Diffusion with forward models: Solving stochastic inverse problems without direct supervision

A Tewari, T Yin, G Cazenavette… - Advances in …, 2023 - proceedings.neurips.cc
Denoising diffusion models are a powerful type of generative models used to capture
complex distributions of real-world signals. However, their applicability is limited to …

Craftsman: High-fidelity mesh generation with 3d native generation and interactive geometry refiner

W Li, J Liu, R Chen, Y Liang, X Chen, P Tan… - arxiv preprint arxiv …, 2024 - arxiv.org
We present a novel generative 3D modeling system, coined CraftsMan, which can generate
high-fidelity 3D geometries with highly varied shapes, regular mesh topologies, and detailed …

DORSal: Diffusion for Object-centric Representations of Scenes et al

A Jabri, S van Steenkiste, E Hoogeboom… - arxiv preprint arxiv …, 2023 - arxiv.org
Recent progress in 3D scene understanding enables scalable learning of representations
across large datasets of diverse scenes. As a consequence, generalization to unseen …

Geometric Neural Process Fields

W Yin, Z **ao, J Shen, Y Chen, CGM Snoek… - arxiv preprint arxiv …, 2025 - arxiv.org
This paper addresses the challenge of Neural Field (NeF) generalization, where models
must efficiently adapt to new signals given only a few observations. To tackle this, we …

DORSal: Diffusion for Object-centric Representations of Scenes

A Jabri, S van Steenkiste, E Hoogeboom… - The Twelfth International … - openreview.net
Recent progress in 3D scene understanding enables scalable learning of representations
across large datasets of diverse scenes. As a consequence, generalization to unseen …