ViSTA: Visual Storytelling using Multi-modal Adapters for Text-to-Image Diffusion Models
2026IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) · Oral
CV
Research Focus
Multimodal generative AI, controllable visual generation, diffusion models, multimodal reasoning, LLM/MLLM post-training, and agentic AI systems.
Education
- Georgetown University, Ph.D. in Computer Science, Computer Vision Group, 2021–May 2027 (expected). Advisor: Sarah A. Bargal.
- Georgetown University, M.S. in Computer Science, 2019–2021.
- Beihang University, B.E. in Computer Science and Technology, 2015–2019.
Industry Research Experience
- Siemens, Data and AI Team, Cloud & Industrial GenAI Intern, May–Aug. 2026.
- Siemens Digital Industries Software, GenX Team, Cloud & Industrial GenAI Intern, May–Aug. 2025.
- ByteDance, Applied Machine Learning Team, Research Scientist Intern, May–Aug. 2023.
Additional Research Experience
- Beihang University, Engineering Research Center, Undergraduate Thesis: Automatic Scoring Method for Essay Review, Sep. 2018–Jun. 2019.
- University of California, Berkeley, Visual Computing Lab, Eyeglasses-free Vision Correcting Display, Jul.–Sep. 2018.
- Beihang University, Institute of Advanced Computing Technology, Real-Time Video Dehazing Based on Spatio-Temporal Markov Random Fields, Jan.–Apr. 2017.
Technical Skills
- Research: Multimodal generative AI, diffusion models, visual storytelling, vision-language and multimodal LLMs, agentic LLMs, reinforcement learning and post-training, text-to-SQL.
- Programming & frameworks: Python, C++; PyTorch, TensorFlow, Hugging Face Transformers, Diffusers; Apache Calcite.
- Languages: Mandarin Chinese (native), English (fluent).
Publications
Sidecar: Training-Free Semantic Reuse for Character-Consistent Free-form Visual Storytelling
2026Under review
FreeStory: Training-Free Character Consistency for Free-Form Visual Storytelling
2026Under review
D-Feat Occlusions: Diffusion Features for Robustness to Partial Visual Occlusions in Object Recognition
2025CVPR Workshop on Uncertainty Quantification for Computer Vision
Predicting Missing Response with BERT Model in Process Data
2024International Meeting of the Psychometric Society (IMPS)
SEINE: SEgment-based Indexing for NEural Information Retrieval
2022SIGIR Workshop on Reaching Efficiency in Neural Information Retrieval (ReNeuIR)
Do We Really Need Everything Everywhere All at Once? Query-Specific Fine-Tuning for Transformer-Based Neural Retrievers
2022Text REtrieval Conference (TREC)
GazBy: Gaze-Based BERT Model to Incorporate Human Attention in Neural Information Retrieval
2022ACM SIGIR International Conference on Theory of Information Retrieval (ICTIR)
Professional Service & Honors
- Reviewer: NeurIPS, ICLR, WACV, BMVC, SIGIR, CIKM, ICTIR, TOIS (2020–present).
- Teaching assistant: Science of AI (2026); Deep Learning (2025–2026); Deep Reinforcement Learning (2021–2023), Georgetown University.
- Honors: 2026 Conference Travel Grant; 2024 Outstanding Teaching Assistant.