RefCloth: Transferable Adversarial Tablecloths for Vision-Language-Action Models
Organizations: School of Artificial Intelligence, University of Chinese Academy of Sciences · Institute of Automation, Chinese Academy of Sciences · University of Chinese Academy of Sciences
Abstract
Closed-loop vision-language-action (VLA) policies rely continuously on multi-view cameras, yet existing visual attacks typically hijack sensor streams with 2D patches or alter target object meshes that require retraining when manipulation objects change. We introduce RefCloth, a framework that optimizes one persistent adversarial tablecloth on a frozen source policy and deploys it zero-shot across unseen tasks and target models without modifying task objects or cameras. Throughout manipulation, static third-person and moving wrist views observe this shared world-bound texture under evolving occlusions. To generalize across shifting surface views while bridging the gap between differentiable and native renderers, RefCloth combines bounded texture parameterization with shared surface UV transformations and a same-path clean reference that removes static rendering offsets at initialization. Across four LIBERO suites in native simulation, one frozen texture transfers from OpenVLA-OFT to and , reducing success relative to an unoptimized cloth by 62.50 and 43.75 percentage points on average and far exceeding amplitude- and variation-matched random controls. Dual-view audits and target closed-loop camera interventions show that moving wrist exposure drives over twice the standalone drop of the main view with strong synergy, while ablations confirm the roles of UV diversification and path calibration.
Figures & tables
| Method | Carrier | Obj.-agn. | 3D-Dual | Reuse | No-Tgt. | Ctrl. |
|---|---|---|---|---|---|---|
| UPA [ 32 ] , EDPA [ 53 ] | 2D patch | ✓ | – | ✓ | ✓ | – |
| Tex3D [ 7 ] | 3D mesh | – | – | – | ✓ | – |
| UniTexture [ 10 ] | 3D mesh | – | ✓ | ✓ | – | – |
| TRAP [ 19 ] | CoT cloth | ✓ | – | – | – | – |
| RefCloth (Ours) | 3D cloth | ✓ | ✓ | ✓ | ✓ | ✓ |
| Suite | Original | Plain | RefCloth | |
|---|---|---|---|---|
| Spatial | 94 | 89 | 12 | 77 |
| Object | 96 | 88 | 30 | 58 |
| Goal | 90 | 83 | 9 | 74 |
| Long | 79 | 43 | 2 | 41 |
| Mean | 89.75 | 75.75 | 13.25 | 62.50 |
| Texture | : | : |
|---|---|---|
| Random 101 | 57.75/18.00 | 87.75/5.00 |
| Random 102 | 53.50/22.25 | 85.00/7.75 |
| Random 103 | 56.75/19.00 | 86.50/6.25 |
| Random mean | 56.00/19.75 | 86.42/6.33 |
| RefCloth | 13.25/62.50 | 49.00/43.75 |
| Suite | Main | Wrist | Main | Wrist |
|---|---|---|---|---|
| Spatial | 23.5 | 41.3 | 23.7 | 39.7 |
| Object | 11.9 | 35.7 | 11.9 | 34.4 |
| Goal | 22.4 | 35.6 | 22.4 | 37.9 |
| Long | 17.9 | 33.4 | 18.7 | 22.8 |
| Condition | Succ. (%) | (pp) | Steps | vs. Plain |
|---|---|---|---|---|
| Plain | 98 | 0 | 32.4 | — |
| Main-only (M) | 88 | 10 | 36.8 | .00885 |
| Wrist-only (W) | 75 | 23 | 44.3 | |
| Both (M+W) | 34 | 64 | 66.1 |
| Method | S | O | G | L | Mean |
|---|---|---|---|---|---|
| Native ref. | 9/80 | 28/60 | 21/62 | 4/39 | 15.50/60.25 |
| EDPA | 23/66 | 14/74 | 18/65 | 2/41 | 14.25/61.50 |
| UPA-RFAS | 27/62 | 35/53 | 32/51 | 7/36 | 25.25/50.50 |
| RefCloth | 12/77 | 30/58 | 9/74 | 2/41 | 13.25/62.50 |
| Method | Surface | ||
|---|---|---|---|
| Original | None | 359 (89.75) | 376 (94.00) |
| Random | Patches | 290 (72.50) | 346 (86.50) |
| EDPA | Patches | 197 (49.25) | 273 (68.25) |
| UPA-RFAS | Patch | 253 (63.25) | 334 (83.50) |
| RefCloth | Cloth | 53 (13.25) | 196 (49.00) |
| Tasks | UV | Spatial | Object | Goal | Long |
|---|---|---|---|---|---|
| One | Off | 49/40 | 65/23 | 63/20 | 19/24 |
| One | On | 14/75 | 6/82 | 17/66 | 1/42 |
| Four | Off | 45/44 | 61/27 | 40/43 | 13/30 |
| Four | On | 12/77 | 30/58 | 9/74 | 2/41 |
Appendix figures & tables29 assets
Supplementary material from the paper’s appendix.
Appendix
| Method | Surface | Source | Training pool | Updates | Batch | Optimizer / step |
|---|---|---|---|---|---|---|
| RefCloth | Cloth | OpenVLA-OFT | 296 frames | 5,000 | 8 | Sign / 0.05 |
| Native reference | Cloth | OpenVLA-OFT | Same pool | 5,000 | 8 | Sign / 0.05 |
| EDPA | Cloth | OpenVLA-OFT | Same pool | 5,000 | 8 | Sign / 0.05 |
| UPA-RFAS | Cloth | OpenVLA-OFT | Same pool | 5,000 | 8 | Sign / 0.05 |
| EDPA | Two patches | OpenVLA-OFT | 303 no-cloth frames | 1,200 | 2 | Sign / |
| UPA-RFAS | One patch | OpenVLA | Spatial RLDS | 2,000 | 2 | AdamW / |
| Target | Suite | Original | Plain | Random cloth | Native ref. | EDPA | UPA-RFAS | RefCloth |
|---|---|---|---|---|---|---|---|---|
| Spatial | 94 | 89 | 62 | 9 | 23 | 27 | 12 | |
| Object | 96 | 88 | 73 | 28 | 14 | 35 | 30 | |
| Goal | 90 | 83 | 79 | 21 | 18 | 32 | 9 | |
| Long | 79 | 43 | 36 | 4 | 2 | 7 | 2 | |
| Spatial | 98 | 98 | 97 | 67 | 77 | 83 | 60 | |
| Object | 99 | 100 | 92 | 80 | 57 | 55 | 38 |
| Suite | Original | Random | EDPA | UPA-RFAS |
|---|---|---|---|---|
| Spatial | 94 | 87 | 59 | 74 |
| Object | 96 | 81 | 62 | 74 |
| Goal | 90 | 74 | 50 | 61 |
| Long | 79 | 48 | 26 | 44 |
| Total | 359 | 290 | 197 | 253 |
| Target | Suite | Pairs | ||
|---|---|---|---|---|
| Spatial | 11 | 18/7 | .216 | |
| Object | 5/21 | .020 | ||
| Goal | 9 | 13/4 | .216 | |
| Long | 0 | 2/2 | 1.000 | |
| Spatial | 17 | 26/9 | .038 | |
| Object | 19 | 31/12 | .038 |
| Target | Suite | Pairs | ||
|---|---|---|---|---|
| Spatial | 15 | 22/7 | .0407 | |
| Object | 5 | 19/14 | .974 | |
| Goal | 23 | 27/4 | .000272 | |
| Long | 5 | 6/1 | .375 | |
| Spatial | 23 | 30/7 | .00134 | |
| Object | 17 | 29/12 | .0460 |
| Method / Comparator | Configuration | Successes / 100 | (pp) | Pairs | |
|---|---|---|---|---|---|
| RefCloth (Ours) | Attenuated grid ( ) | 71 | 29 | — | — |
| Native reference | Batch 8, spatial TV | 92 | 8 | 21/0 | |
| Native reference | Batch 1, spatial TV | 84 | 16 | 14/1 | .00195 |
| Random cloth | Fixed, unoptimized | 92 | 8 | 23/2 | |
| EDPA | amplitude penalty | 77 | 23 | 15/9 | .307 |
| Target | Suite | Plain | RefCloth | Random 101 | Random 102 | Random 103 | Maximum |
|---|---|---|---|---|---|---|---|
| Spatial | 89 | 12 | 65 | 62 | 65 | ||
| Object | 88 | 30 | 67 | 64 | 68 | ||
| Goal | 83 | 9 | 73 | 65 | 71 | ||
| Long | 43 | 2 | 26 | 23 | 23 | ||
| Spatial | 98 | 60 | 99 | 99 | 98 | ||
| Object | 100 | 38 | 90 | 93 | 90 |
| Spatial tasks | Plain | Native ref. | RefCloth |
|---|---|---|---|
| All ten | 99/100 | 64/100 | 59/100 |
| Six held out | 59/60 | 35/60 | 28/60 |
| Seed | Native | Cloth-EDPA | RefCloth | ||
| 42 | 25.75 | 24.00 | 35.50 | ||
| 7 | 27.00 | 27.25 | 45.50 | ||
| 66 | 27.25 | 30.00 | 45.25 | ||
| 88 | 30.50 | 32.25 | 32.00 | ||
| Mean | 27.62 | 28.38 | 39.56 | ||
| SD | 2.03 | 3.56 | 6.86 | 8.03 | 8.11 |
| Seed | Scope | Native | M N | N M | Matched |
|---|---|---|---|---|---|
| 7 | Spatial | 31.0 | 44.0 | 62.0 | 38.0 |
| Object | 20.0 | 29.0 | 31.0 | 62.0 | |
| 8 | Spatial | 29.0 | 18.0 | 28.0 | 63.0 |
| Object | 38.0 | 35.0 | 39.0 | 40.0 | |
| 9 | Spatial | 45.0 | 62.0 | 44.0 | 38.0 |
| Object | 43.0 | 40.0 | 40.0 | 50.0 |
| Native | Matched | |||||
|---|---|---|---|---|---|---|
| Batch | Updates | Visits | S | O | S | O |
| 8 | 5,000 | 40,000 | 67/31 | 80/20 | 60/38 | 38/62 |
| 1 | 40,000 | 40,000 | 85/13 | 72/28 | 54/44 | 43/57 |
| Suite | Plain ( ) | Random 101 | RefCloth ( Tab. 2 ) | ( ) | X-VLA-Self | VLM-only |
|---|---|---|---|---|---|---|
| Spatial | 98.20 (491/500) | 97.40 (487/500) | 97.80 (489/500) | (.688) | 0.40 (2/500) | 16.00 (16/100) |
| Object | 99.60 (498/500) | 98.60 (493/500) | 89.80 (449/500) | ( ) | 12.40 (62/500) | 7.00 (7/100) |
| Goal | 97.80 (489/500) | 97.60 (488/500) | 94.80 (474/500) | (.0107) | 18.40 (92/500) | 13.00 (13/100) |
| Long | 95.00 (475/500) | 95.20 (476/500) | 83.20 (416/500) | ( ) | 51.80 (259/500) | 41.00 (41/100) |
| Mean | 97.65 (1953/2000) | 97.20 (1944/2000) | 91.40 (1828/2000) | 20.75 (415/2000) | 19.25 (77/400) |
| Condition | Succ. | Rate (%) | (pp) | |
|---|---|---|---|---|
| Original ( ) | 118 | 92.19 | — | — |
| Plain ( ) | 105 | 82.03 | 0.00 | — |
| Random 101 | 92 | 71.88 | 10.16 | .00443 |
| Random 102 | 93 | 72.66 | 9.38 | .00754 |
| RefCloth ( Tab. 2 ) | 86 | 67.19 | 14.84 | |
| RefCloth-RT | 84 | 65.62 | 16.41 |
| adjust_bottle ( ) | beat_block_hammer ( ) | place_empty_cup ( ) | |||||||
| Condition | Rate (%) | (pp) | Rate (%) | (pp) | Rate (%) | (pp) | |||
| Original ( ) | 85.16 | — | — | 75.00 | — | — | 53.00 | — | — |
| Plain ( ) | 53.12 | 0.00 | — | 57.00 | 0.00 | — | 19.00 | 0.00 | — |
| Random 101 | 46.09 | 7.03 | .136 | 42.00 | 15.00 | .0275 | 6.00 | 13.00 | |
| RefCloth ( Tab. 2 ) | 14.06 | 39.06 | 2.00 | 55.00 | 3.00 | 16.00 | |||
| RefCloth-RT | 31.25 | 21.88 | 26.00 | 31.00 | 3.00 | 16.00 | |||
| Target | Suite | Successes / 30 | Queries / episode | Main support (%) | Wrist (%) | |
|---|---|---|---|---|---|---|
| Spatial | 26 | 29.5 | 23.48 [22.86,24.48] | 41.27 | 1.000 | |
| Object | 26 | 36.9 | 11.86 [11.42,12.19] | 35.72 | 1.000 | |
| Goal | 22 | 39.4 | 22.41 [21.42,22.50] | 35.59 | 1.000 | |
| Long | 11 | 71.1 | 17.93 [15.03,23.57] | 33.38 | 1.000 | |
| Spatial | 30 | 22.2 | 23.72 [23.12,24.61] | 39.66 | 1.000 | |
| Object | 29 | 31.2 | 11.95 [11.55,12.15] | 34.36 | 1.000 |
| Target | View | Marker support | Plain–hidden | RefCloth–Plain | RMS |
|---|---|---|---|---|---|
| Main | 14.58 [14.14,14.91] | 11.75 [11.27,12.05] | 11.98 [11.51,12.17] | 10.70 | |
| Wrist | 39.04 [32.50,52.38] | 36.00 [30.49,48.07] | 36.01 [31.02,47.02] | 23.20 | |
| Main | 14.63 [14.29,14.88] | 11.85 [11.41,12.03] | 12.01 [11.67,12.17] | 10.69 | |
| Wrist | 37.79 [31.99,45.06] | 34.67 [29.81,41.13] | 34.47 [29.77,39.39] | 22.92 |
| Target | Plain | RefCloth | (pp) | Induced | Recovered | |
|---|---|---|---|---|---|---|
| 26/30 | 11/30 | 50.0 | 15 | 0 | ||
| 29/30 | 9/30 | 66.7 | 20 | 0 |
| Pairwise contrast | (pp) | Pairs | Exact | |
|---|---|---|---|---|
| Plain vs. Main | 11 / 1 | .00635 | .00885 | |
| Plain vs. Wrist | 24 / 1 | |||
| Plain vs. Both | 64 / 0 | |||
| Main vs. Wrist | 16 / 3 | .00443 | .00885 | |
| Main vs. Both | 55 / 1 | |||
| Wrist vs. Both | 41 / 0 |
| Texture | Main | Wrist |
|---|---|---|
| Plain | 0.678 | 0.376 |
| Random 101 | 0.775 | 0.583 |
| Random 102 | 0.785 | 0.693 |
| Random 103 | 0.775 | 0.557 |
| Native ref. | 1.467 | 0.692 |
| UV off | 0.884 | 0.516 |
| Step | (%) | ||||
|---|---|---|---|---|---|
| 0 | 0.00089 | 0.00097 | 0.005 | 53.7 | 19 / 32 |
| 1,875 | 0.04923 | 0.05757 | 0.054 | 57.2 | 23 / 32 |
| 3,750 | 0.06104 | 0.06564 | 0.030 | 52.7 | 18 / 32 |
| 5,000 | 0.06222 | 0.06337 | 0.046 | 50.4 | 16 / 32 |
| Suite | Task | Initial | Plain | Random | RefCloth | Plain | Random | RefCloth |
|---|---|---|---|---|---|---|---|---|
| Spatial | 0 | 10 | 85 S | 77 S | 400 F | 79 S | 75 S | 95 S |
| Spatial | 2 | 10 | 114 S | 110 S | 400 F | 92 S | 100 S | 348 S |
| Spatial | 0 | 12 | 75 S | 166 S | 400 F | 90 S | 82 S | 103 S |
| Goal | 5 | 10 | 125 S | 311 S | 400 F | 169 S | 126 S | 400 F |
| Goal | 7 | 10 | 81 S | 100 S | 400 F | 75 S | 78 S | 76 S |
| Direction | Rel. Error | |||
|---|---|---|---|---|
| Snapshot (32 states) | Matched | Native ref. | Matched | Native ref. |
| Archived ( Fig. 3 b) | 0.8612 | 0.9393 | 0.5029 | 0.3437 |
| Fresh (PNG) | 0.8644 | 0.9382 | 0.4900 | 0.3459 |
| Fresh (float) | 0.8643 | 0.9381 | 0.4899 | 0.3460 |
Adversarial Attacks on VLA Models
Adversarial attacks on vision-language-action (VLA) models study how crafted inputs, poisoned data, or physical cues can disrupt robot decisions and safety.
See all 34 papers