基于多模态潜变量回放与增量调整的终身模仿学习
Lifelong Imitation Learning with Multimodal Latent Replay and Incremental Adjustment
在多模态潜空间而非原始像素上做经验回放，可大幅压缩视觉-语言-状态表征的存储开销，为视觉策略的持续学习提供可复用的记忆机制；角度间隔约束正则化任务嵌入演化的思路也可迁移到视觉表征的增量学习。
原文依据 operating entirely in a multimodal latent space, where compact representations of visual, linguistic, and robot's state information are stored and reused
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论我们提出一个终身模仿学习框架，能够在现实的内存与数据约束下，跨顺序任务实现持续的策略精炼。我们的方法不同于传统的经验回放，而是完全在多模态潜变量空间中运行，其中视觉、语言与机器人状态信息的紧凑表征被存储并复用于支持未来的学习。为进一步稳定适应过程，我们引入一种增量特征调整机制，通过角度间隔约束对任务嵌入的演化进行正则化，从而保持任务间的可区分性。我们的方法在 LIBERO 基准上取得了新的最优结果，与先前领先方法相比，AUC 提升 10-17 点，遗忘最多减少 65%。消融实验证实了各组成部分的有效性，显示出相对其他策略的一致增益。代码见：https://github.com/yfqi/lifelong_mlr_ifa。
英文原文摘要 980 字符
We introduce a lifelong imitation learning framework that enables continual policy refinement across sequential tasks under realistic memory and data constraints. Our approach departs from conventional experience replay by operating entirely in a multimodal latent space, where compact representations of visual, linguistic, and robot's state information are stored and reused to support future learning. To further stabilize adaptation, we introduce an incremental feature adjustment mechanism that regularizes the evolution of task embeddings through an angular margin constraint, preserving inter-task distinctiveness. Our method establishes a new state of the art in the LIBERO benchmarks, achieving 10-17 point gains in AUC and up to 65% less forgetting compared to previous leading methods. Ablation studies confirm the effectiveness of each component, showing consistent gains over alternative strategies. The code is available at: https://github.com/yfqi/lifelong_mlr_ifa.