Dynam3D: Dynamic Layered 3D Tokens Empower VLM for Vision-and-Language Navigation

NeurIPSOral2025

Authors: Zihan Wang, Seungjun Lee, Gim Hee Lee
Affiliation: National University of Singapore
Venue: NeurIPS 2025
Track: Oral

TL;DR

A multi-level patch-instance-zone 3D representation model with a hierarchical dynamic online update strategy for embodied navigation.

Opening excerpt from the authors’ abstract. source

Read the paper

Topics

3d

← All NeurIPS 2025 Oral papers · Browse the whole archive