2026-07-07 04:00 UTCOriginal source2 min readUpdated: 2026-07-07 08:30 UTC

TACO: TActile World Model as a Self-COrrector for Scalable VLA Post-Training

TACO is a tactile-aware world-model-driven framework for scalable VLA post-training in contact-rich manipulation. It converts real failures into imagined visuo-tactile corrections via a Recognize-Imagine-Label loop, achieving a 44% absolute success rate improvement over the base policy.

SourcearXiv RoboticsAuthor: Shengbang Liu, Yueru Jia, Yuyang Yan, Jiaming Liu, Xinran Zhang, Qiuxuan Feng, Yandong Guo, Shiji Zhou, Boxin Shi, Shanghang Zhang

Article intelligence

EngineersAdvanced

Key points

VLA models struggle with contact-rich tasks due to vision-only limitations; tactile-aware corrective post-training efficiently improves recovery.
TACO uses a unified progress-action model and a visuo-tactile generation model to recognize failures and imagine corrections.
Knowledge-insulated tactile adaptation with advantage-conditioned training enables learning from imagined corrections without degrading visual-language priors.
On real-world contact-rich manipulation tasks, TACO improves absolute success rate by 44%, outperforming the variant without tactile adaptation by 32%.

Why it matters

This matters because VLA models struggle with contact-rich tasks due to vision-only limitations; tactile-aware corrective post-training efficiently improves recovery.

Technical impact

May affect model selection, inference cost, product capability, and evaluation benchmarks.

This panel is AI-generated and reviewed for accuracy.

-->

[Submitted on 3 Jul 2026]

Title:TACO: TActile World Model as a Self-COrrector forScalable VLA Post-Training

View a PDF of the paper titled TACO: TActile World Model as a Self-COrrector forScalable VLA Post-Training, by Shengbang Liu and 9 other authors

View PDF HTML (experimental)

Abstract:Vision-Language-Action (VLA) models have shown promising generalization in robotic manipulation, but they still struggle with contact-rich tasks, where minor contact perturbations can cause unrecoverable failures that are hard to detect from vision alone. Since these failures are localized rather than task-level semantic errors, tactile-aware corrective post-training offers an efficient way to improve recovery. However, scaling such supervision through human intervention is costly. Recent works have explored world models to synthesize imagined rollouts for policy improvement, but vision-only world models may produce visually plausible yet contact-inconsistent trajectories. We therefore introduce TACO, a tactile-aware world-model-driven framework for scalable VLA post-training in contact-rich manipulation. Given real robot rollouts, TACO follows a Recognize-Imagine-Label loop with a tactile-aware world model: a unified progress-action model recognizes failure-adjacent states using progress estimates, a visuo-tactile generation model imagines local correction segments, and the progress-action model labels them with executable corrective actions. To incorporate tactile corrective supervision into VLA post-training, TACO combines knowledge-insulated tactile adaptation with advantage-conditioned training, enabling the policy to learn from imagined corrections without degrading pretrained visual-language priors. These components enable TACO to convert real-world failures into imagined visuo-tactile corrections for iterative VLA post-training. Experiments on real-world contact-rich manipulation tasks show that TACO achieves 44% absolute success rate improvement over the base policy and 32% over the policy without knowledge-insulated tactile adaptation.

Subjects:

Robotics (cs.RO)

Cite as: arXiv:2607.02840 [cs.RO]

(or arXiv:2607.02840v1 [cs.RO] for this version)

https://doi.org/10.48550/arXiv.2607.02840

arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Shengbang Liu [view email] [v1] Fri, 3 Jul 2026 00:23:30 UTC (30,144 KB)

Full-text links:

Access Paper:

View a PDF of the paper titled TACO: TActile World Model as a Self-COrrector forScalable VLA Post-Training, by Shengbang Liu and 9 other authors

View PDF

HTML (experimental)

TeX Source

view license

Current browse context:

cs.RO

new | recent | 2026-07

Change to browse by:

References & Citations

NASA ADS

Google Scholar

Semantic Scholar

Data provided by:

Bibliographic Tools

Bibliographic and Citation Tools

Bibliographic Explorer Toggle

Bibliographic Explorer (What is the Explorer?)

Connected Papers Toggle

Connected Papers (What is Connected Papers?)

Litmaps Toggle

Litmaps (What is Litmaps?)

scite.ai Toggle

scite Smart Citations (What are Smart Citations?)

Code, Data, Media

Code, Data and Media Associated with this Article

alphaXiv Toggle

alphaXiv (What is alphaXiv?)

Links to Code Toggle

CatalyzeX Code Finder for Papers (What is CatalyzeX?)

DagsHub Toggle

DagsHub (What is DagsHub?)

GotitPub Toggle

Gotit.pub (What is GotitPub?)

Huggingface Toggle

Hugging Face (What is Huggingface?)

ScienceCast Toggle

ScienceCast (What is ScienceCast?)

Demos

Replicate Toggle

Replicate (What is Replicate?)

Spaces Toggle

Hugging Face Spaces (What is Spaces?)

Spaces Toggle

TXYZ.AI (What is TXYZ.AI?)