COMPASS: Contrastive Multimodal Pretraining for Autonomous Systems
This addresses the problem of limited data and task-specific model design for autonomous systems, though it appears incremental as it builds on existing pretraining and multimodal approaches.
The authors tackled the challenge of learning generalizable representations for autonomous systems by introducing COMPASS, a contrastive multimodal pretraining pipeline that constructs a multimodal graph and maps signals into factorized spatio-temporal latent spaces, achieving generalization across drone navigation, vehicle racing, and visual odometry tasks in unseen environments and real-world data.
Learning representations that generalize across tasks and domains is challenging yet necessary for autonomous systems. Although task-driven approaches are appealing, designing models specific to each application can be difficult in the face of limited data, especially when dealing with highly variable multimodal input spaces arising from different tasks in different environments.We introduce the first general-purpose pretraining pipeline, COntrastive Multimodal Pretraining for AutonomouS Systems (COMPASS), to overcome the limitations of task-specific models and existing pretraining approaches. COMPASS constructs a multimodal graph by considering the essential information for autonomous systems and the properties of different modalities. Through this graph, multimodal signals are connected and mapped into two factorized spatio-temporal latent spaces: a "motion pattern space" and a "current state space." By learning from multimodal correspondences in each latent space, COMPASS creates state representations that models necessary information such as temporal dynamics, geometry, and semantics. We pretrain COMPASS on a large-scale multimodal simulation dataset TartanAir \cite{tartanair2020iros} and evaluate it on drone navigation, vehicle racing, and visual odometry tasks. The experiments indicate that COMPASS can tackle all three scenarios and can also generalize to unseen environments and real-world data.