IntentVLA: Short-Horizon Intent Modeling for Aliased Robot Manipulation

Shijie Lian, Bin Yu, Xiaopeng Lin, Zhaolong Shen, Laurence Tianruo Yang, Yurun Jin, Haishan Liu, Changti Wu, Hang Yuan, Cong Huang, Kai Chen · · 2026

Summary

Abstract Summary

Robot imitation data are often multimodal: similar visual-language observations may be followed by different action chunks because human demonstrators act with different short-horizon intents, task phases, or recent context. Existing frame-conditioned VLA policies infer each chunk from the current observation and instruction alone, so under partial observability they may resample different intents across adjacent replanning steps, leading to inter-chunk conflict and unstable execution. We introduce IntentVLA, a history-conditioned VLA framework that encodes recent visual observations into a compact short-horizon intent representation and uses it to condition chunk generation. We further introduce AliasBench, a 12-task ambiguity-aware benchmark on RoboTwin2 with matched training data and evaluation environments that isolate short-horizon observation aliasing. Across AliasBench, SimplerEnv, LIBERO, and RoboCasa, IntentVLA improves rollout stability and outperforms strong VLA baselines

Key Points

Introduces a new dataset or benchmark
Provides open-source code or data

AliasBench: Ambiguity-Aware Benchmark Design

To evaluate whether a policy can resolve aliased observations from recent context, we build AliasBench on top of RoboTwin2 Chen et al. (2025) . AliasBench contains 12 manipulation tasks together with matched simulation training data and held-out evaluation environments.

The benchmark targets an underexplored gap in current VLA evaluation: most standard benchmarks measure whether a policy can complete a manipulation task, but they rarely isolate whether the policy can maintain a consistent decision when the current observation is aliased.

AliasBench: Ambiguity-Aware Benchmark Design — IntentVLA: Short-Horizon Intent Modeling for Aliased Robot Manipulation — Figure 4: Overview of IntentVLA. A Qwen3-VL backbone encodes the current image and language instruction, while a frozen VGGT-1B history encoder extracts recent visual evidence. IntentVLA fuses the history tokens with the current visual-language context through gated cross-attention, appends a compact short-horizon intent token, and conditions a DiT-based flow-matching action head for chunk generation.

Method

We model manipulation as a partially observable decision process with latent state $s_{t}\in\mathcal{S}$, observation $o_{t}\in\mathcal{O}$, and action $a_{t}\in\mathcal{A}$. This viewpoint is used only to motivate why recent observations can disambiguate the current frame.

At time step $t$, the robot observes $o_{t}$, receives a language instruction $\ell$, and predicts a future action chunk where $H$ is the chunk horizon and $d_{a}$ is the action dimension.

Experiment

We begin with AliasBench, which directly tests the failure mode identified in the introduction. On this benchmark, we compare against Qwen3-VL-GR00T and several history-as-extra-context baselines that feed multiple past frames directly into the Qwen backbone. We then evaluate on SimplerEnv Li et al. (2024c) , LIBERO Liu et al. (2023) , RoboCasa-GR1 Tabletop Tasks Nasiriany et al. (2024) ; Bjorck et al.

(2025) , and a real-world snack-cleanup task to test whether the same design transfers beyond the controlled ambiguity benchmark.

Sources and Demonstrations

Figures are reproduced from Lian et al. See the full paper for experimental details and the project page, when the paper links one, for demonstrations and videos.