NestedVLA：用嵌套超网络为 VLA 机器人模型按任务生成技能参数
NestedVLA: Learning to Consolidate and Generate Skills for Vision-Language-Action Models
一篇 VLA 机器人论文，思路是超网络按任务生成参数而不做参数合并，三个基准上最强基线平均高出 41 个百分点，代码也开源了。
NestedVLA 提出"先独立学技能、事后整合"的 VLA 训练范式：嵌套超网络根据目标任务生成自适应动作专家参数，与冻结的基础 VLA 组合使用，同时单独保留压缩的 VLM 残差。在 MetaWorld、ManiSkill、CALVIN 三个 RL 基准上，该方法平均比最强模型合并基线高 41 个百分点。在 ManiSkill 上它接近逐任务微调的性能差距（1 个百分点以内），并大幅缩小了在 CALVIN 上的差距。代码已在 GitHub 开源（WanyuGroup/NeurIPS2026-NestedVLA）。
NestedVLA: Learning to Consolidate and Generate Skills for Vision-Language-Action Models
Vision--Language--Action (VLA) models promise generalist robotic agents, yet consolidating skills learned across heterogeneous tasks and environments remains difficult: training data is fragmented across platforms, joint multi-task training suffers from interference, and model merging tends to introduce parameter conflicts or require architectural changes. We instead advocate a paradigm in which skills are first learned independently and then consolidated post hoc. To this end, we propose NestedVLA, which replaces parameter aggregation with parameter generation: a nested hypernetwork, conditioned on the target task, synthesizes task-adaptive action-expert parameters that compose with a frozen base VLA. Task-specific compressed VLM residuals are retained separately. This formulation enables flexible task-conditioned skill reuse while mitigating interference. Across three RL benchmarks (MetaWorld, ManiSkill, CALVIN), NestedVLA outperforms the strongest model-merging baseline by 41 pp on average, matches per-task fine-tuning within 1 pp on ManiSkill, and substantially closes the gap on CALVIN, pointing toward a storage-efficient route to skill consolidation. Code is available at https://github.com/WanyuGroup/NeurIPS2026-NestedVLA.
