改进动作离散化方法，让机器人模仿学习更稳定可靠。
Action Tokenizer Matters in In-Context Imitation Learning
- 提出LipVQ-VAE模型，通过权重归一化保证动作空间平滑性。
- 在高保真仿真中性能提升超5.3%，实机验证轨迹更顺滑。
- 适合关注机器人模仿学习与动作表示的开发者和研究者。
上下文模仿学习（ICIL）是一种新范式，使机器人无需重新训练即可从示范中泛化到未见任务。良好的动作表示是有效捕捉示范信息的关键，但动作分词器（将动作离散化并编码的过程）在ICIL中仍缺乏研究。本文首次系统评估现有动作分词方法，发现其虽能有效编码动作轨迹，却未能保持时间上的平滑性，而这是机器人稳定执行的关键。为此，我们提出LipVQ-VAE，一种通过权重归一化在潜在动作空间施加Lipschitz条件的变分自编码器。通过将原始动作输入的平滑性约束传播至量化后的潜在码本，LipVQ-VAE生成更稳定、更平滑的动作。集成至ICIL后，在高保真模拟器中性能提升超过5.3%，真实世界实验也证实其能产生更平稳、更可靠的轨迹。代码与检查点见https://action-tokenizer-matters.github.io/
原文摘要 · Abstract (English)
In-context imitation learning (ICIL) is a new paradigm that enables robots to generalize from demonstrations to unseen tasks without retraining. A well-structured action representation is the key to capturing demonstration information effectively, yet action tokenizer (the process of discretizing and encoding actions) remains largely unexplored in ICIL. In this work, we first systematically evaluate existing action tokenizer methods in ICIL and reveal a critical limitation: while they effectively encode action trajectories, they fail to preserve temporal smoothness, which is crucial for stable robotic execution. To address this, we propose LipVQ-VAE, a variational autoencoder that enforces the Lipschitz condition in the latent action space via weight normalization. By propagating smoothness constraints from raw action inputs to a quantized latent codebook, LipVQ-VAE generates more stable and smoother actions. When integrating into ICIL, LipVQ-VAE improves performance by more than 5.3% in high-fidelity simulators, with real-world experiments confirming its ability to produce smoother, more reliable trajectories. Code and checkpoints are available at https://action-tokenizer-matters.github.io/
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。