RSS 2026 Robotics Paper Awards Announced
量子位 · wechat · 2026-07-16
RSS 2026 paper awards have been announced: out of 708 submissions and 210 accepted papers, three Best Paper awards were given to FlashSAC, Muninn, and NeuralActuator.
- FlashSAC: An off-policy reinforcement learning method for high-dimensional robot control. It improves training stability by reducing gradient updates, scaling up model and experience pool sizes, and constraining norms. In various simulators and Sim2Real experiments with humanoid walking, it compresses training time from hours to minutes.
- Muninn: Adds a caching mechanism to trajectory diffusion models, reusing intermediate results without retraining. It achieves up to 4.6x speedup while preserving performance and safety metrics.
- NeuralActuator: Uses Transformer + differentiable simulation for actuator modeling to predict torque, external contact forces, and motor states. This helps low-cost robotic arms sense forces without dedicated force/torque sensors. It also includes the release of the NAD dataset.
Additionally, the FinalList features works on bionic robot facial mechanisms (related to Tsinghua AIR/BAAI/Galaxy Universal), action tokenizer OAT, extrinsic dexterous manipulation in cluttered scenes DAPL, GPU-accelerated robust trajectory optimization cuNRTO, and differentiable fluid-robot multiphysics simulation for robotic eels. The conference also announced the Test of Time Award for the EuRoC dataset and highly compliant underactuated robot hands, along with presentation topics from four EarlyCareerSpotlight scholars.
More from Embodied
- ECCV26 Oral: Flow Matching Enables Single-Stage Multi-View Point Cloud Registration — ducha_aiki · 2026-09-11
- Polish developers build iPhone app that detects nearby Meta smart glasses — Low-Honeydew6483 · 2026-09-11
- Ant's Afu health AI hits 150M users, unveils AI+hardware health alliance at Bund Summit — APPSO · 2026-09-11
- Johns Hopkins Launches Full-Stack Hands-on Robot Learning Class with SO-101 Arm Kits — _krishna_murthy · 2026-09-11
- SyncWorld: In-Context Robot World Model Simulates Unseen Views and Embodiments Zero-Shot — ChongZzZhang · 2026-09-11
- A 3D Pose Dataset for Dogs Released — ducha_aiki · 2026-09-11