The paper introduces M2Tok, a Multi-head Multi-codebook Action Tokenizer that reduces reconstruction loss for continuous action signals by decomposing latent features into multiple heads and assigning independent codebooks to each. This design expands representational expressivity, leading to lower reconstruction error and higher success rates in Vision‑Language‑Action models evaluated on RoboTwin, Simpler‑Env, and zero‑shot real‑world tasks. Ablation studies confirm the effectiveness of both multi‑head and multi‑codebook mechanisms.
By Chunpu Xu, Zhixuan Liang, Yuhao Zhang, Chi-Min Chan, Jessie Wang, Yang Xiao, Mengkang Hu, Xiaokang Yang, Yao Mu
arXiv:2606. 30113v1 Announce Type: cross Abstract: Discrete action tokenization provides a compact interface for autoregressive VLA policies, but accurately recovering continuous robot actions from discrete codes remains challenging.
By Tengyue Jiang, Chunpu Xu, Jiayue Kang, Yao Mu
The paper "ActionCodec: What Makes for Good Action Tokenizers" addresses the lack of design guidance for action tokenization in Vision‑Language‑Action (VLA) models. By deriving information‑theoretic principles—maximizing temporal token overlap, reducing vocabulary redundancy, boosting multimodal mutual information, and ensuring token independence—the authors propose ActionCodec, a tokenizer that improves training efficiency and VLA performance. Experiments show that a SmolVLM2‑2.2B fine‑tuned with ActionCodec reaches a 95.5 % success rate on LIBERO, rising to 97.4 % with architectural tweaks, setting a new state‑of‑the‑art for VLA models without robotics pre‑training.
By Zibin Dong, Yicheng Liu, Shiduo Zhang, Baijun Ye, Yifu Yuan, Fei Ni, Jingjing Gong, Xipeng Qiu, Hang Zhao, Yinchuan Li, Jianye Hao
ActionPiece rethinks how actions are tokenized for autoregressive vision‑language‑action models by introducing physical rank consistency (PRC) to evaluate relational fidelity of reconstructed actions. The method jointly supervises representation learning and quantization to preserve local physical distance rankings, improving both PRC and policy success. Experiments on LIBERO, LIBERO‑Plus, SimplerEnv, and VLA‑Arena show significant gains over baseline tokenizers.
By Shijie Lian, Bin Yu, Zhaolong Shen, Xiaopeng Lin, Yichao Du, Zhirui Zhang, Laurence T. Yang, Kai Chen
arXiv:2608. 10484v1 Announce Type: cross Abstract: Action verbs describe not only the physical outcomes of actions, but also how those actions are performed.
By Li Wenjie, Yash Jangir, Ignacy Stepka, Yash Agarwal, Marion Kipsang, Yonatan Bisk
arXiv:2606. 14752v1 Announce Type: cross Abstract: Modern Vision-Language-Action (VLA) models must bridge pretrained vision-language reasoning and precise continuous robot control.
By Xirui Kang, Yanpei Shi, Lucy Liang, Roy Gan, Dongxiu Liu, Pushi Zhang, Danpeng Chen, Xiaoyi Qin, Yinan Zheng, Jinliang Zheng, Hao Wang, Xianyuan Zhan, Hang Su
The paper investigates how different action tokenization methods affect closed‑loop control in autoregressive vision‑language‑action models. It compares analytical, linear, and nonlinear representations, showing that lower reconstruction error does not guarantee better policy performance. The study highlights the need to evaluate tokenization on multiple criteria, including sequence predictability and decoder stability, rather than relying solely on reconstruction fidelity.
By Yuxin Yang, Gaohan He, Changxue Guan, Hangming Liu
MambaAct introduces a compact Mamba-based action expert to replace the transformer-based expert in flow‑matching vision‑language‑action models, preserving the pretrained visual‑language prefix encoder while reducing the action expert size by tenfold. The new design employs a lightweight Context Compressor to embed vision‑language conditioning into each Mamba block, eliminating repeated token‑level cross‑attention. Experiments show a 64.8 % reduction in denoising latency on LIBERO and 58.8 % in real‑world deployment, with a modest 1.08 % increase in mean success rate across multiple fine‑tuning seeds.
By Khuyen Pham, Tran Nguyen Le, Edoardo Fazzari, Cesare Stefanini
The paper introduces Direction-Scale Decomposition (DSD), an action representation that separates translation and rotation increments into direction and scale components before tokenization. DSD is evaluated with uniform binning and a B-spline tokenizer (BEAST) in both simulation and real-world manipulation tasks, showing improved success rates on LIBERO and SimplerEnv, especially under mixed-dataset training. Real-robot experiments confirm performance gains with and without robotics pretraining, supporting DSD as an effective representation for discrete-token vision-language-action models.
By Yufei Duan, Hang Yin, Alberta Longhini, Chao Tang, Danica Kragic
IMLE‑VLA replaces the iterative action head in vision‑language‑action policies with a single‑step conditional generator trained via conditional Implicit Maximum Likelihood Estimation (cIMLE). This eliminates multi‑step sampling, boosting inference frequency by 3.67× (55 Hz vs. 15 Hz) and achieving the highest average success rate (98.0 %) on the 40‑task LIBERO benchmark while maintaining robustness under perturbations. Real‑world tests on a Franka Emika Panda show smoother, faster motions and a 3.9×–6.6× reduction in inference time per episode.
By Kian Hosseinkhani (Simon Fraser University), Qinhe Peng (University of Pennsylvania), George Shramko (Simon Fraser University), Mehran Aghabozorgi (Simon Fraser University), Jianing Qian (University of Pennsylvania), Tristan Engst (Simon Fraser University), Alireza Moazeni (Simon Fraser University), Dinesh Jayaraman (University of Pennsylvania), Ke Li (Simon Fraser University, Canada CIFAR AI Chair)
arXiv:2607. 21670v1 Announce Type: cross Abstract: Action tokenization maps continuous robot action chunks to discrete tokens and has become an important interface for modern visuomotor policies.
By Chaoqi Liu, Yue Zhao, Haonan Chen, Xiaoshen Han, Jiawei Gao, Ehsan Adeli, Yilun Du
Vision-language-action (VLA) models commonly adopt an LLM-centric $V \to L \to A$ pathway, where visual observations are projected into the representation space of a large language model before being decoded into robot actions. Although effective, this design incurs substantial computation and memory overhead at every policy invocation.