Research Projects

From Navigation to Refinement: Revealing the Two-Stage Nature of Flow-based Diffusion Models through Oracle Velocity Haoming Liu, Jinnuo Liu, Yanhao Li, Liuyang Bai, Yunkai Ji, Yuanhe Guo, Shenji Wan, Hongyi Wen cvpr2026 2025/09-2026/04

TL;DR: Derives an oracle flow-matching velocity that exposes early navigation for layout and generalization, then later refinement for detail and memorization.

In Proceedings of The IEEE/CVF Conference on Computer Vision and Pattern Recognition 2026

Project Page Paper

ImageGem: In-the-wild Generative Image Interaction Dataset for Generative Model Personalization Yuanhe Guo^*, Linxi Xie^*, Zhuoran Chen, Kangrui Yu, Ryan Po, Guandao Yang, Gordon Wetzstein, Hongyi Wen iccv2025 2024/05-2025/09

💡 Following up work of GEMRec, but focused on LoRA

TL;DR: Introduces a large-scale real-world interaction dataset for learning fine-grained user preferences in image generation, retrieval, recommendation, and LoRA personalization.

In Proceedings of the 2025 International Conference on Computer Vision

Project Page Code & Dataset Paper

Diffusion Cocktail: Fused Generation from Diffusion Models Haoming Liu, Yuanhe Guo, Hongyi Wen 2023/11-2024/05

TL;DR: Mixes content and style signals across domain-specific diffusion models without training, enabling controllable generations unavailable from a single checkpoint.

Extension for A1111 Webui has been released as sd-webui-ditail

Project Page Code Paper Demo

GEMRec: Towards Generative Model Recommendation Yuanhe Guo, Haoming Liu, Hongyi Wen wsdm2024 2023/05-2023/12

TL;DR: Frames generative model recommendation as personalized prompt-model retrieval and generated item ranking, supported by the GEMRec-18K interaction dataset.

In Proceedings of the 17th ACM International Conference on Web Search and Data Mining

Code Paper Dataset Demo

The Verse NO.1555 - Virtual Campus Chang Liu, Xu Li, Li Zhang, Dainyu Lin, Sinuo Chen, Yuahng Zheng, Shengyang Peng, Anthony Oury, Yuanhe Guo, Xingchen Zhang 2022/09-2023/05

TL;DR: A project of making a virtual NYU Shanghai century avenue campus, as part of NYU Shanghai 10th anniversary celebration

Media reports for the project:
[NYU Shanghai News]Creating A Digital Memory Map of Century Avenue Campus
[NYUSHDIC]Verse: No.1555

AR Audio Reactive Visual Effect for Auditorium Performance in Realtime Yuanhe Guo, Morui Yu, Marcela Godoy 2022/06-2022/09

TL;DR: A summer project funded by Dean's Undergraduate Research Fund (DURF)

Awarded as Fall 2022 Undergraduate Research Symposium Best Presentation in STEM & Media

Showcase:
A Collaboration with NYU Shanghai Reads and NYU Shanghai Performance Team

Course Projects

Deep Learning - Final Competition Yuanhe Guo, Jianing Zhang, Haotong Wu, Alfredo Canziani, Mengye Ren, Yann LeCun Leaderboard: 2nd 2024/04-2024/05

TL;DR: Predicts future segmentation masks from video by combining U-Net pseudo-labels with a SimVP-style world model, ranking 2nd on the course leaderboard.

Course project for CSCI-GA 2572 Deep Learning, 2024 Spring

Code

Computer Vision - Applying ViT in Generalized Few-shot Semantic Segmentation Liyuan Geng, Jinhong Xia, Yuanhe Guo, Rob Fergus 2023/11-2023/12

TL;DR: Applied Vision Transformers to generalized few-shot semantic segmentation for adapting segmentation models to novel visual categories.

Course project for CSCI-GA 2271 Computer Vision, 2023 Fall

Code Paper

Database - Air Ticket Reservation System Yuanhe Guo, Jinhong Xia, Lihua Xu 2023/04-2023/05

TL;DR: Built a database-backed air ticket reservation system covering flight search, booking, ticket management, and user workflows.

Course project for CSCI-SHU 213 Database Systems, 2023 Spring

Code

Machine Learning - Fine-Tuning a Diffusion Model For Doodles Generation Yuanhe Guo, Yiling Cao, Li Guo 2023/04-2023/05

TL;DR: Fine-tuned a diffusion model to generate doodle-style images from prompts and explore lightweight personalization for visual generation.

Course project for CSCI-SHU 360 Machine Learning, 2023 Spring

Code Video

Introduction to Computer Science and Data Science - GameWiz, an Online Multiplayer Quiz Game Yuanhe Guo, Yiling Cao, Xianbin Gu 2022/04-2022/05

TL;DR: A socket-based desktop Multiplayer quiz game.

Course project for CSCI-SHU 101 Introduction to Computer Science and Data Science

Code Video