3D Foundation Models Workshop

Introduction

Textual and 2D visual foundation models, like GPT, BERT, DINO, and StableDiffusion, have revolutionized the fields of natural language processing and 2D computer vision with broad applications across academia and industry. However, the community still awaits the establishment of a 3D foundation model, one that could lead to diverse downstream applications in areas ranging from 3D content creation to AR/VR, robotics, and autonomous driving. This workshop aims to bring together researchers from various domains of 3D computer vision, facilitating discussions around the development of 3D foundation models. In particular, we investigate the questions including, but not limited to:

What types of datasets are necessary for 3D foundation models?
On which 3D tasks should such models be trained?
What are the consensus on the basic architecture and 3D representations for understanding and synthesis tasks?
What applications and capabilities can they unlock?

We invite a diverse group of top experts in the field to share their recent research outcomes and future visions in this regard, with a focus on scalable, generalizable, and adaptable 3D AI frameworks. We hope this workshop will inspire and accelerate the future innovations in 3D foundation models and their broad applications.

Workshop Schedule

Start Time (PDT)	End Time (PDT)	Event
8:50 AM	9:00 AM	Opening remarks
9:00 AM	9:30 AM	Aniruddha Kembhavi From bits to backdrops: Sourcing and Scoring 3D Foundation Models
9:30 AM	10:00 AM	Ruoshi Liu 3D Foundation Models for Physical Intelligence
10:00 AM	10:15 AM	Coffee break
10:15 AM	10:45 AM	Peter Wonka, Biao Zhang Towards Training a Large 3D Generative Model
10:45 AM	11:15 AM	Minghua Liu Understanding and Generating 3D Objects in an Open World
11:15 AM	11:45 AM	Andrea Vedaldi Towards a 3D foundation
11:45 AM	1:30 PM	Lunch break
1:30 PM	2:00 PM	Jun Gao 3D Representations and Algorithms for Leveraging Foundation Models
2:00 PM	2:30 PM	Angela Dai From Quantity to Quality for 3D Understanding
2:30 PM	3:00 PM	Xiaoguang Han How to Prepare Data for 3D Generative Foundation Model?
3:00 PM	3:15 PM	Coffee break
3:15 PM	3:45 PM	Hao Tan Large Reconstruction Models
3:45 PM	4:15 PM	Jérome Revaud From CroCo to DUSt3R: A Paradigm Change in 3D Vision?
4:15 PM	5:00 PM	~~Panel discussion~~ (Cancelled)