- down period of times. Experiment results show that the framework captures the workload variation firmly with a very short delay. We also find out that the cloud platform shows the best elasticity in repeat workload mode due to its recurring and predictable feature. Finally, we discover the length of the cool- down period should be properly set up in order to balance system stability and good elasticity.
A night image enhancement algorithm based on guided filtering

Xuxin Tang, Zhijiang Li, and Yuhang Chen

Advanced Graphic Communications and Media Technologies, 2017

Abs PDF

Colored night images lead to various problems in real life, such as low dynamic range, detail blurring and so on, which are waiting to be solved. For this reason, it proposes an algorithm for night image enhancement based on guided filtering in the paper. This image process includes three main steps. Firstly, by transforming to HVS color space, the targeted image is separated coarsely into two images—base layer image and detail layer image in the result of guided filtering. In the second step, the base layer and detail layer are processed by different strategies selectively and respectively to lighten the whole, but remain details as much as possible. Eventually, the base layer, which has been previously increased contrast, is added detail layer to keep the details and margins. The result image can be retained after reverting to the original color space. Our result illustrates the com- parison among result images and images enhanced by other methods, which obviously proves that the algorithm in our paper can obtain relatively better results in the ways of detail enhancement and color fidelity.