Lune

CVPR2026Top-tier venue

GeCo-SRT: Geometry-aware Continual Adaptation for Cross-Task Sim-to-Real Transfer

Wenbo Yu, Wenke Xia, Weitao Zhang, Di Hu

Abstract

Bridging the sim-to-real gap is important for applying low-cost simulation data to real-world robotic systems. However, previous methods are severely limited by treating each transfer as an isolated endeavor, demanding repeated, costly tuning and wasting prior transfer experience. To move beyond isolated sim-to-real, we build a continual cross-task sim-to-real transfer paradigm centered on knowledge accumulation across iterative transfers, thereby enabling effective and efficient adaptation to novel tasks. Thus, we propose GeCo-SRT, a geometryaware continual adaptation method. It utilizes domaininvariant and task-invariant knowledge from local geometric features as a transferable foundation to accelerate adaptation during subsequent sim-to-real transfers. This method starts with a geometry-aware mixture-of-experts module, which dynamically activates experts to specialize in distinct geometric knowledge to bridge observation sim-to-real gap. Further, the geometry-expert-guided prioritized experience replay module preferentially samples from underutilized experts, refreshing specialized knowledge to combat forgetting and maintain robust cross-task performance. Leveraging knowledge accumulated during iterative transfer, the GeCo-SRT method not only achieves 52% average performance improvement over the baseline, but also demonstrates significant data efficiency for new task adaptation with only 1/6 data. We hope this work inspires approaches for efficient, low-cost cross-task sim-to-real transfer. Our code is released at this https URL

Ask Lune

Know this paper in minutes, not an afternoon.

Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.

Questions to start from

Lune can draw on

the full paper · related work

Get the answer

Free to start. No credit card required.

Lineage

Builds onCited byLeft to right by year; the more cited, the higher.

Builds on9

  1. Always Be Dreaming: A New Approach for Data-Free Class-Incremental LearningJames Seale Smith, Yen-Chang Hsu, Jonathan C. Balloch, Yilin Shen et al.ICCV 2021 · 208 citations
  2. Understanding Domain Randomization for Sim-to-real TransferXiaoyu Chen, Jiachen Hu, Chi Jin, Lihong Li et al.ICLR 2022 · 164 citations
  3. Domain-Agnostic Prior for Transfer Semantic SegmentationXinyue Huo, Lingxi Xie, Hengtong Hu, Wengang Zhou et al.CVPR 2022 · 39 citations
  4. ASID: Active Exploration for System Identification in Robotic ManipulationMarius Memmel, Andrew Wagenmaker, Chuning Zhu, Dieter Fox et al.ICLR 2024 · 36 citations
  5. Human-assisted Robotic Policy Refinement via Action Preference OptimizationWenke Xia, Yichu Yang, Hongtao Wu, Xiao Ma et al.NeurIPS 2025 · 17 citations

Related papers

  1. Continual Knowledge Adaptation for Reinforcement LearningJinwu Hu, Zihao Lian, Zhiquan Wen, Chenghao Li et al.NeurIPS 2025 · 8 citations
  2. REvolveR: Continuous Evolutionary Models for Robot-to-robot Policy TransferXingyu Liu, Deepak Pathak, Kris KitaniICML 2022 · 24 citations
  3. Continual Sequence Generation with Adaptive Compositional ModulesYanzhe Zhang, Xuezhi Wang, Diyi YangACL 2022 · 53 citations
  4. SRSA: Skill Retrieval and Adaptation for Robotic Assembly TasksYijie Guo, Bingjie Tang, Iretiayo Akinola, Dieter Fox et al.ICLR 2025
  5. Graph Cross-Domain Continual Fine-Tuning via Orthogonal LoRA Routing with Contrastive Expert SpecializationQianyi Cai, Ziyue Qiao, Minghao Yang, Xiao Luo et al.WWW 2026
Dusk over the sea between two cliffs drawn in fine vertical lines