Industrial ReMEmbR: Scaling Spatiotemporal Memory to Robot Fleets
arXiv cs.AI · 21 Oct 2026, 03:58 AEST
IndustrialReMEmbR is a shared spatiotemporal memory system designed for warehouse robot fleets that reduces computation and storage costs. It generates scene descriptions on demand, allowing robots to encode video clips without uploading raw camera frames.
Key facts
- Warehouse robot fleets continuously observe operational conditions, but preserving these observations as reusable semantic memory incurs substantial computation, communication, and storage costs.
- We present IndustrialRemembr, a shared spatiotemporal memory system that extends NVIDIA ReMEmbR to robot fleets and reduces server-side computation by generating scene descriptions only on demand.
- On each robot, an NVIDIA Jetson encodes video clips into retrieval embeddings with Cosmos-Embed1, vision tokens with Cosmos3-Edge, and reconstruction latents with the Wan2.2 VAE, without uploading raw camera frames.
- The encoding models are re-used to detect people in image patches, zeroing them out to support PII abstraction for GDPR-compliant deployments.
Read aloud
IndustrialReMEmbR is a shared spatiotemporal memory system designed for warehouse robot fleets that reduces computation and storage costs. It generates scene descriptions on demand, allowing robots to encode video clips without uploading raw camera frames. Warehouse robot fleets continuously observe operational conditions, but preserving these observations as reusable semantic memory incurs substantial computation, communication, and storage costs. We present IndustrialRemembr, a shared spatiotemporal memory system that extends NVIDIA ReMEmbR to robot fleets and reduces server-side computation by generating scene descriptions only on demand. On each robot, an NVIDIA Jetson encodes video clips into retrieval embeddings with Cosmos-Embed1, vision tokens with Cosmos3-Edge, and reconstruction latents with the Wan2.two VAE, without uploading raw camera frames. The encoding models are re-used to detect people in image patches, zeroing them out to support PII abstraction for GDPR-compliant deployments.
Sources in this cluster
- Industrial ReMEmbR: Scaling Spatiotemporal Memory to Robot Fleets
arXiv cs.AI · 21 Oct 2026, 03:58 AEST