Metraipolis › archive › dc_ed82bdbf071d

Beyond Reconstruction Error: Analytical and Data-Driven Action Tokenization for Autoregressive Vision-Language-Action Models — arXiv:2609.25820

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.25820
License (as archived)
CC0-1.0
Archived
2026-10-08T15:08:36.299Z
SHA-256 of the archived text
aeaedc3572e97b5cfc3837ead9d4987be65d27e0c521080acc5b6e8ca40f24f3
Stable URL of this page
https://metraipolis.com/documento/dc_ed82bdbf071d

How to cite

Beyond Reconstruction Error: Analytical and Data-Driven Action Tokenization for Autoregressive Vision-Language-Action Models — arXiv:2609.25820 — arXiv (en), https://arxiv.org/abs/2609.25820 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_ed82bdbf071d (Metraipolis archive dc_ed82bdbf071d, sha256 aeaedc3572e97b5cfc3837ead9d4987be65d27e0c521080acc5b6e8ca40f24f3)
BibTeX
@misc{dced82bdbf071d,
  title = {Beyond Reconstruction Error: Analytical and Data-Driven Action Tokenization for Autoregressive Vision-Language-Action Models — arXiv:2609.25820},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.25820},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_ed82bdbf071d; sha256 aeaedc3572e97b5cfc3837ead9d4987be65d27e0c521080acc5b6e8ca40f24f3},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.25820. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Discrete action tokenization is central to autoregressive vision-language-action (VLA) models, yet action representations are often evaluated primarily through reconstruction fidelity. We ask which representation properties actually matter for closed-loop control by comparing fixed analytical, data-driven linear, and nonlinear neural representations under a unified tokenization interface. Across rate-distortion analysis, sequence-modeling diagnostics, and 3,500 LIBERO rollouts, representation rankings change with the evaluation criterion. PCA achieves lower nominal reconstruction error than Temporal-DCT, but produces less predictable token sequences and 3.0 percentage points lower mean seen-task success across three policy-training seeds, with the policy ordering reversing in one seed. In a matched seed-42 ablation, an autoencoder further reduces reconstruction error yet does not yield the strongest policy and exhibits greater sensitivity to discrete token perturbations. These findings show that reconstruction fidelity alone cannot reliably select action representations for autoregressive control, motivating joint evaluation of geometric fidelity, sequence predictability, decoder stability, and closed-loop performance.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-23 in cs.LG).
Categorie arXiv: Robotics (cs.RO) ; Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2609.25820 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-08T14:37:29.550Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-09-23?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).