让冻结的脉冲模型在部署后仍能快速适应新状态，无需更新参数或标签。
SpikeWorld: Fast-State Adaptation for Frozen Spiking World Models
- 用外部路径动态修正状态预测，不更新主模型参数。
- 在未见数据上提升预测准确率最高达30.01%，动作追踪提升24.20%。
- 适合需要快速适应但不能修改模型的实时部署场景。
预测模型在行动后果被观测时接收自监督信号。当动态与语义共享参数时，冻结模型难以适应，而权重更新需优化器状态且可能破坏已学表征。本文提出SpikeWorld，一个145万参数的稀疏脉冲模型，联合训练于异构感知预测、语义理解、图像-文本关联及动作条件动态。部署时所有参数冻结。延迟的下一状态残差通过两条外部路径更新：累计固定池损失选择有界动作修正，路径特定残差矩阵优化下一状态预测。两条路径均不依赖标签、教师输出、奖励、成功信号或真实偏移值。联合优化使动作下一状态均方误差降低17.10%，同时提升多模态预测、语义准确率与图文检索效果。在未见剪切和衰减流上，综合外部状态使整体预测分别提升5.48%和30.01%；固定池动作路径分别提升追踪精度24.20%和3.94%。在包含450条新Meta-World轨迹（每臂75条）的六臂实验中，冻结策略奖励提升7.90（95%置信区间[2.48, 14.06]）；13.33分的成功差异具有描述性（置信区间[0, 40]）。相同感官输入下，模型参数与继承语义输出保持比特级不变。16字节的RLS估计器在直线衰减上获得最高非溯源奖励，表明贡献并非来自优越线性识别，而是其与冻结多模态脉冲检查点的集成。参考代码已公开于https://github.com/Oooorca/SpikeWorld。
原文摘要 · Abstract (English)
A predictive model receives a self-supervised signal whenever the consequence of an action is observed. Using that signal after deployment is difficult when dynamics and semantics share parameters: freezing prevents adaptation, whereas weight updates require optimizer state and may alter the learned representation. Here we introduce SpikeWorld, a 1.45M-parameter sparse spiking model jointly trained for heterogeneous sensory prediction, semantics, image-text binding and action-conditioned dynamics. At deployment, all trained parameters are frozen. Delayed next-state residuals update two external paths: cumulative fixed-bank losses select the bounded action correction, while route-specific residual matrices refine next-state prediction. Neither path uses labels, teacher outputs, rewards, success signals or the true shift value. Joint optimization improves action next-state MSE by 17.10\% while also improving multimodal prediction, semantic accuracy and image-text retrieval. On held-out shear and attenuation streams, the combined external state improves aggregate prediction by 5.48\% and 30.01\%; its fixed-bank action path improves tracking by 24.20\% and 3.94\%, respectively. In a six-arm study comprising 450 new Meta-World trajectories (75 per arm), SpikeWorld raises frozen-policy reward by 7.90 (95\% CI [2.48, 14.06]); the 13.33-point success difference is descriptive (CI [0, 40]). For identical sensory inputs, model parameters and inherited semantic outputs remain bitwise unchanged. A 16-byte RLS estimator obtains the highest non-oracle reward on linear attenuation, showing that the contribution is not superior linear identification, but its integration with a frozen multimodal spiking checkpoint. Reference code is publicly available at https://github.com/Oooorca/SpikeWorld.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。