GEMLAB
Pre-prints
-
Arxiv CLoRA: A Contrastive Approach to Compose Multiple LoRA ModelsPre-print (Arxiv) , 2024 -
Arxiv MIST: Mitigating Intersectional Bias with Disentangled Cross-Attention Editing in Text-to-Image Diffusion ModelsPre-print (Arxiv) , 2024 -
Arxiv GANTASTIC: GAN-based Transfer of Interpretable Directions for Disentangled Image Editing in Text-to-Image Diffusion ModelsPre-print (Arxiv) , 2024
Publications
-
CVPR NoiseCLR: A Contrastive Learning Approach for Unsupervised Discovery of Interpretable Directions in Diffusion ModelsConference on Computer Vision and Pattern Recognition (CVPR) , 2024 -
CVPR CONFORM: Contrast is All You Need For High-Fidelity Text-to-Image Diffusion ModelsConference on Computer Vision and Pattern Recognition (CVPR) , 2024 -
CVPR RAVE: Randomized Noise Shuffling for Fast and Consistent Video Editing with Diffusion ModelsConference on Computer Vision and Pattern Recognition (CVPR) , 2024 -
WACV Fantastic Style Channels and Where to Find Them: A Submodular Framework for Discovering Diverse Directions in GANsWinter Conference on Applications of Computer Vision (WACV) , 2023 -
WACV Text and Image Guided 3D Avatar Generation and ManipulationWinter Conference on Applications of Computer Vision (WACV) , 2023 -
ECCV FairStyle: Debiasing StyleGAN2 with Style Channel ManipulationsEuropean Conference on Computer Vision (ECCV) , 2022 -
CVPR Workshop PaintInStyle: One-Shot Discovery of Interpretable Directions by PaintingConference on Computer Vision and Pattern Recognition (CVPR Workshop) , 2022 -
CVPR Workshop Rank in Style: A Ranking-based Approach to Find Interpretable DirectionsConference on Computer Vision and Pattern Recognition (CVPR Workshop) , 2022 -
ACM CC MIDISpace: Finding Linear Directions in Latent Space for Music Generation (Honorable Mention Award)ACM Creativity & Cognition (ACM CC) , 2022 -
WACV StyleMC: Multi-Channel Based Fast Text-Guided Image Generation and ManipulationWinter Conference on Applications of Computer Vision (WACV) , 2022 -
ICCV LatentCLR: A Contrastive Learning Approach for Unsupervised Discovery of Interpretable DirectionsInternational Conference on Computer Vision (ICCV) , 2021 -
ICCV Workshop Graph2Pix: A Graph-Based Image to Image Translation Framework (Full Paper)Advances of Image Manipulation (ICCV Workshop) , 2021 -
NeurIPS Workshop Controlled Cue Generation for Play Scripts (Best Paper Award)Controllable Generative Modeling in Language and Vision (NeurIPS Workshop) , 2021 -
NeurIPS Workshop Exploring Latent Dimensions of Crowd-sourced CreativityMachine Learning for Creativity and Design (NeurIPS Workshop) , 2021
News
Mar 3, 2024 | Three papers are accepted to CVPR’24! |
---|---|
Jan 1, 2023 | Two papers are accepted to WACV’23! |
Jun 28, 2022 | Midispace received Honorable Mention Award at ACM Creativity & Cognition Conference! |
Jan 28, 2022 | Our undergraduate researchers Cemre and Eylul won the best senior thesis award! |
Dec 13, 2021 | Controlled Cue Generation for Play Scripts is selected as the Best Paper at CTRLGen Workshop at NeurIPS’21! |
Team
Principal Investigator: Dr. Pinar YanardagCurrent PhD Students: Tuna Han Salih Meral, Yusuf Dalva, Kiymet Akdemir, Hidir Yesiltepe
Alumi: Enis Simsar (now at ETH), Oguz Kaan Yuksel (now at EPFL), Dilara Gokay (now at DeepMind), Alara Dirik (now at Imperial College London), Umut Kocasari (now at TUM), Kerem Zaman (now at UNC Chapel Hill), Mert Yuksekgonul (now at Stanford), Zehranaz Canfes (now at TUM), Furkan Atasoy (now at TUM), Berkay Doner (now at EPFL), Elif Sema Balcioglu (now at EPFL), Eylul Yalcinkaya (now at Facebook), Cemre Efe Karakas (now at Amazon), Merve Rabia Barin (now at Koc), Ezgi Gulperi Er (now at Goldman Sachs).