# Pinar Yanardag, Ph.D.

> Assistant Professor of Computer Science at Virginia Tech and Director of GEMLAB. Research on generative AI: making image and video diffusion models controllable, interpretable, and accessible. Previously a postdoc at MIT Media Lab and a Ph.D. from Purdue University.

## About

- Tenure-track Assistant Professor, Department of Computer Science, Virginia Tech (Blacksburg, VA 24061).
- Director of GEMLAB and member of the Sanghani Center for AI and Data Analytics.
- Ph.D. in Computer Science, Purdue University (advisor: S.V.N. Vishwanathan); M.S. at Boğaziçi University (advisor: Ethem Alpaydın); postdoc at MIT Media Lab.
- Recent honor: NSF CAREER Award (2026). Fulbright Fellow and Google Anita Borg Scholar.
- Past life: security researcher and core developer of the Pardus GNU/Linux distribution; CEO of AI Fiction; co-founder of GLITCH (first generative-AI clothing line); Emmy-nominated Creative Director.
- Contact: pinary@vt.edu

## Research

Three thrusts, with project pages:

- Discovering and leveraging latent capabilities for efficient, training-free control of image/video synthesis: CONFORM (CVPR 2024), FluxSpace (CVPR 2025), NoiseCLR (CVPR 2024), RAVE (CVPR 2024), MotionFlow (AAAI 2026), Inverse-DVS (NeurIPS 2025), Infinity-RoPE (CVPR 2026).
- Integrating cross-modal capabilities (VLMs/LLMs with visual generative priors): CREA (NeurIPS 2025), DiffEx (CVPR 2025), C-DPO (NeurIPS 2025), DPP-GRPO (CVPR 2026).
- Personalized, human-aligned, and democratized generative AI: LoRACLR (CVPR 2025), CLoRA (ICCV 2025), LoRAShop (NeurIPS 2025), Plot'n Polish (AAAI 2026), Stylebreeder (NeurIPS 2024).

## Selected publications

- [Infinity-RoPE: Action-Controllable Infinite Video Generation (CVPR 2026)](https://arxiv.org/abs/2511.20649)
- [Diverse Video Generation with DPP-Guided Policy Optimization (CVPR 2026)](https://arxiv.org/abs/2511.20647)
- [CLoRA: A Contrastive Approach to Compose Multiple LoRA Models (ICCV 2025, Highlight)](https://arxiv.org/abs/2403.19776)
- [LoRAShop: Training-Free Multi-Concept Image Generation and Editing (NeurIPS 2025, Spotlight)](https://arxiv.org/abs/2505.23758)

## Links

- [GEMLAB website (full publication list, code, datasets, demos)](https://gemlab-vt.github.io/)
- [GEMLAB GitHub](https://github.com/gemlab-vt)
- [Google Scholar](https://scholar.google.com/citations?user=qzczdd8AAAAJ)
- [X (Twitter)](https://x.com/pinguar)