Generative worlds. Human direction.

Generative Modeling (GEM) Lab is led by Dr. Pinar Yanardag in the Department of Computer Science at Virginia Tech. We build generative models for images, video, and interactive worlds that follow human direction: controllable at a fine-grained level, personal to the concepts you care about, and efficient enough to run over long horizons.

Publications Meet the team NeurIPS’26 Overview CVPR’26 Overview

25+Papers since Fall 2023

7PhD students

7+Industry partners

∞Generative ideas

Research, illustrated.

Our lab has led 25+ top-conference papers since Fall 2023.

A person shows the world model a photo of her dog, and the model projects a world with that dog in it, ready to explore with arrow keys.

SPAWN Your concepts, inside interactive worlds.

From Zero to Hero: Training-Free Custom Concept Spawning in World Models

Kiymet Akdemir, Pinar Yanardag

NeurIPS 2026Paper Project

A tired model slumps under a stack of memory boxes while a researcher with a budget clipboard points to a happy model holding one small memory block.

VideoMLA Compact memory for minute-scale video generation.

VideoMLA: Low-Rank Latent KV Cache for Minute-Scale Autoregressive Video Diffusion

Hidir Yesiltepe, Jiazhen Hu, Tuna Han Salih Meral, Adil Kaan Akan, Kaan Oktay, Hoda Eldardiry, Pinar Yanardag

NeurIPS 2026Paper Project

A determined model strains forward against a chain fixed to an anchor in the ground.

AdaState Evolving anchors for streaming video.

AdaState: Self-Evolving Anchors for Streaming Video Generation

Yusuf Dalva, Pinar Yanardag

NeurIPS 2026Paper Project

A cheerful model gets a bouncing ball right but fails at a cube floating in water and a sponge dripping water.

HiPhy Getting multiple physics principles right at once.

HiPhy: Hierarchical Alignment for Physically-Plausible Multi-Principle Video Generation

Tahira Kazimi, Shubhankar Borse, Munawar Hayat, Fatih Porikli, Pinar Yanardag

NeurIPS 2026Paper Project

More research.

Selected publications, newest first.

RAVEL: Rare Concept Generation and Editing via Graph-driven Relational Guidance

Kavana Venkatesh, Yusuf Dalva, Ismini Lourentzou, Pinar Yanardag

NeurIPS 2026Paper Project
Learn Once, Edit Anywhere: Visual Direction Transfer for Diffusion Models

Yusuf Dalva, Hidir Yesiltepe, Pinar Yanardag

ECCV 2026Paper Project
RankT2I: A Submodular Framework for Discovering Interpretable and Diverse Semantics in Text-to-Image Models

Ritika Allada, Pinar Yanardag

ECCV 2026Paper
∞-RoPE: Action-Controllable Infinite Video Generation Emerges From Autoregressive Self-Rollout

Hidir Yesiltepe, Tuna Han Salih Meral, Adil Kaan Akan, Kaan Oktay, Pinar Yanardag

CVPR 2026Paper Project
Diverse Video Generation with Determinantal Point Process-Guided Policy Optimization

Tahira Kazimi, Connor Dunlop, Pinar Yanardag

CVPR 2026Paper Project
LoRAShop: Training-Free Multi-Concept Image Generation and Editing with Rectified Flow Transformers

Yusuf Dalva, Hidir Yesiltepe, Pinar Yanardag

NeurIPS 2025, SpotlightPaper Project
Personalized Image Editing in Text-to-Image Diffusion Models via Collaborative Direct Preference Optimization

Connor Dunlop, Matthew Zheng, Kavana Venkatesh, Pinar Yanardag

NeurIPS 2025Paper Project
CREA: A Collaborative Multi-Agent Framework for Creative Content Generation with Diffusion Models

Kavana Venkatesh*, Connor Dunlop*, Pinar Yanardag

NeurIPS 2025Paper Project
Dynamic View Synthesis as an Inverse Problem

Hidir Yesiltepe, Pinar Yanardag

NeurIPS 2025Paper Project
CLoRA: A Contrastive Approach to Compose Multiple LoRA Models

Tuna Han Salih Meral*, Enis Simsar*, Federico Tombari, Pinar Yanardag

ICCV 2025, SpotlightPaper Project
LoRAverse: A Submodular Framework to Retrieve Diverse Adapters for Diffusion Models

Mert Sonmezer, Matthew Zheng, Pinar Yanardag

ICCV 2025Paper Project
FluxSpace: Disentangled Semantic Editing in Rectified Flow Transformers

Yusuf Dalva, Kavana Venkatesh, Pinar Yanardag

CVPR 2025Paper Project
LoRACLR: Contrastive Adaptation for Customization of Diffusion Models

Enis Simsar, Thomas Hofmann, Federico Tombari, Pinar Yanardag

CVPR 2025Paper Project
Explaining in Diffusion: Explaining a Classifier Through Hierarchical Semantics with Text-to-Image Diffusion Models

Tahira Kazimi*, Ritika Allada*, Pinar Yanardag

CVPR 2025Paper Project
Stylebreeder: Exploring and Democratizing Artistic Styles through Text-to-Image Models

Matthew Zheng*, Enis Simsar*, Hidir Yesiltepe, Federico Tombari, Joel Simon, Pinar Yanardag

NeurIPS 2024Paper Project
NoiseCLR: A Contrastive Learning Approach for Unsupervised Discovery of Interpretable Directions in Diffusion Models

Yusuf Dalva, Pinar Yanardag

CVPR 2024, OralPaper Project
CONFORM: Contrast is All You Need For High-Fidelity Text-to-Image Diffusion Models

Tuna Han Salih Meral, Enis Simsar, Federico Tombari, Pinar Yanardag

CVPR 2024Paper Project
RAVE: Randomized Noise Shuffling for Fast and Consistent Video Editing with Diffusion Models

Ozgur Kara*, Bariscan Kurtkaya*, Hidir Yesiltepe, James M. Rehg, Pinar Yanardag

CVPR 2024, HighlightPaper Project

Selected collaborations.

Projects led by our collaborators.

Canvas-to-Image: Compositional Image Generation with Multimodal Controls

Yusuf Dalva*, Gordon Guocheng Qian*, Maya Goldenberg, Tsai-Shien Chen, Kfir Aberman, Sergey Tulyakov, Pinar Yanardag, Kuan-Chieh Jackson Wang

SIGGRAPH 2026Paper Project
ConceptAttention: Diffusion Transformers Learn Highly Interpretable Features

Alec Helbling, Tuna Han Salih Meral, Benjamin Hoover, Pinar Yanardag, Duen Horng Chau

ICML 2025Paper Project

News

Recent milestones from the lab.

Sep 25, 2026Five papers accepted to NeurIPS 2026.
Mar 3, 2026Four papers accepted to CVPR 2026. See an overview of our CVPR presence.
Sep 25, 2025Four papers accepted to NeurIPS 2025, main conference.
Feb 3, 2025Three papers accepted to CVPR 2025.
Apr 3, 2024Two papers selected as Oral and Highlight at CVPR 2024.
Mar 3, 2024Three papers accepted to CVPR 2024.

Team

GEM Lab is led by Dr. Pinar Yanardag, with a group of PhD researchers at Virginia Tech.

Pinar Yanardag
Principal Investigator
Hidir Yesiltepe
PhD Student
Yusuf Dalva
PhD Student
Tuna Han Salih Meral
PhD Student
Kiymet Akdemir
PhD Student
Enis Simsar
PhD Student, co-advised at ETH Zurich
Tahira Kazimi
PhD Student
Connor Dunlop
PhD Student
Future Gem
We’re hiring. PhD students and collaborators welcome. Contact us

Alumni

Oguz Kaan Yuksel EPFL
Dilara Gokay DeepMind
Alara Dirik Imperial College London
Umut Kocasari TUM
Kerem Zaman UNC Chapel Hill
Mert Yuksekgonul Stanford
Zehranaz Canfes TUM
Furkan Atasoy TUM
Berkay Doner EPFL
Elif Sema Balcioglu EPFL
Eylul Yalcinkaya Meta
Cemre Efe Karakas Amazon
Merve Rabia Barin Purdue
Ezgi Gulperi Er Goldman Sachs
Ritika Allada US Government
Matthew Zheng JP Morgan

Before Virginia Tech, our group was CATLAB at Boğaziçi University. Visit the site to see our earlier research on GAN controllability and interpretability.

Let’s collaborate

We work across industry and academia, and we’re always open to new partnerships.

GEM Lab partners with leading companies and research labs on generative-AI research, internships, and joint projects. If our work resonates with yours, we’d love to hear from you.

NSF
Google Research
Adobe Research
Snap Inc.
Qualcomm Research
fal
Deloitte

Email us to collaborate

Connect

Drop us an email or find us on social.