AI Paper Atlas

Research map / Vision-language and multimodal / Robot action and navigation

TDSR-VLA: Transition-aware Denoising Sequence Representations for Vision-Language-Action

Dong-Woo Kim, KEUNHO SONG, Seungmin Lee, Hwanhee Ju, Eun Cha, Daekyum Kim

ECCV 2026 · Vision-language and multimodal › Robot action and navigation · approach: Vision-language-action foundation models · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use USR instead of SR as keys in subgoal layers (ablation)TDSR-VLA · OpenArm real-worldaverage score: 0.70 -> 0.83Improved
use three subgoal injection layers instead of zero (ablation)TDSR-VLA · OpenArm real-worldaverage scoreImproved
increase subgoal injection layers from six to nine (ablation)TDSR-VLA · OpenArm real-worldaverage scoreHurt
three versus six subgoal layers (ablation)TDSR-VLA · OpenArm real-worldaverage scoreNo clear effect
planner-side transition cues plus subgoal image versus GR00T N1.5GR00T N1.5 · OpenArm real-worldaverage score: 0.74 -> 0.83Improved
TDSR-VLA versus GR00T N1.5 on object-cooperative tasksGR00T N1.5 · OpenArm object-cooperativeaverage score: 0.70 -> 0.88Improved
TDSR-VLA versus GR00T N1.5 on Dual Cup SortingGR00T N1.5 · OpenArm Dual Cup Sortingscore: 0.58 -> 0.55Hurt
TDSR-VLA versus OpenVLAOpenVLA · LIBERO Goalsuccess rate: 79.2 -> 82.0Improved
TDSR-VLA versus OpenVLAOpenVLA · LIBERO Longsuccess rate: 53.7 -> 56.0Improved
TDSR-VLA versus pi0 without pretrainingpi0 · LIBERO Goalsuccess rate: 89.0 -> 82.0Hurt
TDSR-VLA versus pi0 without pretrainingpi0 · LIBERO Longsuccess rate: 48.0 -> 56.0Improved
add planner subgoal generation before actingTDSR-VLA · Per-chunk latencylatencyHurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy