Utonia: Toward One Encoder for All Point Clouds

2026-03-03

AI Summary

This work addresses the challenge of constructing a unified self-supervised encoder capable of processing heterogeneous point cloud data from diverse sources—including remote sensing, outdoor LiDAR, indoor RGB-D scans, CAD models, and monocular videos. The authors propose Utonia, the first cross-domain self-supervised point cloud Transformer encoder, which learns a unified and transferable 3D representation space through joint training on multi-source data. Its key innovations lie in cross-domain feature alignment and sparse 3D modeling, enabling, for the first time, unified self-supervised representation learning across multiple domains. Utonia demonstrates emergent capabilities arising from cross-domain co-training and achieves significant performance gains in downstream tasks spanning 3D perception, robotic manipulation, and spatial reasoning within vision-language models.

Technology Category

• Machine Learning: Unsupervised & Self-Supervised Learning
• Computer Vision: Multi-modal Vision
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Systems and Infrastructure for Web, Mobile and WoT: Web applications in cross-disciplinary domains and verticals such as mixed reality, smart cities, and digital health

Abstract

We dream of a future where point clouds from all domains can come together to shape a single model that benefits them all. Toward this goal, we present Utonia, a first step toward training a single self-supervised point transformer encoder across diverse domains, spanning remote sensing, outdoor LiDAR, indoor RGB-D sequences, object-centric CAD models, and point clouds lifted from RGB-only videos. Despite their distinct sensing geometries, densities, and priors, Utonia learns a consistent representation space that transfers across domains. This unification improves perception capability while revealing intriguing emergent behaviors that arise only when domains are trained jointly. Beyond perception, we observe that Utonia representations can also benefit embodied and multimodal reasoning: conditioning vision-language-action policies on Utonia features improves robotic manipulation, and integrating them into vision-language models yields gains on spatial reasoning. We hope Utonia can serve as a step toward foundation models for sparse 3D data, and support downstream applications in AR/VR, robotics, and autonomous driving.

Problem

• point cloud
• unified encoder
• cross-domain representation
• 3D foundation model
• self-supervised learning

Innovation

• Utonia
• self-supervised learning
• point cloud transformer
• cross-domain representation
• foundation model