Pith. sign in

REVIEW 3 major objections 6 minor 59 references

Attribute-Based Robotic Grasping with Data-Efficient Adaptation

T0 review · 3 major / 6 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read The paper claims that an end-to-end visual-textual grasping model, pretrained self-supervised on 36 simulated basic blocks and adapted with unlabeled target images plus one successful grasp per target, attains 86.0% instance grasping…

desk verdict Solid empirical follow-up; the adversarial adaptation gain is likely inflated because the unlabeled images come from the exact test objects. read the letter →

arxiv 2501.02149 v1 pith:ZVPSF7KK submitted 2025-01-04 cs.RO cs.AI

Yang Yang , Houjian Yu , Xibai Lou , Yuanhao Liu , Changhyun Choi This is my paper · ORCID
classification cs.ROcs.AI
keywords roboticgraspinginstanceobjectattributesmultimodallearningdomainadaptationfew-shotself-supervisedaffordances
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

The paper tries to establish that a robot can learn to grasp a target object described by attributes such as color, shape, and category name, and then adapt to entirely new objects with almost no task-specific data. It claims that pretraining on 36 simple simulated blocks, with no human labels beyond the simulation itself, is enough to learn generic attribute representations that transfer to unseen objects. Two cheap adaptation steps are then proposed: adversarial adaptation using unlabeled images, and one-grasp adaptation using a single successful grasp of the target. The central experimental assertion is that this combination reaches 86.0% instance grasping success on simulated novel objects and 81.7% on real-world objects, beating the compared baselines by large margins. A sympathetic reading is that this makes text-commanded grasping practical for a robot entering a new environment.

What carries the argument

The load-bearing machinery is a gated-attention fusion of a convolutional visual encoder and a deep-averaging text encoder, followed by a fully convolutional affordance decoder that is evaluated over $N=6$ rotations of the fused feature map. The joint embedding is trained with a triplet loss that encodes the equation of object persistence: the visual embedding difference between the pre-grasp and post-grasp heightmaps is enforced to equal the text embedding of the query, written as $\phi_{v,\text{vec}}(v_{\text{pre}}) - \phi_{v,\text{vec}}(v_{\text{post}}) \approx \phi_t(t)$. For adaptation, a gradient reversal layer makes the image encoder learn domain-invariant features from augmented unlabeled images, while one-grasp adaptation rotates a single successful grasp into $N$ training views and fine-tunes the whole encoder-decoder with the grasping loss $L_{\text{grasp}}$.

What would settle it

Run the same pipeline with a genuine natural-language interface that produces free-form nouns, synonyms, and unseen attribute words, and measure instance grasping success on the same target objects; a large drop from the reported 81.7% would expose the system's dependence on hand-chosen phrases rather than on attribute learning itself.

Watch

Extended reading notes

Core claim

The core claim is that object attributes can serve as the shared currency between vision and language for instance grasping, and that a metric space learned from the persistence of objects before and after grasping gives the model a structure that supports very-shot adaptation. With only basic synthetic objects as pretraining data, the model localizes the target by text-conditioned attention and predicts pixel-wise grasping affordances; after combining adversarial adaptation on unlabeled images and end-to-end fine-tuning on one rotated grasp, it grasps novel targets 81.7% of the time in the real world. The paper further asserts that both adaptation mechanisms are independently useful, that their gains are additive, and that the one-grasp method outperforms a zero-shot foundation-model baseline at grounding attribute descriptions in multi-object scenes.

Load-bearing premise

The load-bearing premise is that every user query can be reduced to the paper's small attribute vocabulary (color, shape, and category name) and that the text encoder correctly grounds those words, because the experiments never evaluate a real language parser.

Editorial extensions

If this is right

  • If the 86.0% and 81.7% rates hold, a robot can be deployed for a new object with roughly one grasp's worth of interaction and a handful of unlabeled images, rather than a large labeled dataset.
  • The additive gains of adversarial and one-grasp adaptation imply that recognition-level transfer and manipulation-level fine-tuning address different failure modes and can be stacked.
  • The structured attribute metric space is what makes one-grasp adaptation work; without the metric loss, the adaptation gains are much smaller.
  • Rotation augmentation of one grasp is sufficient to make the model approximately rotation-invariant for both recognition and grasping.
  • The comparison with a zero-shot foundation model suggests that text-image matching alone does not reliably pick the target in clutter, while the adapted model's affordance map does.

Reading between the lines

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

  • One extension the authors leave implicit is whether the same pretrain-and-adapt recipe transfers to other manipulation primitives such as suction, pushing, or placing, which they state should in principle be possible but do not test.
  • Because the experiments use hand-chosen attribute phrases, a testable boundary is whether a real language parser producing synonyms, negation, or novel attribute words would preserve the reported success rates.
  • If the attribute metric space is as generic as claimed, attribute words absent from the pretraining vocabulary, such as 'striped' or 'shiny', should require a new one-grasp adaptation rather than working zero-shot; this is a direct, checkable consequence.
  • The paper's comparison with a zero-shot foundation model hints that fine-tuning such a model with the same one-grasp data might combine broad pretraining with task-specific grounding, but that combination is not tested.
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.

Referee Report

3 major / 6 minor

Summary. The paper proposes an end-to-end attribute-based instance grasping system. A visual-textual encoder-decoder is pre-trained in simulation on 36 basic colored/shaped blocks, using self-supervised grasp outcomes and a triplet loss that enforces object persistence between pre- and post-grasp scene embeddings. To transfer to novel objects and real scenes, the paper introduces two data-efficient adaptations: adversarial adaptation, which uses unlabeled target-domain images with object-level augmentation, and one-grasp adaptation, which fine-tunes the whole model from one successful grasp of a target object with rotation augmentation. Experiments in simulation and on a real robot report a generic model achieving 72.1% and 63.1% instance grasping success on simulated novel and real novel objects, respectively, and an adapted Adversarial+One-Grasp model achieving 86.0% and 81.7%, outperforming several baselines. The paper concludes that these results are obtained on unknown objects.

Significance. If the claims hold, the paper is a valuable contribution to data-efficient language-conditioned robotic grasping. The self-supervised pretraining on basic blocks, the object-persistence metric learning, and the two stackable adaptation methods are coherent and well-motivated. The experimental scale is substantial: 1,200 and 3,400 simulation test cases for basic and novel objects, respectively, and 252 real-world trials per method. The ablations of the two augmentation schemes (Tables IV and V) and the attention/confusion-matrix analyses are informative. I agree with the reader's main concern: the adversarial adaptation evaluation is confounded because the unlabeled target set Dt is the same set of objects used for testing. This weakens the headline generalization claim. The natural-language assumption is also not evaluated. These issues are fixable with additional experiments and careful rephrasing, so the central idea remains defensible.

major comments (3)
  1. [Sec. V-A, VII-B, VII-C, Table III, Fig. 13] The target-domain dataset Dt is defined in Sec. V-A as unlabeled images of 'all conceivable objects in the target domain,' and in the experiments this set coincides exactly with the 34 simulated novel objects (Sec. VII-B) and 21 real-world objects (Sec. VII-C) on which the adapted models are later tested. The Adversarial and Adversarial+One-Grasp models in Table III and Fig. 13 have therefore already seen every test instance through unlabeled images before the reported trials. The concluding statement in Sec. VIII that these results are 'both on unknown objects' is misleading: the adversarial component is evaluated only on objects contained in its adaptation set, and the one-grasp component is per-target supervision by design. The observed gains (72.1 to 86.0 in simulation; 63.1 to 81.7 in the real world) conflate domain-invariant feature learning with instance memorization. The manuscript should add an object-disjoint evaluation in which Dt is built from a held-out object set and the test objects are different instances of the same domain, reporting those numbers separately. This is essential to support the central generalization claim.
  2. [Sec. III, VII-B, VII-C, VIII] Assumption 1, which states that a language attribute parser extracts color, shape, and category keywords from query text, is never evaluated. All experiments use hand-chosen attribute phrases for each object, so the reported grasping success rates assume an oracle parser. The abstract and conclusion describe the system as responding to user commands and query text without this caveat. The authors should either evaluate an actual parser on free-form commands or explicitly state in the abstract and conclusion that all results assume the parser succeeds; without this, the natural-language interface claim is not supported.
  3. [Table III, Fig. 13, Sec. VII-E] The adaptation results are reported as point estimates without confidence intervals or significance tests. The margins that motivate the combined Adversarial+One-Grasp method over One-Grasp alone are modest (86.0% vs 83.7% in simulation; 81.7% vs 76.6% in the real world). With 252 real-world trials per method, a difference of 5.1 percentage points is within sampling variability, so the claim that the combined method is 'consistently the best' is not statistically supported. Please provide confidence intervals, exact trial counts, or repeated-run statistics for the adaptation comparisons.
minor comments (6)
  1. [Sec. IV-C, Algorithm 1] The training description says the model is trained for 5k iterations and then 'we replay the entire data for 100 epochs'; please clarify whether the 100 epochs occur once after the 5k online iterations or interleaved with collection, as this affects reproducibility.
  2. [Sec. V-A] The phrase 'all conceivable objects in the target domain' is imprecise; please specify how the object set for Dt is selected and whether it is intended to be the same as or disjoint from the test objects.
  3. [Sec. VII-D] The CLIP comparison is asymmetric: the proposed method is evaluated on multi-object workspace images while CLIP is evaluated on single-object crops. This makes the comparison favorable to the proposed method and should be acknowledged or corrected.
  4. [Sec. V-B] The sentence 'The token embedding of the object name is initialized properly to keep the embedding vector of the query text unchanged' needs details; for reproducibility, please specify how out-of-vocabulary category names are embedded in the Word2Vec CBOW vocabulary.
  5. [Eqs. (1) and (5)] Equations (1) and (5) contain typographical artifacts (NsX and |T |X) that should be corrected.
  6. [Fig. 10] The bar label 'One-Grasp, w/o attr.' does not correspond to a method name in Table III; please align the label with a described baseline.

Circularity Check

0 steps flagged · score 2.0 of 10

No circular derivation; the main empirical claims are self-contained, with only a validation-protocol overlap that is not circularity.

full rationale

The paper's core claims are empirical and are not derived from their inputs by construction. Pre-training is self-supervised from simulator interaction data (Algorithm 1), with the multimodal embedding supervised by the object-persistence constraint, not by the test results. The grasping label q_e is assigned as the attribute similarity between the query and the grasped object (Sec. IV-C), which is a reward/ground-truth definition used for training, not a prediction equivalent to the evaluation metric. The two adaptation methods, adversarial adaptation and one-grasp adaptation, are concrete loss-based updates (Eqs. 9-10 and Sec. V-B) and are evaluated by physical grasping trials; they do not reduce by definition to the reported success rates. The only self-citation, reference [1] for One-Grasp, is used as a baseline and as prior description of that adaptation component, but the method is also fully described in the present paper, so the citation is not load-bearing. The most significant concern is evaluative rather than circular: the adversarial target dataset Dt is described as containing all conceivable objects in the target domain (Sec. V-A), and the experiments then test on the same 34 simulated novel objects and 21 real objects that supplied adaptation images (Secs. VII-B and VII-C), so the adapted results may overstate generalization to unseen instances. This is a validation-protocol issue, not a circular derivation, and it does not make the claimed 86.0%/81.7% numbers equivalent to the model's training inputs by construction.

Assumptions & free parameters 7 free parameters · 6 assumptions · 0 invented entities

The method's training signal comes from simulator interaction, not from a closed-form derivation. The central learning assumptions are that attributes are expressible as a small dictionary, that visual scene changes before and after grasping align with text embeddings, that domain-adversarial invariance transfers to the decoder, and that simulator-generated labels are reliable. No new physical entities are introduced. Many hyperparameters are named but not given values, which matters for replication.

free parameters (7)
  • triplet margin alpha = not reported
    Controls the margin in the attribute metric loss L_attr (Eq. 5); hand-chosen but the value is not stated.
  • metric loss weight lambda_a = not reported
    Balances grasp motion loss and attribute metric loss in Eq. 6; value not reported.
  • background mask weight lambda_M = not reported
    Weights background pixels in the grasp loss Eq. 1; value not reported.
  • GRL reversal constant lambda_r = not reported
    Sets gradient reversal strength in Eq. 8 during adversarial adaptation; value not reported.
  • exploration epsilon = not reported
    Used in epsilon-greedy data collection in Algorithm 1; value not reported.
  • number of grasp orientations N = 6
    Rotates fused features by N=6 (30-degree increments) for grasp angle search and one-grasp data augmentation.
  • object-level augmentation parameters = not reported
    Scaling, flipping, rotation, and visual jitter magnitudes for ObjectAug are described qualitatively but not quantified.
assumptions (6)
  • domain assumption Query text is parsed into keywords of object attributes by an external parser.
    Assumption 1 in Section III; the entire system assumes color, shape, and category words can be extracted from natural language, and no parser is evaluated.
  • domain assumption Objects are stably placed within the workspace and there is no stacking between objects.
    Assumption 2 in Section III; the perception and top-down grasp primitive depend on this placement condition.
  • domain assumption Object persistence: the visual embedding difference before and after grasping equals the textual embedding of the grasped object.
    Used in Section IV-B to supervise the multimodal metric space via triplet loss; this is a modeling assumption about how scene changes correspond to object descriptions.
  • standard math Word2Vec CBOW embeddings plus a deep averaging network adequately represent attribute text.
    The text encoder in Section IV-A relies on a standard continuous bag-of-words model; the pretrained word vectors and their coverage of attribute terms are not detailed.
  • domain assumption Domain-adversarial training with a gradient reversal layer learns domain-invariant features that improve grasp decoding.
    Section V-A assumes that making pooled visual features indistinguishable across domains transfers to better grasp affordance predictions, although the adversarial loss only sees the global average pooled vector.
  • domain assumption Simulator-generated ground-truth grasp labels and attribute similarity labels are reliable.
    Section IV-C derives labels automatically from the simulation, assuming the simulator knows which object was grasped and its attributes.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Attribute-Based Robotic Grasping with Data-Efficient Adaptation." pith.science (2026). https://pith.science/paper/ZVPSF7KK

@misc{pith2026250102149,
  author       = {Pith},
  title        = {Pith review of: Attribute-Based Robotic Grasping with Data-Efficient Adaptation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ZVPSF7KK}},
  note         = {Machine review of arXiv:2501.02149}
}
read the original abstract

Robotic grasping is one of the most fundamental robotic manipulation tasks and has been the subject of extensive research. However, swiftly teaching a robot to grasp a novel target object in clutter remains challenging. This paper attempts to address the challenge by leveraging object attributes that facilitate recognition, grasping, and rapid adaptation to new domains. In this work, we present an end-to-end encoder-decoder network to learn attribute-based robotic grasping with data-efficient adaptation capability. We first pre-train the end-to-end model with a variety of basic objects to learn generic attribute representation for recognition and grasping. Our approach fuses the embeddings of a workspace image and a query text using a gated-attention mechanism and learns to predict instance grasping affordances. To train the joint embedding space of visual and textual attributes, the robot utilizes object persistence before and after grasping. Our model is self-supervised in a simulation that only uses basic objects of various colors and shapes but generalizes to novel objects in new environments. To further facilitate generalization, we propose two adaptation methods, adversarial adaption and one-grasp adaptation. Adversarial adaptation regulates the image encoder using augmented data of unlabeled images, whereas one-grasp adaptation updates the overall end-to-end model using augmented data from one grasp trial. Both adaptation methods are data-efficient and considerably improve instance grasping performance. Experimental results in both simulation and the real world demonstrate that our approach achieves over 81% instance grasping success rate on unknown objects, which outperforms several baselines by large margins.

Figures

Figures reproduced from arXiv: 2501.02149 by the authors.

Figure 1
Figure 1. Attribute-based instance grasping. Various objects of generic attributes are placed in the workspace, and we propose to grasp a target object by describing its attributes, e.g., “Please give me the apple, a red sphere.”. be similarly commanded utilizing such object attributes (e.g., color, shape, and category name, etc.), it would allow better generalization capability for novel objects than using a discrete set of … view at source ↗
Figure 2
Figure 2. Overview of affordances and attribute learning. The workspace image and query text are encoded separately and fused using gated-attention. The fusion matrix Fatt is rotated by N orientations for different grasping angles and then fed into the grasping affordances decoder. The decoder learns to predict pixel-wise scores of target grasping success, and we run the ϵ-greedy grasping policy and obtain the image vpost aft… view at source ↗
Figure 4
Figure 4. Examples of basic objects. Synthetic objects of various colors and shapes are used for learning object attributes and grasping affordances. To ensure shape attribute learning, we include objects having random textures. Algorithm 1 Online Data Collection Initialize bounded buffer B Notations: ϵ-greedy policy πϵ, our model ϕ, image v, text t, mask M, action a, and label q¯e 1: while collecting data do 2: reset the sim… view at source ↗
Figures from the paper (8 more)
Figure 5
Figure 5. Figure 5: Testing objects in simulation and the real world. We use the testing objects that share similar attributes with the training objects. See Appendix for more details. for learning object attributes. To deal with sparse rewards in target-driven grasping, the hindsight exp…
Figure 8
Figure 8. Figure 8: One-grasp data augmentation. We place the target object solely on the workspace and collect one successful grasp. To enrich the distribution of grasping angles, we rotate the collected data and synthesize a few samples for one-grasp adaptation. One rotation example of …
Figure 9
Figure 9. Figure 9: Visualization of attention and grasping. (a) shows attention of our encoder, and (b) shows heatmaps of grasping affordances for different targets (described by the query text). The maps of attention and grasping are consistent even for the novel objects. 3) to evaluate…
Figure 10
Figure 10. Figure 10: The adaptation gains of instance grasping made by dif￾ferent adaptation methods. The plot shows the effectiveness of our One-Grasp and Adversarial+One-Grasp adaptation approach, which achieve significant adaptation gains. 3) NoMetric takes as input images and text, an…
Figure 11
Figure 11. Figure 11: Visualization of grasping maps before and after adap￾tation. (a) shows the grasping affordances from the generic model trained only with simulated basic objects, and (b) to (d) show the affordances from the adapted models after Adversarial, One-Grasp, and Adversarial+…
Figure 12
Figure 12. Figure 12: The visualization of confusion matrices. The comparison between the two confusion matrices shows the effectiveness of our adaptation method trained with a minimum amount of data. various objects. To evaluate the performance of CLIP, we segment and crop randomly placed…
Figure 13
Figure 13. Figure 13: Adaptation performance in the designed testing environments. The instance grasping success rate of four approaches in the four testing environments. The plot shows the effectiveness of the proposed two data-efficient adaptation methods, Adversarial and One-Grasp adapt…
Figure 14
Figure 14. Figure 14: Example images of the designed testing environments for ablative analysis of adaptation. The environments are created to test the adaption performance of the grasping models as the degree of domain shift increases. TABLE IV: Ablations for Object-Level Augmentation (%)…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

59 extracted references · 53 canonical work pages

  1. [1]

    Attribute-based robotic grasping with one-grasp adaptation,

    Y . Yang, Y . Liu, H. Liang, X. Lou, and C. Choi, “Attribute-based robotic grasping with one-grasp adaptation,” in IEEE International Conference on Robotics and Automation (ICRA) . IEEE, 2021

    work page 2021

  2. [2]

    Multi-task domain adaptation for deep learning of instance grasping from simulation,

    K. Fang, Y . Bai, S. Hinterstoisser, S. Savarese, and M. Kalakrishnan, “Multi-task domain adaptation for deep learning of instance grasping from simulation,” in IEEE International Conference on Robotics and Automation (ICRA). IEEE, 2018, pp. 3516–3523

    work page 2018

  3. [3]

    A deep learning approach to grasping the invisible,

    Y . Yang, H. Liang, and C. Choi, “A deep learning approach to grasping the invisible,” IEEE Robotics and Automation Letters , vol. 5, no. 2, pp. 2232–2239, 2020

    work page 2020

  4. [4]

    Covariate shift and local learning by distribution matching,

    J. Qui ˜nonero-Candela, M. Sugiyama, A. Schwaighofer, and N. Lawrence, “Covariate shift and local learning by distribution matching,” pp. 131–160, 2008

    work page 2008

  5. [5]

    Analysis of representations for domain adaptation,

    S. Ben-David, J. Blitzer, K. Crammer, F. Pereira et al. , “Analysis of representations for domain adaptation,” Advances in Neural Information Processing Systems (NIPS) , vol. 19, p. 137, 2007

    work page 2007

  6. [6]

    Learning visual affordances for robotic manipulation,

    A. Zeng, “Learning visual affordances for robotic manipulation,” Ph.D. dissertation, Princeton University, 2019

    2019

  7. [7]

    An overview of 3d object grasp synthesis algorithms,

    A. Sahbani, S. El-Khoury, and P. Bidaud, “An overview of 3d object grasp synthesis algorithms,” Robotics and Autonomous Systems, vol. 60, no. 3, pp. 326–336, 2012

    work page 2012

  8. [8]

    Data-driven grasp synthesis—a survey,

    J. Bohg, A. Morales, T. Asfour, and D. Kragic, “Data-driven grasp synthesis—a survey,” IEEE Transactions on Robotics , vol. 30, no. 2, pp. 289–309, 2013

    work page 2013

Show all 59 references
  1. [9]

    Learning to generate 6-dof grasp poses with reachability awareness,

    X. Lou, Y . Yang, and C. Choi, “Learning to generate 6-dof grasp poses with reachability awareness,” in IEEE International Conference on Robotics and Automation (ICRA) . IEEE, 2020, pp. 1532–1538

    2020

  2. [10]

    Learning synergies between pushing and grasping with self-supervised deep reinforcement learning,

    A. Zeng, S. Song, S. Welker, J. Lee, A. Rodriguez, and T. Funkhouser, “Learning synergies between pushing and grasping with self-supervised deep reinforcement learning,” in IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) . IEEE, 2018, pp. 4238–4245

    2018

  3. [11]

    Self-supervised interactive object segmentation through a singulation-and-grasping approach,

    H. Yu and C. Choi, “Self-supervised interactive object segmentation through a singulation-and-grasping approach,” in Computer Vision– ECCV 2022: 17th European Conference, Tel Aviv, Israel, October 23–27, 2022, Proceedings, Part XXXIX . Springer, 2022, pp. 621–637

    2022

  4. [12]

    Mechanical search: Multi-step retrieval of a target object occluded by clutter,

    M. Danielczuk, A. Kurenkov, A. Balakrishna, M. Matl, D. Wang, R. Mart´ın-Mart´ın, A. Garg, S. Savarese, and K. Goldberg, “Mechanical search: Multi-step retrieval of a target object occluded by clutter,” in International Conference on Robotics and Automation (ICRA) . IEEE, 2019...

    2019

  5. [13]

    Grasp2vec: Learning object representations from self-supervised grasping,

    E. Jang, C. Devin, V . Vanhoucke, and S. Levine, “Grasp2vec: Learning object representations from self-supervised grasping,” in Conference on Robot Learning (CoRL) , 2018, pp. 99–112

    2018

  6. [14]

    End- to-end learning of semantic grasping,

    E. Jang, S. Vijayanarasimhan, P. Pastor, J. Ibarz, and S. Levine, “End- to-end learning of semantic grasping,” in Conference on Robot Learning (CoRL), 2017, pp. 119–132

    2017

  7. [15]

    Ccan: Constraint co-attention network for instance grasping,

    J. Cai, X. Tao, H. Cheng, and Z. Zhang, “Ccan: Constraint co-attention network for instance grasping,” in IEEE International Conference on Robotics and Automation (ICRA) . IEEE, 2020, pp. 8353–8359

    2020

  8. [16]

    Describing objects by their attributes,

    A. Farhadi, I. Endres, D. Hoiem, and D. Forsyth, “Describing objects by their attributes,” in IEEE Conference on Computer Vision and Pattern Recognition (CVPR). IEEE, 2009, pp. 1778–1785

    2009

  9. [17]

    Object localization using color, texture and shape,

    Y . Zhong and A. K. Jain, “Object localization using color, texture and shape,” Pattern Recognition, vol. 33, no. 4, pp. 671–684, 2000

    2000

  10. [18]

    Attribute based object identification,

    Y . Sun, L. Bo, and D. Fox, “Attribute based object identification,” in IEEE International Conference on Robotics and Automation (ICRA) . IEEE, 2013, pp. 2096–2103

    2013

  11. [19]

    Affordance prediction via learned object attributes,

    T. Hermans, J. M. Rehg, and A. Bobick, “Affordance prediction via learned object attributes,” in IEEE International Conference on Robotics and Automation (ICRA): Workshop on Semantic Perception, Mapping, and Exploration. Citeseer, 2011, pp. 181–184

    2011

  12. [20]

    Online learn- ing of object representations by appearance space feature alignment,

    S. Pirk, M. Khansari, Y . Bai, C. Lynch, and P. Sermanet, “Online learn- ing of object representations by appearance space feature alignment,” in IEEE International Conference on Robotics and Automation (ICRA) . IEEE, 2020

    2020

  13. [21]

    Grounding language attributes to objects using bayesian eigenobjects,

    V . Cohen, B. Burchfiel, T. Nguyen, N. Gopalan, S. Tellex, and G. Konidaris, “Grounding language attributes to objects using bayesian eigenobjects,” in IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) . IEEE, 2019, pp. 1187–1194

    2019

  14. [22]

    Interactive text2pickup networks for natural language-based human–robot collaboration,

    H. Ahn, S. Choi, N. Kim, G. Cha, and S. Oh, “Interactive text2pickup networks for natural language-based human–robot collaboration,” IEEE Robotics and Automation Letters , vol. 3, no. 4, pp. 3308–3315, 2018

    2018

  15. [23]

    Cad2rl: Real single-image flight without a single real image,

    F. Sadeghi and S. Levine, “Cad2rl: Real single-image flight without a single real image,” in Robotics: Science and Systems (RSS) , 2017

    2017

  16. [24]

    Domain randomization for transferring deep neural networks from simulation to the real world,

    J. Tobin, R. Fong, A. Ray, J. Schneider, W. Zaremba, and P. Abbeel, “Domain randomization for transferring deep neural networks from simulation to the real world,” in IEEE/RSJ international conference on intelligent robots and systems (IROS) . IEEE, 2017, pp. 23–30

    2017

  17. [25]

    Domain ran- domization and generative models for robotic grasping,

    J. Tobin, L. Biewald, R. Duan, M. Andrychowicz, A. Handa, V . Kumar, B. McGrew, A. Ray, J. Schneider, P. Welinder et al. , “Domain ran- domization and generative models for robotic grasping,” in IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) . IEEE,...

    2018

  18. [26]

    Active domain randomization,

    B. Mehta, M. Diaz, F. Golemo, C. J. Pal, and L. Paull, “Active domain randomization,” in Conference on Robot Learning (CoRL) . PMLR, 2020, pp. 1162–1176

    2020

  19. [27]

    A survey on transfer learning,

    S. J. Pan and Q. Yang, “A survey on transfer learning,” IEEE Transac- tions on Knowledge and Data Dngineering , vol. 22, no. 10, pp. 1345– 1359, 2009

    2009

  20. [28]

    Adversarial discrim- inative domain adaptation,

    E. Tzeng, J. Hoffman, K. Saenko, and T. Darrell, “Adversarial discrim- inative domain adaptation,” in IEEE conference on computer vision and pattern recognition (CVPR) , 2017, pp. 7167–7176

    2017

  21. [29]

    Deep grasp adaptation through domain transfer,

    Y . Chen, J. Jiang, R. Lei, Y . Bekiroglu, F. Chen, and M. Li, “Deep grasp adaptation through domain transfer,” in IEEE International Conference on Robotics and Automation (ICRA) , 2023

    2023

  22. [30]

    One-shot learning of object categories,

    L. Fei-Fei, R. Fergus, and P. Perona, “One-shot learning of object categories,” IEEE Transactions on Pattern Analysis and Machine In- telligence, vol. 28, no. 4, pp. 594–611, 2006

    2006

  23. [31]

    Siamese neural networks for one-shot image recognition,

    G. Koch, R. Zemel, and R. Salakhutdinov, “Siamese neural networks for one-shot image recognition,” in ICML Deep Learning Workshop, vol. 2. Lille, 2015

    2015

  24. [32]

    Prototypical networks for few- shot learning,

    J. Snell, K. Swersky, and R. Zemel, “Prototypical networks for few- shot learning,” in Advances in Neural Information Processing Systems (NIPS), 2017, pp. 4077–4087

    2017

  25. [33]

    Few-shot adver- sarial domain adaptation,

    S. Motiian, Q. Jones, S. Iranmanesh, and G. Doretto, “Few-shot adver- sarial domain adaptation,” in Advances in Neural Information Processing Systems (NIPS), 2017, pp. 6670–6680

    2017

  26. [34]

    A baseline for few-shot image classification,

    G. S. Dhillon, P. Chaudhari, A. Ravichandran, and S. Soatto, “A baseline for few-shot image classification,” in International Conference on Learning Representations (ICLR) , 2019

    2019

  27. [35]

    Referitgame: Referring to objects in photographs of natural scenes,

    S. Kazemzadeh, V . Ordonez, M. Matten, and T. Berg, “Referitgame: Referring to objects in photographs of natural scenes,” in Conference on Empirical Methods in Natural Language Processing (EMNLP), 2014, pp. 787–798

    2014

  28. [36]

    Imagenet: A large-scale hierarchical image database,

    J. Deng, W. Dong, R. Socher, L.-J. Li, K. Li, and L. Fei-Fei, “Imagenet: A large-scale hierarchical image database,” in IEEE conference on computer vision and pattern recognition (CVPR) . IEEE, 2009, pp. 248–255

    2009

  29. [37]

    Deep residual learning for image recognition,

    K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for image recognition,” in IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016, pp. 770–778

    2016

  30. [38]

    Deep unordered composition rivals syntactic methods for text classification,

    M. Iyyer, V . Manjunatha, J. Boyd-Graber, and H. Daum ´e III, “Deep unordered composition rivals syntactic methods for text classification,” in Association for Computational Linguistics (ACL) , 2015, pp. 1681– 1691

    2015

  31. [39]

    Rectified linear units improve restricted boltzmann machines,

    V . Nair and G. E. Hinton, “Rectified linear units improve restricted boltzmann machines,” in International conference on machine learning (ICML), 2010, pp. 807–814

    2010

  32. [40]

    Efficient estimation of word representations in vector space,

    T. Mikolov, K. Chen, G. Corrado, and J. Dean, “Efficient estimation of word representations in vector space,” arXiv preprint arXiv:1301.3781 , 2013. xiv

    2013 arXiv

  33. [41]

    Gated-attention architectures for task-oriented language grounding,

    D. S. Chaplot, K. M. Sathyendra, R. K. Pasumarthi, D. Rajagopal, and R. Salakhutdinov, “Gated-attention architectures for task-oriented language grounding,” in Thirty-Second AAAI Conference on Artificial Intelligence (AAAI), 2018

    2018

  34. [42]

    Fully convolutional networks for semantic segmentation,

    J. Long, E. Shelhamer, and T. Darrell, “Fully convolutional networks for semantic segmentation,” in IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2015, pp. 3431–3440

    2015

  35. [43]

    Learning visual affordances with target-orientated deep q-network to grasp objects by harnessing environmental fixtures,

    H. Liang, X. Lou, Y . Yang, and C. Choi, “Learning visual affordances with target-orientated deep q-network to grasp objects by harnessing environmental fixtures,” in IEEE International Conference on Robotics and Automation (ICRA) . IEEE, 2021, pp. 2562–2568

    2021

  36. [44]

    Network in network,

    M. Lin, Q. Chen, and S. Yan, “Network in network,” arXiv preprint arXiv:1312.4400, 2013

    2013 arXiv

  37. [45]

    Learning deep features for discriminative localization,

    B. Zhou, A. Khosla, A. Lapedriza, A. Oliva, and A. Torralba, “Learning deep features for discriminative localization,” in IEEE Conference on Computer Vision and Pattern Recognition (CVPR) , 2016, pp. 2921– 2929

    2016

  38. [46]

    Facenet: A unified em- bedding for face recognition and clustering,

    F. Schroff, D. Kalenichenko, and J. Philbin, “Facenet: A unified em- bedding for face recognition and clustering,” in IEEE Conference on Computer Vision and Pattern Recognition (CVPR) , 2015, pp. 815–823

    2015

  39. [47]

    R. S. Sutton and A. G. Barto, Reinforcement learning: An introduction . MIT press, 2018

    2018

  40. [48]

    Hindsight expe- rience replay,

    M. Andrychowicz, F. Wolski, A. Ray, J. Schneider, R. Fong, P. Welinder, B. McGrew, J. Tobin, O. P. Abbeel, and W. Zaremba, “Hindsight expe- rience replay,” in Advances in Neural Information Processing Systems (NIPS), 2017, pp. 5048–5058

    2017

  41. [49]

    Generative adversarial nets,

    I. Goodfellow, J. Pouget-Abadie, M. Mirza, B. Xu, D. Warde-Farley, S. Ozair, A. Courville, and Y . Bengio, “Generative adversarial nets,” Advances in Neural Information Processing Systems (NIPS) , vol. 27, 2014

    2014

  42. [50]

    Domain-adversarial training of neural networks,

    Y . Ganin, E. Ustinova, H. Ajakan, P. Germain, H. Larochelle, F. Lavio- lette, M. Marchand, and V . Lempitsky, “Domain-adversarial training of neural networks,” The Journal of Machine Learning Research , vol. 17, no. 1, pp. 2096–2030, 2016

    2016

  43. [51]

    M-adda: Unsupervised domain adaptation with deep metric learning,

    I. H. Laradji and R. Babanezhad, “M-adda: Unsupervised domain adaptation with deep metric learning,” in Domain adaptation for visual understanding. Springer, 2020, pp. 17–31

    2020

  44. [52]

    Spatial transformer networks,

    M. Jaderberg, K. Simonyan, A. Zisserman et al. , “Spatial transformer networks,” in Advances in Neural Information Processing Systems (NIPS), 2015, pp. 2017–2025

    2015

  45. [53]

    Revisiting data augmentation for rotational invariance in convolutional neural networks,

    F. Quiroga, F. Ronchetti, L. Lanzarini, and A. F. Bariviera, “Revisiting data augmentation for rotational invariance in convolutional neural networks,” in International Conference on Modelling and Simulation in Management Sciences . Springer, 2018, pp. 127–141

    2018

  46. [54]

    V-rep: A versatile and scalable robot simulation framework,

    E. Rohmer, S. P. Singh, and M. Freese, “V-rep: A versatile and scalable robot simulation framework,” in IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) . IEEE, 2013, pp. 1321–1326

    2013

  47. [55]

    The ycb object and model set: Towards common benchmarks for manipulation research,

    B. Calli, A. Singh, A. Walsman, S. Srinivasa, P. Abbeel, and A. M. Dollar, “The ycb object and model set: Towards common benchmarks for manipulation research,” in International Conference on Advanced Robotics (ICAR). IEEE, 2015, pp. 510–517

    2015

  48. [56]

    Chopra, Introduction to google sketchup

    A. Chopra, Introduction to google sketchup. John Wiley & Sons, 2012

    2012

  49. [57]

    Visp for visual servoing: a generic software platform with a wide class of robot control skills,

    ´E. Marchand, F. Spindler, and F. Chaumette, “Visp for visual servoing: a generic software platform with a wide class of robot control skills,” IEEE Robotics & Automation Magazine, vol. 12, no. 4, pp. 40–52, 2005

    2005

  50. [58]

    Moveit![ros topics],

    S. Chitta, I. Sucan, and S. Cousins, “Moveit![ros topics],” IEEE Robotics & Automation Magazine , vol. 19, no. 1, pp. 18–19, 2012

    2012

  51. [59]

    Learning transferable visual models from natural language supervision,

    A. Radford, J. W. Kim, C. Hallacy, A. Ramesh, G. Goh, S. Agarwal, G. Sastry, A. Askell, P. Mishkin, J. Clark et al., “Learning transferable visual models from natural language supervision,” in International conference on machine learning . PMLR, 2021, pp. 8748–8763. Yang Yang ...

    2021

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.