Aykut Erdem

Can Your Model Separate Yolks with a Water Bottle? Benchmarking Physical Commonsense Understanding in Video Generation Models

Geometry-Aligned Tangent-Plane Diffusion Transformers for 360° Panorama Generation

Learning to Think Like a Cartoon Captionist: Humor Understanding With Multimodal Reasoning Models

CLIPAway: Harmonizing focused embeddings for removing objects via diffusion models

ViLMA: A Zero-Shot Benchmark for Linguistic and Temporal Grounding in Video-Language Models

Compositional Generalization in Multimodal Foundation Models