Graph-based representations for Spatial-AI | Andrew Davison | Tartan SLAM Series

A presentation by Andrew Davison as part of the Tartan SLAM Series.
Series overviews and links can be found on our webpage: https://theairlab.org/tartanslamseries/

Abstract: To enable the next generation of smart robots and devices which can truly interact with their environments, Simultaneous Localisation and Mapping (SLAM) will progressively develop into a general real-time geometric and semantic `Spatial AI' perception capability. Andrew will give many examples from their work on gradually increasing visual SLAM capability over the years. However, much research must still be done to achieve true Spatial AI performance. A key issue is how estimation and machine learning components can be used and trained together as we continue to search for the best long-term scene representations to enable intelligent interaction. Further, to enable the performance and efficiency required by real products, computer vision algorithms must be developed together with the sensors and processors which form full systems, and Andrew will cover research on vision algorithms for non-standard visual sensors and graph-based computing architectures.

Outline:
0:00 - Intro
2:00 - Visual SLAM-Enabled Products and Systems
3:49 - SLAM to Spatial AI & Potential Products
6:57 - Current Gap for Spatial AI systems
7:30 - FutureMapping
8:24 - Rearrangement: A Challenge for Embodied AI
9:39 - Overview of MonoSLAM, ElasticFusion, SemanticFusion
14:21 - Semantic SLAM Computation Graph
16:47 - SLAM meets Deep Learning
18:38 - New Representations for Spatial AI
23:04 - iMAP
25:41 - Object-based Representations
32:55 - Hardware for Spatial AI
38:19 - Finding the Graphs in Spatial AI
40:47 - Gaussian Belief Propagation for Spatial AI
45:10 - Conclusion
46:17 - Q&A

AirLab Links:
Website: https://theairlab.org
Twitter:   / airlabcmu  
LinkedIn:   / the-air-lab-at-carnegie-mellon-university  
Facebook:   / airlabcmu  
Medium:   / airlabcmu  

RPL Links:
Website: https://rpl.ri.cmu.edu/
Twitter:   / rpl_cmu