Beyond Static Perception: Integrating Temporal Context into VLMs for Cloth Folding

Adri\`a Colom\'e; Carme Torras; Oriol Barbany

REVIEW 4 major objections 5 minor 31 references

Beyond Static Perception: Integrating Temporal Context into VLMs for Cloth Folding

T0 review · 4 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Incorporating a keyframe history into a vision-language model improves its ability to recognize complex cloth states and predict the next folding action.

desk verdict Honest, well-scoped ablation of the authors' own BiFold model showing keyframe context beats consecutive/no context on offline pixel-space AP, but the physical cloth-folding claim is not yet supported without closed-loop evaluation. read the letter →

arxiv 2505.07600 v1 pith:YBF76FOX submitted 2025-05-12 cs.RO cs.CV

Oriol Barbany , Adrià Colomé , Carme Torras This is my paper · ORCID

classification cs.ROcs.CV

keywords clothfoldingvision-languagemodelstemporalcontextkeyframeslow-rankadaptationdeformableobjectmanipulationattentionmapsactionprediction

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper argues that a vision-language model cannot fold clothes reliably from a single snapshot, because crumpled and self-occluded garments make the next pick and place ambiguous. It analyzes BiFold, a VLM policy that conditions each action on the current instruction and observation plus a short history of keyframes, and claims that this history is the main driver of accurate action prediction. On the BiFold dataset, adding three keyframes raises average precision at 5-pixel tolerance from 40.3 to 75.1, while using consecutive raw frames (34.6) performs worse than no context at all. The paper also presents evidence that low-rank fine-tuning makes the model's image and text features align spatially across frames, so attention lands on task-relevant cloth regions rather than background.

What carries the argument

The load-bearing object is the keyframe-conditioned policy $\pi_\theta(a_t \mid \ell_t, o_t, o_{t-1}, \ldots, o_{t-H})$ with $H=3$, where each observation is encoded by the same image encoder, prepended with a learned token, and combined with modality and positional embeddings before a transformer encoder fuses all tokens via cross-modal attention. The output tokens for the current observation are decoded by convolutional networks into probability distributions over pick and place positions for both arms. Keyframes rather than raw frames matter because high-frequency consecutive frames are redundant; sampling before each action gives the model genuinely new information about the garment's evolving state.

What would settle it

Run the trained BiFold policy with $H=0$ and $H=3$ on a physical robot or a dynamics-enabled cloth simulator and measure the fraction of successfully completed folds; if keyframe history does not improve real folding success, the claim fails. A sharper check: feed the model the same current frame repeated as all three keyframes; if average precision is unchanged, then temporal state change, not extra tokens, is what the model actually uses.

Watch

Extended reading notes

Core claim

BiFold's central claim is that temporal context, specifically keyframes sampled before each action rather than raw consecutive frames, improves the model's ability to recognize and interpret complex object states during cloth folding. The paper supports this by comparing three variants: no context, consecutive-frame context, and keyframe context. Keyframe context dominates on pixel-space average precision (AP5 of 75.1 versus 40.3 for no context and 34.6 for consecutive frames), and the authors interpret the consecutive-frame failure as feature dilution from redundant, noisy tokens. Qualitatively, attention maps show that task-relevant text tokens such as 'bottom' and 'fold' attend to consistent garment regions across the current observation and past keyframes, and PCA visualizations show that fine-tuned features track the cloth across time while pretrained features do not.

Load-bearing premise

The BiFold dataset and its pixel-space average-precision scores faithfully track real cloth-folding competence, even though no physical-robot or simulation success rate is reported.

Editorial extensions

If this is right

VLM policies for deformable object manipulation should use sparsely sampled keyframes as temporal context; dense frame histories can actively hurt accuracy.
Explicit mesh or keypoint state representations are not required: a fine-tuned VLM can implicitly encode garment state from RGB observations and history.
Parameter-efficient fine-tuning (0.29% of encoder parameters) is enough to repurpose a pretrained vision-language encoder for language-conditioned cloth folding.
Text-to-image attention maps learned without segmentation supervision can localize task-relevant cloth regions such as fold lines and leg openings.
A short keyframe history should help the policy recover from failed manipulation steps, since the past observations reveal how the cloth arrived at its current state.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The same keyframe-conditioning mechanism could transfer to other deformable-object tasks such as bed-making, bagging, or surgical tissue manipulation, where self-occlusion and history matter similarly.
Because the paper reports only detection-style metrics on an offline dataset, the decisive next test is physical-robot or simulation success rate; a $H=3$ versus $H=0$ comparison on completed folds would directly test the claim.
The consecutive-frame failure suggests an adaptive frame-selection module, choosing the most informative past observations, could further improve accuracy with fewer tokens.
If the temporal alignment shown in attention maps reflects genuine state tracking, then adding explicit correspondence losses such as normalized object coordinates could push the model further, a direction the paper itself mentions.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

This is an interpretability companion to the authors' BiFold paper, not a new method, and it is upfront about that. The genuinely new piece is the three-way ablation: keyframe history, consecutive history, and no history. The numbers in Table I are striking—keyframes reach AP5 75.1 versus 40.3 with no context and 34.6 with consecutive frames—and the gap is internally consistent across AP10, AP50, and the quantile metric. The explanation that consecutive frames add redundant, noisy tokens is plausible and testable. The PCA and attention visualizations are also consistent with the claim that fine-tuning and temporal context focus the model on task-relevant cloth regions.

What I like: the paper is clearly written, the control conditions are the right ones for the stated claim, and the authors do not oversell the work as a new method. For a workshop-level analysis, this is solid and useful.

The soft spots are real but not disqualifying for an ablation study. The biggest one: the metrics are pixel-space AP and quantiles computed on the authors' own canonicalized VR-Folding dataset. Nothing in the paper shows that a 5-pixel or 10-pixel AP gain means a better grasp or fold on a physical garment. There is no real-robot success rate, no simulation rollout, no crumpled or recovery scenarios, even though the abstract motivates the work with exactly those cases. The offline metric could be rewarding the model for retrieving canonical action coordinates rather than reasoning about deformable state, and the attention maps do not establish a causal link to manipulation success. Table I also lacks error bars, episode counts, and significance tests, so the headline gap could be smaller under protocol variation. The self-referential nature—their model, their dataset, their interpretation—is fine for an ablation, but it limits how far the conclusions can travel. No code, weights, or evaluation data are visible in the manuscript.

I agree with the reader's conditional verdict and with the stress-test note. The central claim about physical cloth-folding state recognition is not fully supported; the claim about offline pixel-space accuracy is supported. For a serious venue, the paper needs a closed-loop evaluation, even a modest one, plus basic statistics. For a workshop or a position-style venue, it is acceptable as is. I would not cite it as evidence for physical benefit, but I would cite it as a careful internal analysis of temporal context in a VLM policy.

My recommendation: send it to peer review rather than desk-rejecting, but expect heavy revision before it becomes a fully convincing contribution.

Referee Report

4 major / 5 minor

Summary. This workshop paper analyzes BiFold, a vision-language-action model for bimanual cloth folding, and argues that temporal context in the form of keyframes improves the model's ability to perceive and interpret garment state. The authors compare BiFold (which conditions on three keyframe observations) against two ablations: no temporal context and consecutive-frame context. They report pixel-space Average Precision (AP) at 5, 10, and 50 px thresholds and an output quantile metric on the BiFold dataset, finding a large advantage for keyframe context (AP5 75.1 vs 40.3 vs 34.6). They also present qualitative PCA visualizations of image-encoder features and text-to-image attention maps, claiming that fine-tuning improves task-specific feature localization and that attention is temporally consistent. The paper does not report real-robot or simulation success rates, and all experiments are conducted on flattened, canonically aligned garment configurations, despite the abstract motivating crumpled garments and failure recovery.

Significance. If the central claim holds, the paper would provide a useful and falsifiable result for deformable-object manipulation: that the choice of temporal context (keyframes rather than consecutive frames) is decisive for language-conditioned action prediction, and that a parameter-efficient fine-tuned VLM can implicitly encode garment state. The comparison of keyframe vs consecutive vs no context is a clean ablation that goes beyond merely adding history, and the internally consistent ordering in Table I is suggestive. The paper also contributes a purely automatic dataset pipeline (from VR-Folding) and a publicly described architecture, which supports reproducibility. However, the significance is tempered by the fact that the evaluation is entirely offline on the authors' own dataset and metrics; no evidence connects the reported AP gains to physical folding success. The qualitative attention and PCA analyses are plausible but self-referential, and they are not backed by quantitative grounding. The core load-bearing claim therefore requires either closed-loop validation or a carefully narrowed wording.

major comments (4)

[Section III.A, Table I] Table I reports AP5/AP10/AP50 and quantile values without error bars, confidence intervals, significance tests, or the number of evaluation episodes. The headline comparison (BiFold AP5 75.1 vs w/o context 40.3 vs consecutive 34.6) could be within protocol variation if the evaluation set is small or the model is stochastic. Please report the number of evaluation actions/episodes, multiple seeds or checkpoints, and at least a paired bootstrap or standard error. This is load-bearing for claim (ii), which rests on the superiority of keyframe context.
[Abstract and Section II] The abstract motivates the temporal-context design by explicitly mentioning 'crumpled garments or recovery from failed manipulations,' yet Section II states that the system focuses on 'flattened configurations' and Section III.A evaluates only the BiFold dataset, which is built from canonically aligned flattened demonstrations. None of the experiments test crumpled starting states, mid-folding failures, or recovery scenarios. Claim (ii) is framed as improving 'complex object states,' but the evaluated states are all flattened and canonical. Either add experiments on crumpled/recovery configurations or narrow the claim to the flattened setting.
[Section III.A, metrics] The AP and quantile metrics are computed in the canonical pixel space of the BiFold dataset, which is generated by automatic action segmentation and canonical alignment per garment category. The paper does not show that an AP5 gain of 35 points corresponds to a physically better grasp or fold. There is no real-robot success rate, no simulation rollout, and no measure of task completion. The keyframe history could be helping the model retrieve canonical action coordinates rather than reason about the deformable state. Please provide a closed-loop evaluation (simulation or physical) or explicitly validate that the offline AP metric correlates with manipulation success.
[Section III.B and III.C] The qualitative evidence for claims (i) and (ii) consists of PCA visualizations of the model's own features and attention maps produced by the same model. These are consistent with temporal focus, but they do not by themselves establish a causal link to action-prediction quality, and the paper does not quantify the degree of temporal consistency or spatial localization. In addition, Section III.B contains an incomplete sentence: 'we retain only those patches whose first principal component is non-negative To address (ii) and evaluate the impact of temporal context...' — the methodological description of the PCA retention threshold and token-level aggregation is missing, making the visualization procedure not fully reproducible.

minor comments (5)

[Section III.B] Fix the sentence fragment at the end of the PCA threshold description ('non-negative To address (ii)') and describe precisely how PCA is performed: which layers/patches are used, how the retention threshold is applied, and whether the visualization shows a single example or an average over a dataset.
[Table I] Report the dataset split (training/validation/test sizes), the number of seeds, and training hyperparameters (LoRA rank, learning rate, epochs) so that the AP numbers can be reproduced or compared.
[Fig. 4] Specify whether the attention maps are averaged over attention heads and layers or shown for a single head/layer, and whether they are from one representative trial or aggregated over multiple scenarios.
[Section II] Since this paper analyzes BiFold, a self-contained description of the model would help: the original paper is cited as an arXiv preprint [1], but key architecture details such as the number of context frames H=3, the LoRA application points, and the decoder structure are only summarized. A short appendix or a link to released code would improve reproducibility.
[Section III.B] The claim that pre-trained SigLIP features 'struggle to capture semantic correspondences across time steps' is based on qualitative PCA plots. Please state whether any quantitative comparison (e.g., cosine similarity of corresponding patches across frames) was computed, or present the PCA as illustrative only.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the keyframe-context claim rests on reported ablations and visualizations, not on a self-referential construction.

full rationale

The paper's central claim (ii) — that keyframe history improves state recognition — is supported by the Table I ablation comparing BiFold with keyframe context, without context, and with consecutive frames. These are three distinct input configurations of the same policy, and the reported AP/quantile metrics are computed against ground-truth pick and place positions in the BiFold dataset. The outcome metric is not defined in terms of the model's own outputs, and no quantity is fitted and then renamed as a prediction. The qualitative attention analysis in Fig. 4 is additional evidence rather than a circular derivation. The paper does rely heavily on the authors' prior BiFold paper for both the model and the dataset, and this self-citation is prominent; however, it is used as provenance for the architecture and data pipeline, not as the evidential basis for the current comparative claims. The current claims stand on the ablations and visualizations reported in this manuscript. The absence of a described train/test split, the lack of error bars or significance tests, and the absence of real-robot or closed-loop validation are legitimate correctness and generalization concerns, but they do not amount to circularity under the definition used here. No step in the paper equates an input with a derived result by construction, and no load-bearing argument reduces to an unverified self-citation.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

No new physical or architectural entities are introduced. The analysis concerns BiFold, a model described in the authors' prior paper [1]; this paper contributes ablations and interpretability analyses only.

free parameters (2)

context length H = 3
Number of past keyframes fed to the policy; set to 3 in the base configuration and used in all experiments, Section II.
PCA retention threshold = first principal component non-negative
Analysis-only parameter applied when visualizing patch features, following DINOv2; it shapes the qualitative figures in Section III.B.

assumptions (4)

domain assumption The BiFold dataset, generated by an automatic pipeline from VR-Folding demonstrations, is a valid training and evaluation benchmark for cloth folding.
Section II states the pipeline is fully automatic and canonical alignment enables language instruction generation; the paper provides no independent validation that the extracted action and language pairs are correct.
domain assumption PCA projections and attention maps faithfully expose the model's internal semantic representations.
Sections III.B and III.C draw claims (i) and (ii) from color-coded PCA features and text-to-image attention maps on selected frames.
domain assumption Keyframes sampled before each action execution are the most informative temporal context, and H=3 suffices to capture state changes.
Section II defines the policy with H=3 and justifies keyframes over raw frames; no sensitivity analysis over H is presented.
standard math Pretrained SigLIP features, transformer cross-modal attention, and LoRA behave as standard machine-learning machinery.
The method section relies on standard transformer and LoRA formulations from references [17], [18], and [19].

how reviews work

0 comments

Cite this review

Pith. "Pith review of Beyond Static Perception: Integrating Temporal Context into VLMs for Cloth Folding." pith.science (2026). https://pith.science/paper/YBF76FOX

@misc{pith2026250507600,
  author       = {Pith},
  title        = {Pith review of: Beyond Static Perception: Integrating Temporal Context into VLMs for Cloth Folding},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/YBF76FOX}},
  note         = {Machine review of arXiv:2505.07600}
}

read the original abstract

Manipulating clothes is challenging due to their complex dynamics, high deformability, and frequent self-occlusions. Garments exhibit a nearly infinite number of configurations, making explicit state representations difficult to define. In this paper, we analyze BiFold, a model that predicts language-conditioned pick-and-place actions from visual observations, while implicitly encoding garment state through end-to-end learning. To address scenarios such as crumpled garments or recovery from failed manipulations, BiFold leverages temporal context to improve state estimation. We examine the internal representations of the model and present evidence that its fine-tuning and temporal context enable effective alignment between text and image regions, as well as temporal consistency.

Figures

Figures reproduced from arXiv: 2505.07600 by the authors.

**Figure 1.** Motivation: In multi-step tasks like cloth folding, the state can become visually ambiguous due to self-occlusions or compounding errors from past actions. Relying solely on the current observation can make the next action ill-defined (left). By incorporating temporal context (right), the model gains critical information enabling more accurate perception and decision-making. Abstract—Manipulating clothes is challe… view at source ↗

**Figure 2.** BiFold architecture: BiFold fine-tunes a pre-trained SigLIP [17] with LoRA [18] to obtain image and text features. The same image encoder is used for the current and past observations. Each token sequence is prepended with a different learned token, and we add information about the modality and the position inside the sequence by adding a learned embedding and sinusoidal positional encodings. All tokens are concaten… view at source ↗

**Figure 4.** Text→Image attention: Attention scores from text to image patches to the current observation and keyframes. As shown in [PITH_FULL_IMAGE:figures/full_fig_p003_4.png] view at source ↗

Figures from the paper (1 more)

**Figure 3.** Figure 3: Image encoder features: While pre-trained SigLIP features offer aligned language and visual embeddings, they struggle to capture semantic correspondences across time steps. In contrast, BiFold’s fine-tuned version focuses more effectively on the manipulated cloth, enha…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

31 extracted references · 24 canonical work pages

[1]

BiFold: Bimanual Cloth Folding with Language Guidance,

O. Barbany, A. Colom ´e, and C. Torras, “BiFold: Bimanual Cloth Folding with Language Guidance,” arXiv:2501.16458, 2025

arXiv 2025
[2]

TRTM: Template-based Reconstruction and Target-oriented Manipulation of Crumpled Cloths,

W. Wang, G. Li, M. Zamora, and S. Coros, “TRTM: Template-based Reconstruction and Target-oriented Manipulation of Crumpled Cloths,” inIEEE International Conference on Robotics and Automation (ICRA), 2024

work page 2024
[3]

Learning Language-Conditioned Deformable Object Manipulation with Graph Dynamics,

Y . Deng, K. Mo, C. Xia, and X. Wang, “Learning Language-Conditioned Deformable Object Manipulation with Graph Dynamics,” inIEEE International Conference on Robotics and Automation (ICRA), 2024

work page 2024
[4]

Foldsformer: Learning Sequential Multi-Step Cloth Manipulation With Space-Time Attention,

K. Mo, C. Xia, X. Wang, Y . Deng, X. Gao, and B. Liang, “Foldsformer: Learning Sequential Multi-Step Cloth Manipulation With Space-Time Attention,”IEEE Robotics and Automation Letters, vol. 8, no. 2, pp. 760–767, 2023

work page 2023
[5]

Fab- ricflownet: Bimanual cloth manipulation with a flow-based policy,

T. Weng, S. Bajracharya, Y . Wang, K. Agrawal, and D. Held, “Fab- ricflownet: Bimanual cloth manipulation with a flow-based policy,” in Conference on Robot Learning (CoRL), 2021

work page 2021
[6]

Gar- mentTracking: Category-Level Garment Pose Tracking,

H. Xue, W. Xu, J. Zhang, T. Tang, Y . Li, W. Du, R. Ye, and C. Lu, “Gar- mentTracking: Category-Level Garment Pose Tracking,” inIEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), June 2023, pp. 21 233–21 242

work page 2023
[7]

Garmentnets: Category-level pose estimation for garments via canonical space shape completion,

C. Chi and S. Song, “Garmentnets: Category-level pose estimation for garments via canonical space shape completion,” inIEEE International Conference on Computer Vision (ICCV), 2021

work page 2021
[8]

Quadratic dynamic matrix control for fast cloth manipulation,

E. Caldarelli, A. Colom ´e, C. Ocampo-Martinez, and C. Torras, “Quadratic dynamic matrix control for fast cloth manipulation,” in2023 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2023, pp. 8178–8185

work page 2023

Show all 31 references

[9]

Mesh-based dynamics with occlusion reasoning for cloth manipulation,

Z. Huang, X. Lin, and D. Held, “Mesh-based dynamics with occlusion reasoning for cloth manipulation,” inRobotics: Science and Systems (RSS), 2022

2022
[10]

Learning visible connectivity dynamics for cloth smoothing,

X. Lin, Y . Wang, Z. Huang, and D. Held, “Learning visible connectivity dynamics for cloth smoothing,” inConference on Robot Learning (CoRL), 2022, pp. 256–266

2022
[11]

Deformable surface reconstruc- tion via riemannian metric preservation,

O. Barbany, A. Colom ´e, and C. Torras, “Deformable surface reconstruc- tion via riemannian metric preservation,”Computer Vision and Image Understanding, vol. 249, p. 104155, 2024

2024
[12]

Cloth-Splatting: 3D State Estimation from RGB Supervision for Deformable Objects,

A. Longhini, M. B ¨usching, B. P. Duisterhof, J. Lundell, J. Ichnowski, M. Bj ¨orkman, and D. Kragic, “Cloth-Splatting: 3D State Estimation from RGB Supervision for Deformable Objects,” inConference on Robot Learning (CoRL), 2024

2024
[13]

Dynamic 3d gaussian tracking for graph-based neural dynamics modeling,

M. Zhang, K. Zhang, and Y . Li, “Dynamic 3d gaussian tracking for graph-based neural dynamics modeling,” inConference on Robot Learning (CoRL), 2024

2024
[14]

Reconstruction of Manipulated Garment with Guided Deformation Prior,

R. Li, C. Dumery, Z. Deng, and P. Fua, “Reconstruction of Manipulated Garment with Guided Deformation Prior,” inAdvances in Neural Information Processing Systems (NeurIPS), 2024

2024
[15]

Meshudf: Fast and differentiable meshing of unsigned distance field networks,

B. Guillard, F. Stella, and P. Fua, “Meshudf: Fast and differentiable meshing of unsigned distance field networks,” inEuropean Conference on Computer Vision (ECCV), 2022

2022
[16]

Learning keypoints for robotic cloth manipulation using synthetic data,

T. Lips, V .-L. De Gusseme, and F. Wyffels, “Learning keypoints for robotic cloth manipulation using synthetic data,”IEEE Robotics and Automation Letters, vol. 9, no. 7, pp. 6528–6535, 2024

2024
[17]

Sigmoid loss for language image pre-training,

X. Zhai, B. Mustafa, A. Kolesnikov, and L. Beyer, “Sigmoid loss for language image pre-training,” inIEEE International Conference on Computer Vision (ICCV), 2023, pp. 11 975–11 986

2023
[18]

LoRA: Low-rank adaptation of large language models,

E. J. Hu, Y . Shen, P. Wallis, Z. Allen-Zhu, Y . Li, S. Wang, L. Wang, and W. Chen, “LoRA: Low-rank adaptation of large language models,” inInternational Conference on Learning Representations (ICLR), 2022

2022
[19]

Attention is All you Need,

A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, L. u. Kaiser, and I. Polosukhin, “Attention is All you Need,” inNeural Information Processing Systems (NeurIPS), vol. 30, 2017

2017
[20]

RT-1: Robotics Transformer for Real-World Control at Scale,

A. Brohan, N. Brown, J. Carbajal, Y . Chebotar, J. Dabis, C. Finn, K. Gopalakrishnan, K. Hausman, A. Herzog, J. Hsu, J. Ibarz, B. Ichter, A. Irpan, T. Jackson, S. Jesmonth, N. Joshi, R. Julian, D. Kalashnikov, Y . Kuang, I. Leal, K.-H. Lee, S. Levine, Y . Lu, U. Malla, D. Manj...

2022 arXiv
[21]

Learning Transferable Visual Models From Natural Language Supervi- sion,

A. Radford, J. W. Kim, C. Hallacy, A. Ramesh, G. Goh, S. Agarwal, G. Sastry, A. Askell, P. Mishkin, J. Clark, G. Krueger, and I. Sutskever, “Learning Transferable Visual Models From Natural Language Supervi- sion,” inInternational Conference on Machine Learning (ICML), 2021

2021
[22]

OpenVLA: An Open-Source Vision-Language-Action Model,

M. Kim, K. Pertsch, S. Karamcheti, T. Xiao, A. Balakrishna, S. Nair, R. Rafailov, E. Foster, G. Lam, P. Sanketi, Q. Vuong, T. Kollar, B. Burchfiel, R. Tedrake, D. Sadigh, S. Levine, P. Liang, and C. Finn, “OpenVLA: An Open-Source Vision-Language-Action Model,”arXiv preprint ar...

2024 arXiv
[23]

Spoc: Imitating shortest paths in simulation enables effective navigation and manipulation in the real world,

K. Ehsani, T. Gupta, R. Hendrix, J. Salvador, L. Weihs, K.-H. Zeng, K. P. Singh, Y . Kim, W. Han, A. Herrasti, R. Krishna, D. Schwenk, E. VanderBilt, and A. Kembhavi, “Spoc: Imitating shortest paths in simulation enables effective navigation and manipulation in the real world,...

2024
[24]

Low-rank few-shot adaptation of vision- language models,

M. Zanella and I. Ben Ayed, “Low-rank few-shot adaptation of vision- language models,” inIEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), 2024, pp. 1593–1603

2024
[25]

Leveraging Large Language Models for Multimodal Search,

O. Barbany, M. Huang, X. Zhu, and A. Dhua, “Leveraging Large Language Models for Multimodal Search,” inIEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), 2024

2024
[26]

Reflective planning: Vision-language models for multi-stage long-horizon robotic manipulation,

Y . Feng, J. Han, Z. Yang, X. Yue, S. Levine, and J. Luo, “Reflective planning: Vision-language models for multi-stage long-horizon robotic manipulation,” arXiv:2502.16707, 2025

2025 arXiv
[27]

Octo: An open-source generalist robot policy,

Octo Model Team, D. Ghosh, H. Walke, K. Pertsch, K. Black, O. Mees, S. Dasari, J. Hejna, C. Xu, J. Luo, T. Kreiman, Y . Tan, L. Y . Chen, P. Sanketi, Q. Vuong, T. Xiao, D. Sadigh, C. Finn, and S. Levine, “Octo: An open-source generalist robot policy,” inRobotics: Science and S...

2024
[28]

RT-2: Vision- Language-Action Models Transfer Web Knowledge to Robotic Control,

A. Brohan, N. Brown, J. Carbajal, Y . Chebotar, X. Chen, K. Choroman- ski, T. Ding, D. Driess, A. Dubey, C. Finn, P. Florence, C. Fu, M. G. Arenas, K. Gopalakrishnan, K. Han, K. Hausman, A. Herzog, J. Hsu, B. Ichter, A. Irpan, N. Joshi, R. Julian, D. Kalashnikov, Y . Kuang, I....

2023 arXiv
[29]

DINOv2: Learning robust visual features without supervision,

M. Oquab, T. Darcet, T. Moutakanni, H. V . V o, M. Szafraniec, V . Khali- dov, P. Fernandez, D. HAZIZA, F. Massa, A. El-Nouby, M. Assran, N. Ballas, W. Galuba, R. Howes, P.-Y . Huang, S.-W. Li, I. Misra, M. Rabbat, V . Sharma, G. Synnaeve, H. Xu, H. Jegou, J. Mairal, P. Labatu...

2024
[30]

Normalized Object Coordinate Space for Category-Level 6D Object Pose and Size Estimation,

H. Wang, S. Sridhar, J. Huang, J. Valentin, S. Song, and L. J. Guibas, “Normalized Object Coordinate Space for Category-Level 6D Object Pose and Size Estimation,” inIEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), June 2019

2019
[31]

Elastictok: Adaptive tokenization for image and video,

W. Yan, V . Mnih, A. Faust, M. Zaharia, P. Abbeel, and H. Liu, “Elastictok: Adaptive tokenization for image and video,” inInternational Conference on Learning Representations (ICLR), 2025

2025

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.

[1] [1]

BiFold: Bimanual Cloth Folding with Language Guidance,

O. Barbany, A. Colom ´e, and C. Torras, “BiFold: Bimanual Cloth Folding with Language Guidance,” arXiv:2501.16458, 2025

arXiv 2025

[2] [2]

TRTM: Template-based Reconstruction and Target-oriented Manipulation of Crumpled Cloths,

W. Wang, G. Li, M. Zamora, and S. Coros, “TRTM: Template-based Reconstruction and Target-oriented Manipulation of Crumpled Cloths,” inIEEE International Conference on Robotics and Automation (ICRA), 2024

work page 2024

[3] [3]

Learning Language-Conditioned Deformable Object Manipulation with Graph Dynamics,

Y . Deng, K. Mo, C. Xia, and X. Wang, “Learning Language-Conditioned Deformable Object Manipulation with Graph Dynamics,” inIEEE International Conference on Robotics and Automation (ICRA), 2024

work page 2024

[4] [4]

Foldsformer: Learning Sequential Multi-Step Cloth Manipulation With Space-Time Attention,

K. Mo, C. Xia, X. Wang, Y . Deng, X. Gao, and B. Liang, “Foldsformer: Learning Sequential Multi-Step Cloth Manipulation With Space-Time Attention,”IEEE Robotics and Automation Letters, vol. 8, no. 2, pp. 760–767, 2023

work page 2023

[5] [5]

Fab- ricflownet: Bimanual cloth manipulation with a flow-based policy,

T. Weng, S. Bajracharya, Y . Wang, K. Agrawal, and D. Held, “Fab- ricflownet: Bimanual cloth manipulation with a flow-based policy,” in Conference on Robot Learning (CoRL), 2021

work page 2021

[6] [6]

Gar- mentTracking: Category-Level Garment Pose Tracking,

H. Xue, W. Xu, J. Zhang, T. Tang, Y . Li, W. Du, R. Ye, and C. Lu, “Gar- mentTracking: Category-Level Garment Pose Tracking,” inIEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), June 2023, pp. 21 233–21 242

work page 2023

[7] [7]

Garmentnets: Category-level pose estimation for garments via canonical space shape completion,

C. Chi and S. Song, “Garmentnets: Category-level pose estimation for garments via canonical space shape completion,” inIEEE International Conference on Computer Vision (ICCV), 2021

work page 2021

[8] [8]

Quadratic dynamic matrix control for fast cloth manipulation,

E. Caldarelli, A. Colom ´e, C. Ocampo-Martinez, and C. Torras, “Quadratic dynamic matrix control for fast cloth manipulation,” in2023 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2023, pp. 8178–8185

work page 2023