How Compositional Generalization and Creativity Improve as Diffusion Models are Trained

v1v2v3 (latest)

How Compositional Generalization and Creativity Improve as Diffusion Models are Trained

17 February 2025

Alessandro Favero

Antonio Sclocchi

Francesco Cagnetta

Pascal Frossard

ArXiv (abs)PDF HTML

Papers citing "How Compositional Generalization and Creativity Improve as Diffusion Models are Trained"

8 / 8 papers shown

Title
GuessBench: Sensemaking Multimodal Creativity in the Wild Zifeng Zhu Shangbin Feng Herun Wan Ningnan Wang Minnan Luo Yulia Tsvetkov MLLM CoGe VLM 77 0 0 01 Jun 2025
What Do You Need for Diverse Trajectory Stitching in Diffusion Planning? Quentin Clark Florian Shkurti 766 0 0 23 May 2025
Bigger Isn't Always Memorizing: Early Stopping Overparameterized Diffusion Models Alessandro Favero Antonio Sclocchi Matthieu Wyart DiffM 63 0 0 22 May 2025
Scaling Laws and Representation Learning in Simple Hierarchical Languages: Transformers vs. Convolutional Architectures Francesco Cagnetta Alessandro Favero Antonio Sclocchi Matthieu Wyart 90 0 0 11 May 2025
Of All StrIPEs: Investigating Structure-informed Positional Encoding for Efficient Music Generation Manvi Agarwal Changhong Wang Gaël Richard 65 0 0 07 Apr 2025
Simplified and Generalized Masked Diffusion for Discrete Data Jiaxin Shi Kehang Han Zehao Wang Arnaud Doucet Michalis K. Titsias DiffM 201 105 0 17 Jan 2025
An analytic theory of creativity in convolutional diffusion models Mason Kamb Surya Ganguli DiffM 79 18 0 28 Dec 2024
Probing the Latent Hierarchical Structure of Data via Diffusion Models Antonio Sclocchi Alessandro Favero Noam Itzhak Levi Matthieu Wyart DiffM 121 5 0 17 Oct 2024