Pith. sign in

REVIEW 3 major objections 5 minor 67 references

UM-Adapt: Unsupervised Multi-Task Adaptation Using Adversarial Cross-Task Distillation

T0 review · 3 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read UM-Adapt uses cross-task coherence and contour preservation to adapt multiple dense prediction tasks to a new domain without target labels.

desk verdict The framework and ablations are solid, but the adversarial-energy mechanism and the semi-supervised SOTA claims are not backed by the reported numbers. read the letter →

arxiv 1908.03884 v3 pith:4CP7T3PP submitted 2019-08-11 cs.CV

Jogendra Nath Kundu , Nishank Lakkakula , R. Venkatesh Babu This is my paper · ORCID
classification cs.CV
keywords unsuperviseddomainadaptationmulti-tasklearningcross-taskdistillationdensepredictionsemanticsegmentationdepthestimationsurfacenormaladversarialenergy
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 show that a vision model can be adapted from synthetic to real images for several spatial prediction tasks at once—depth, semantic segmentation, and surface normals—without any labels on the real images. Its wager is that the relationships between tasks are themselves a usable signal: a network trained to predict one task map from another on synthetic data can be reused to check and improve the model's real-image predictions. On top of that, a contour-preserving regularizer keeps the adapted features spatially faithful to the input image. If this works, models trained entirely on rendered scenes could be dropped into real environments and fine-tuned with unlabeled photos, reducing the need for expensive annotation of dense outputs.

What carries the argument

The cross-task distillation module is the load-bearing mechanism: for each task $t_i$, a frozen encoder-decoder network $\mathcal{N}_{\to t_i}$ is trained on synthetic ground-truth maps to predict that task from the other tasks' maps, and the distillation loss compares the base model's direct prediction with the transferred prediction, thereby balancing tasks and serving as an energy function in a GAN-style alternating update. The second regularizer, CCR, is a shallow decoder trained to reconstruct the input image's contour map (ground truth from a standard edge-detection algorithm), and its frozen loss preserves spatial content in the target encoder. The architecture is partially shared: source and target encoders share up to Res-4, and only the target's Res-5 parameters are updated during adaptation.

What would settle it

Run the no-adversarial variant on a target domain where the spatial alignment between tasks differs from the synthetic source (for instance, synthetic indoor scenes adapted to real outdoor road scenes) and monitor the distillation loss against held-out task accuracy; if the loss decreases while accuracy stalls or drops, the assumed cross-task energy manifold is not transferring.

Watch

Extended reading notes

Core claim

The central claim is that unsupervised domain adaptation for multiple dense prediction tasks can be driven by two regularities: contour-based content regularization (CCR), which preserves the spatial contour structure of the input in the adapted features, and cross-task distillation, where small encoder-decoder task-transfer networks trained only on synthetic ground-truth task maps are reused both to balance task performance and, adversarially, as an energy function that detects domain shift. In this setup only the Res-5 block of a target encoder branch is trainable, with no target annotations. The paper reports that this yields top transfer results on ImageNet classification, comparable PASCAL VOC 2007 detection with a smaller backbone, and that the semi-supervised extension outperforms fully supervised multi-task learning on both NYUD and Cityscapes.

Load-bearing premise

The load-bearing premise is that task-transfer networks trained only on synthetic ground-truth maps define a meaningful energy manifold for target-domain predictions, so minimizing the distillation loss pushes the target encoder toward true task outputs rather than merely satisfying a synthetic-data prior.

Editorial extensions

If this is right

  • Only a small set of target encoder parameters needs to be adapted to transfer all three dense prediction tasks to a new domain without labels.
  • Contour-based content regularization improves over feature-consistency regularization for structured prediction adaptation.
  • Using the frozen task-transfer networks alone, without any extra discriminator, already yields stable unsupervised adaptation in the no-adversarial variant.
  • The semi-supervised variant matches or exceeds fully supervised multi-task learning baselines on NYUD and Cityscapes using far fewer target labels.
  • The adapted representation transfers to novel tasks such as ImageNet classification and PASCAL VOC detection, even with a smaller backbone.

Reading between the lines

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

  • The design suggests that a domain discriminator can be replaced by task-coherency losses whenever the output tasks share spatial structure, a principle that could generalize to optical flow, edges, keypoints, or other aligned predictions.
  • The method's effectiveness likely depends on how correlated the tasks are; adding a task with weak spatial alignment to the others could dilute the distillation signal, and that dependency is testable.
  • The contour regularizer could be lifted out and reused as a cheap content-preserving loss in single-task unsupervised adaptation.
  • Because the no-adversarial variant is simpler and reported as more stable, the adversarial energy update may be an optional refinement rather than a necessity in production settings.
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 / 5 minor

Summary. The paper proposes UM-Adapt, a unified framework for unsupervised multi-task adaptation of spatially-structured tasks (monocular depth, surface normals, semantic segmentation) under synthetic-to-real domain shift. The two proposed regularizers are contour-based content regularization (CCR), which matches contour maps reconstructed from the shared encoder against a contour prior, and cross-task distillation, in which frozen task-transfer networks map one task output to another and the discrepancy between direct and transferred predictions is used as a training signal. The paper additionally proposes an adversarial variant, UM-Adapt-(Adv.), in which the task-transfer networks are treated as energy functions and updated to score source ground-truth tuples low and target prediction tuples high, while the target encoder is updated to produce low-energy tuples. Experiments cover PBRS-to-NYUD adaptation (depth, normals, segmentation), GTA5-to-KITTI/Cityscapes adaptation, a semi-supervised variant UM-Adapt-S, and transfer learning to ImageNet classification and PASCAL VOC detection.

Significance. If the central claims were fully borne out, the paper would make a meaningful contribution: a single multi-task adaptation framework with a simple content regularizer (CCR) that visibly improves over the FCF baseline on all NYUD tasks, and a cross-task distillation objective that improves over the joint-discriminator baseline. The non-adversarial components are coherent and the ablations show consistent gains. However, the paper's distinctive advertised mechanism, the adversarial energy-based cross-task distillation, is not empirically supported: the differences between UM-Adapt-(noAdv.) and UM-Adapt-(Adv.) in Tables 1-3 are within run-to-run noise, with no error bars or multiple seeds. Furthermore, the headline semi-supervised SOTA claims rest on comparisons against fully-supervised baselines using far more labeled target data (30k vs. 795 for NYUD). The transfer-learning SOTA claim is based on a 0.21% top-5 improvement over a different backbone with no significance analysis. The non-adversarial core is credible, but the paper's framing and headline claims overstate what the evidence shows.

major comments (3)
  1. [Section 3.3.2 / Algorithm 2; Tables 1-3] The adversarial cross-task distillation is the paper's second and third stated contributions, yet the reported data do not demonstrate that it helps. Across Tables 1-3, UM-Adapt-(Adv.) versus UM-Adapt-(noAdv.) differ only by depth rel 0.178 vs. 0.175, normals mean 22.2 vs. 22.2, and segmentation mIoU 0.214 vs. 0.221. These differences are small and no error bars or multiple runs are reported. Section 4.2 itself states that UM-Adapt-(noAdv.) is 'significantly stable and faster' because it avoids updating the task-transfer networks. The paper therefore fails to support the claim that the adversarial energy mechanism drives the reported performance; the gains over UM-Adapt-B(CCR) could be attributed entirely to the non-adversarial distillation loss. Please either provide statistically meaningful evidence (multiple seeds, significance tests) for the advantage of the adversarial update, or reframe the contribution as non-adversarial cross-task distillation plus CCR, which the current numbers support.
  2. [Section 4.2; Tables 6-7] The claim that UM-Adapt-S 'outperforms the current fully-supervised multi-task learning state-of-the-art' is based on non-comparable supervision budgets. In Table 6, Kendall et al. and GradNorm are trained with 30k labeled NYUD samples, while UM-Adapt-S uses 795 labeled samples; the additional rows in Table 5 indicate UM-Adapt-S uses 500 labeled Cityscapes images, whereas the comparison methods in Table 7 are trained on full supervision. A semi-supervised method with far fewer labels can outperform a fully-supervised method, but the abstract and Section 4.2 present this as an unqualified SOTA result without acknowledging the label-count difference. Please report comparisons under matched label budgets, or clearly qualify the claim and state the supervision counts for each baseline.
  3. [Section 4.3; Table 8] The transfer-learning claim of 'state-of-the-art ... on ImageNet classification' rests on a 69.51 vs. 69.30 top-5 accuracy difference against Doersch et al. with a ResNet-101 backbone, while UM-Adapt-S uses a ResNet-50. This 0.21-point gap is well within typical run-to-run noise and no error bars or multiple runs are provided. On PASCAL detection UM-Adapt-S is actually lower than the best baseline (70.02 vs. 70.53). The claim should be softened to 'comparable' or supported with more rigorous comparisons including uncertainty estimates and matched backbones.
minor comments (5)
  1. [Abstract / Section 4.3] The abstract's 'state-of-the-art transfer learning results on ImageNet classification' is not justified by the single comparison against Doersch et al.; the evaluation uses only one prior method family and no significance analysis.
  2. [Section 3.2] The notation 'GT con.' is introduced without explicitly stating that it is the output of a standard contour detection algorithm (HED); please define it at first use.
  3. [Section 4.1 / Supplementary] Important hyperparameters appear only in the supplementary: the pseudo-softmax noise scale (truncated normal with std 0.075) and the gradient mask proportion (50%). These are part of the method and should be stated in the main text or at least referenced explicitly in the main experimental section.
  4. [Throughout] There are numerous typos and inconsistent spellings: 'Muti-task learning' in the introduction, 'sematic' instead of 'semantic' in Tables 3 and 5, and 'Liana et al.' instead of 'Laina et al.' in the experimental setting. These should be corrected in a polished revision.
  5. [Figure 2] Figure 2 is dense and difficult to read at publication resolution; the task-transfer network architectures are only in the supplementary and should be summarized in the caption or main text.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the UM-Adapt objective is a standard adversarial consistency regularizer evaluated on held-out target benchmarks, and no prediction reduces by construction to a fitted input.

full rationale

The paper's unsupervised adaptation (Algorithm 2) is a minimax game rather than a derivation: the task-transfer networks N->ti are trained on synthetic source ground-truth output maps, and the target branch Mt is updated using the transfer-consistency energy L_G^ti = L_ti(Yhat^{->ti}_t, Yhat^ti_t) plus the contour-based self-supervised loss L_CCR. Target labels are never used in the unsupervised variant, and the NYUD, Cityscapes, ImageNet, and PASCAL VOC numbers are measured on held-out target data, so the reported results are not fitted inputs renamed as predictions. CCR uses contour pseudo-labels obtained from the input image by a standard contour detector, and the contour decoder is frozen, so it is a self-supervised regularizer, not a target-label fit. Self-citations to AdaDepth [29] provide architectural initialization and an ablation baseline, but the central claim does not rest on an unverified uniqueness theorem or on a self-citation chain. The paper itself notes that UM-Adapt-(Adv.) gives only minimal improvement over UM-Adapt-(noAdv.) and that noAdv. is 'significantly stable and faster'; this is an effect-size and robustness concern, not a circularity. No equation in the paper reduces to its own input by construction, so no circular step can be exhibited.

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

The central claim rests primarily on the empirical effectiveness of the proposed regularizers, with no mathematical derivation. The free parameters are training hyperparameters (alpha and the unreported lambda) and ad hoc distribution tricks; the axioms are domain assumptions about contour alignment, transferability of the task-transfer networks, and the sufficiency of Res-5-only adaptation. No invented entities are introduced.

free parameters (4)
  • Distillation loss weight alpha = 10
    Reported in Section 4.1 'Training details' (alpha = 10). Chosen by hand; no sensitivity study.
  • CCR loss weight lambda
    Used in Algorithm 2 as lambda L_CCR but its numerical value is not reported in the paper, preventing exact replication.
  • Pseudo-softmax noise scale = truncated Normal std 0.075, mean 1
    Ad hoc trick in Supplementary to represent one-hot segmentation maps; 50% of batches use one-hot and 50% use pseudo-softmax.
  • Gradient mask proportion = 50%
    Supplementary: randomly generated gradient mask layer in 50% of the batch for class balancing.
assumptions (4)
  • domain assumption Majority of image contours align with the contours of task-specific output maps.
    Section 3.2 states this to justify using contour reconstruction (CCR) as a content-preserving regularizer for the latent representation.
  • domain assumption Task-transfer networks trained on source ground-truth maps approximate the cross-task energy manifold for target predictions.
    Section 3.3.2 argues the compressed encoder-decoder allows N_to_ti to generalize to target samples without negative examples; the entire adversarial adaptation relies on this.
  • domain assumption Adapting only the Res-5 block of the target encoder is sufficient to bridge the domain shift.
    Section 3.1.2 and Algorithm 2 restrict trainable target parameters to theta_res5, following AdaDepth; no experiments vary this capacity.
  • standard math Standard stochastic optimization (Adam, WGAN-GP training) converges to a useful solution.
    The paper relies on standard backpropagation and adversarial training stability without proof.

how reviews work

0 comments
Cite this review

Pith. "Pith review of UM-Adapt: Unsupervised Multi-Task Adaptation Using Adversarial Cross-Task Distillation." pith.science (2026). https://pith.science/paper/4CP7T3PP

@misc{pith2026190803884,
  author       = {Pith},
  title        = {Pith review of: UM-Adapt: Unsupervised Multi-Task Adaptation Using Adversarial Cross-Task Distillation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/4CP7T3PP}},
  note         = {Machine review of arXiv:1908.03884}
}
read the original abstract

Aiming towards human-level generalization, there is a need to explore adaptable representation learning methods with greater transferability. Most existing approaches independently address task-transferability and cross-domain adaptation, resulting in limited generalization. In this paper, we propose UM-Adapt - a unified framework to effectively perform unsupervised domain adaptation for spatially-structured prediction tasks, simultaneously maintaining a balanced performance across individual tasks in a multi-task setting. To realize this, we propose two novel regularization strategies; a) Contour-based content regularization (CCR) and b) exploitation of inter-task coherency using a cross-task distillation module. Furthermore, avoiding a conventional ad-hoc domain discriminator, we re-utilize the cross-task distillation loss as output of an energy function to adversarially minimize the input domain discrepancy. Through extensive experiments, we demonstrate superior generalizability of the learned representations simultaneously for multiple tasks under domain-shifts from synthetic to natural environments. UM-Adapt yields state-of-the-art transfer learning results on ImageNet classification and comparable performance on PASCAL VOC 2007 detection task, even with a smaller backbone-net. Moreover, the resulting semi-supervised framework outperforms the current fully-supervised multi-task learning state-of-the-art on both NYUD and Cityscapes dataset.

Figures

Figures reproduced from arXiv: 1908.03884 by the authors.

Figure 1
Figure 1. A schematic diagram to understand the implications [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. An overview of the proposed UM-Adapt architecture for multi-task adaptation. The blue and pink background wide-channel indicates data flow for synthetic and natural domain respectively. On the right we show an illustration of the proposed cross-task distillation module, which is later utilized as an energy-function for adversarial adaptation (Section 3.3.2). an optimal strategy would be to adapt minimally possible p… view at source ↗
Figure 3
Figure 3. An overview of the (a) proposed CCR framework with [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Qualitative comparison of different ablations of [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]
Figure 1
Figure 1. Figure 1: An overview of the proposed [PITH_FULL_IMAGE:figures/full_fig_p011_1.png]
Figure 2
Figure 2. Figure 2: Qualitative comparison of different ablations of [PITH_FULL_IMAGE:figures/full_fig_p014_2.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

67 extracted references · 54 canonical work pages

  1. [1]

    Real-time monocular depth estimation using synthetic data with do- main adaptation via image style transfer

    Amir Atapour-Abarghouei and Toby P Breckon. Real-time monocular depth estimation using synthetic data with do- main adaptation via image style transfer. In CVPR, 2018. 2, 4

    work page 2018

  2. [2]

    Unsupervised pixel- level domain adaptation with generative adversarial net- works

    Konstantinos Bousmalis, Nathan Silberman, David Dohan, Dumitru Erhan, and Dilip Krishnan. Unsupervised pixel- level domain adaptation with generative adversarial net- works. In CVPR, 2017. 2

    work page 2017

  3. [3]

    Multitask learning

    Rich Caruana. Multitask learning. Machine learning , 28(1):41–75, 1997. 3

    work page 1997

  4. [4]

    No more discrimi- nation: Cross city adaptation of road scene segmenters

    Yi-Hsin Chen, Wei-Yu Chen, Yu-Ting Chen, Bo-Cheng Tsai, Yu-Chiang Frank Wang, and Min Sun. No more discrimi- nation: Cross city adaptation of road scene segmenters. In ICCV, 2017. 1, 8

    work page 2017

  5. [5]

    The cityscapes dataset for semantic urban scene understanding

    Marius Cordts, Mohamed Omran, Sebastian Ramos, Timo Rehfeld, Markus Enzweiler, Rodrigo Benenson, Uwe Franke, Stefan Roth, and Bernt Schiele. The cityscapes dataset for semantic urban scene understanding. In CVPR,

  6. [6]

    Domain adaptation for visual applications: A comprehensive survey

    Gabriela Csurka. Domain adaptation for visual applications: A comprehensive survey. arXiv preprint arXiv:1702.05374,

    arXiv

  7. [7]

    Multi-task self- supervised visual learning

    Carl Doersch and Andrew Zisserman. Multi-task self- supervised visual learning. In ICCV, 2017. 8

    work page 2017

  8. [8]

    Predicting depth, surface nor- mals and semantic labels with a common multi-scale convo- lutional architecture

    David Eigen and Rob Fergus. Predicting depth, surface nor- mals and semantic labels with a common multi-scale convo- lutional architecture. In ICCV, 2015. 1, 3, 6

    work page 2015

Show all 67 references
  1. [9]

    Depth map prediction from a single image using a multi-scale deep net- work

    David Eigen, Christian Puhrsch, and Rob Fergus. Depth map prediction from a single image using a multi-scale deep net- work. In NIPS, pages 2366–2374, 2014. 3, 6, 7

    2014

  2. [10]

    Unsupervised domain adaptation by backpropagation

    Yaroslav Ganin and Victor Lempitsky. Unsupervised domain adaptation by backpropagation. In ICML, 2015. 2

    2015

  3. [11]

    Domain-adversarial train- ing of neural networks

    Yaroslav Ganin, Evgeniya Ustinova, Hana Ajakan, Pas- cal Germain, Hugo Larochelle, Franc ¸ois Laviolette, Mario Marchand, and Victor Lempitsky. Domain-adversarial train- ing of neural networks. The Journal of Machine Learning Research, 17(1):2096–2030, 2016. 2

    2016

  4. [12]

    Bastiaan Kleijn, Mengjie Zhang, and David Balduzzi

    Muhammad Ghifary, W. Bastiaan Kleijn, Mengjie Zhang, and David Balduzzi. Domain generalization for object recog- nition with multi-task autoencoders. In ICCV, 2015. 1

    2015

  5. [13]

    Bros- tow

    Cl ´ement Godard, Oisin Mac Aodha, and Gabriel J. Bros- tow. Unsupervised monocular depth estimation with left- right consistency. In CVPR, 2017. 7

    2017

  6. [14]

    Generative adversarial nets

    Ian Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron Courville, and Yoshua Bengio. Generative adversarial nets. In NIPS, 2014. 2

    2014

  7. [15]

    Gretton, AJ

    A. Gretton, AJ. Smola, J. Huang, M. Schmittfull, KM. Borg- wardt, and B. Sch ¨olkopf. Covariate shift and local learning by distribution matching. MIT Press. 2

  8. [16]

    Heterogeneous face attribute estimation: A deep multi-task learning approach

    Hu Han, Anil K Jain, Shiguang Shan, and Xilin Chen. Heterogeneous face attribute estimation: A deep multi-task learning approach. IEEE transactions on pattern analysis and machine intelligence, 2017. 3

    2017

  9. [17]

    A joint many-task model: Growing a neural network for multiple nlp tasks

    Kazuma Hashimoto, Caiming Xiong, Yoshimasa Tsuruoka, and Richard Socher. A joint many-task model: Growing a neural network for multiple nlp tasks. arXiv preprint arXiv:1611.01587, 2016. 2

    2016 arXiv

  10. [18]

    Distilling the knowledge in a neural network

    Geoffrey Hinton, Oriol Vinyals, and Jeff Dean. Distilling the knowledge in a neural network. 2014. 4

    2014

  11. [19]

    Cycada: Cycle-consistent adversarial domain adapta- tion

    Judy Hoffman, Eric Tzeng, Taesung Park, Jun-Yan Zhu, Phillip Isola, Kate Saenko, Alexei A Efros, and Trevor Dar- rell. Cycada: Cycle-consistent adversarial domain adapta- tion. In ICML, 2018. 2, 4, 7

    2018

  12. [20]

    Fcns in the wild: Pixel-level adversarial and constraint-based adaptation

    Judy Hoffman, Dequan Wang, Fisher Yu, and Trevor Darrell. Fcns in the wild: Pixel-level adversarial and constraint-based adaptation. arXiv preprint arXiv:1612.02649, 2016. 7

    2016 arXiv

  13. [21]

    Conditional generative adversarial network for struc- tured domain adaptation

    Weixiang Hong, Zhenzhen Wang, Ming Yang, and Junsong Yuan. Conditional generative adversarial network for struc- tured domain adaptation. In CVPR, 2018. 2

    2018

  14. [22]

    Phillip Isola, Jun-Yan Zhu, Tinghui Zhou, and Alexei A. Efros. Image-to-image translation with conditional adver- sarial networks. In CVPR, 2017. 2

    2017

  15. [23]

    Analyzing modu- lar cnn architectures for joint depth prediction and semantic segmentation

    Omid Hosseini Jafari, Oliver Groth, Alexander Kirillov, Michael Ying Yang, and Carsten Rother. Analyzing modu- lar cnn architectures for joint depth prediction and semantic segmentation. In ICRA, 2017. 6

    2017

  16. [24]

    Multi-task learning using uncertainty to weigh losses for scene geome- try and semantics

    Alex Kendall, Yarin Gal, and Roberto Cipolla. Multi-task learning using uncertainty to weigh losses for scene geome- try and semantics. In CVPR, 2018. 2, 3, 8

    2018

  17. [25]

    Adam: A method for stochastic optimization

    Diederik Kingma and Jimmy Ba. Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980 ,

    arXiv

  18. [26]

    Recurrent scene parsing with perspective understanding in the loop

    Shu Kong and Charless Fowlkes. Recurrent scene parsing with perspective understanding in the loop. arXiv preprint arXiv:1705.07238, 2017. 6

    2017 arXiv

  19. [27]

    Saliency unified: A deep architecture for simultaneous eye fixation prediction and salient object segmentation

    Srinivas SS Kruthiventi, Vennela Gudisa, Jaley H Dholakiya, and R Venkatesh Babu. Saliency unified: A deep architecture for simultaneous eye fixation prediction and salient object segmentation. In CVPR, 2016. 3

    2016

  20. [28]

    Unsupervised feature learning of human actions as trajectories in pose embedding manifold

    Jogendra Nath Kundu, Maharshi Gor, Phani Krishna Uppala, and Venkatesh Babu Radhakrishnan. Unsupervised feature learning of human actions as trajectories in pose embedding manifold. In WACV, 2019. 1

    2019

  21. [29]

    Adadepth: Unsupervised content congruent adaptation for depth estimation

    Jogendra Nath Kundu, Phani Krishna Uppala, Anuj Pahuja, and R Venkatesh Babu. Adadepth: Unsupervised content congruent adaptation for depth estimation. In CVPR, 2018. 1, 2, 3, 4, 5, 7

    2018

  22. [30]

    Deeper depth prediction with fully convolutional residual networks

    Iro Laina, Christian Rupprecht, Vasileios Belagiannis, Fed- erico Tombari, and Nassir Navab. Deeper depth prediction with fully convolutional residual networks. In 3DV, 2016. 3, 5, 6

    2016

  23. [31]

    Refinenet: Multi-path refinement networks for high- resolution semantic segmentation

    Guosheng Lin, Anton Milan, Chunhua Shen, and Ian D Reid. Refinenet: Multi-path refinement networks for high- resolution semantic segmentation. In CVPR, 2017. 6

    2017

  24. [32]

    Efficient piecewise training of deep structured models for semantic segmentation

    Guosheng Lin, Chunhua Shen, Anton Van Den Hengel, and Ian Reid. Efficient piecewise training of deep structured models for semantic segmentation. In CVPR, 2016. 6

    2016

  25. [33]

    Deep con- volutional neural fields for depth estimation from a single image

    Fayao Liu, Chunhua Shen, and Guosheng Lin. Deep con- volutional neural fields for depth estimation from a single image. In CVPR, 2015. 6

    2015

  26. [34]

    End- to-end multi-task learning with attention

    Shikun Liu, Edward Johns, and Andrew J Davison. End- to-end multi-task learning with attention. arXiv preprint arXiv:1803.10704, 2018. 8

    2018 arXiv

  27. [35]

    Fully convolutional networks for semantic segmentation

    Jonathan Long, Evan Shelhamer, and Trevor Darrell. Fully convolutional networks for semantic segmentation. In CVPR, 2015. 6

    2015

  28. [36]

    Learning transferable features with deep adaptation net- works

    Mingsheng Long, Yue Cao, Jianmin Wang, and Michael Jor- dan. Learning transferable features with deep adaptation net- works. In ICML, 2015. 2

    2015

  29. [37]

    Unsupervised domain adaptation with residual trans- fer networks

    Mingsheng Long, Han Zhu, Jianmin Wang, and Michael I Jordan. Unsupervised domain adaptation with residual trans- fer networks. In NIPS, 2016. 5

    2016

  30. [38]

    Cross-stitch networks for multi-task learning

    Ishan Misra, Abhinav Shrivastava, Abhinav Gupta, and Mar- tial Hebert. Cross-stitch networks for multi-task learning. In CVPR, 2016. 2, 3

    2016

  31. [39]

    Joint semantic segmentation and depth estimation with deep convolutional networks

    Arsalan Mousavian, Hamed Pirsiavash, and Jana Ko ˇseck´a. Joint semantic segmentation and depth estimation with deep convolutional networks. In 3DV. IEEE, 2016. 6

    2016

  32. [40]

    Image to image translation for domain adaptation

    Zak Murez, Soheil Kolouri, David Kriegman, Ravi Ra- mamoorthi, and Kyungnam Kim. Image to image translation for domain adaptation. In CVPR, 2018. 4

    2018

  33. [41]

    Learning features by watching ob- jects move

    Deepak Pathak, Ross Girshick, Piotr Dollar, Trevor Darrell, and Bharath Hariharan. Learning features by watching ob- jects move. In CVPR, 2017. 8

    2017

  34. [42]

    Geonet: Geometric neural network for joint depth and surface normal estimation

    Xiaojuan Qi, Renjie Liao, Zhengzhe Liu, Raquel Urtasun, and Jiaya Jia. Geonet: Geometric neural network for joint depth and surface normal estimation. In CVPR, 2018. 6

    2018

  35. [43]

    Hy- perface: A deep multi-task learning framework for face de- tection, landmark localization, pose estimation, and gender recognition

    Rajeev Ranjan, Vishal M Patel, and Rama Chellappa. Hy- perface: A deep multi-task learning framework for face de- tection, landmark localization, pose estimation, and gender recognition. IEEE Transactions on Pattern Analysis and Ma- chine Intelligence, 2017. 3

    2017

  36. [44]

    Faster r-cnn: Towards real-time object detection with region proposal networks

    Shaoqing Ren, Kaiming He, Ross Girshick, and Jian Sun. Faster r-cnn: Towards real-time object detection with region proposal networks. In NIPS, pages 91–99, 2015. 8

    2015

  37. [45]

    Cross-domain self- supervised multi-task feature learning using synthetic im- agery

    Zhongzheng Ren and Yong Jae Lee. Cross-domain self- supervised multi-task feature learning using synthetic im- agery. In CVPR, 2018. 2, 3

    2018

  38. [46]

    Playing for data: Ground truth from computer games

    Stephan R Richter, Vibhav Vineet, Stefan Roth, and Vladlen Koltun. Playing for data: Ground truth from computer games. In ECCV, 2016. 6

    2016

  39. [47]

    Monocular depth esti- mation using neural regression forest

    Anirban Roy and Sinisa Todorovic. Monocular depth esti- mation using neural regression forest. In CVPR, 2016. 6

    2016

  40. [48]

    Imagenet large scale visual recognition challenge

    Olga Russakovsky, Jia Deng, Hao Su, Jonathan Krause, San- jeev Satheesh, Sean Ma, Zhiheng Huang, Andrej Karpathy, Aditya Khosla, Michael Bernstein, et al. Imagenet large scale visual recognition challenge. International journal of computer vision, 115(3):211–252, 2015. 8

    2015

  41. [49]

    Improved techniques for training gans

    Tim Salimans, Ian Goodfellow, Wojciech Zaremba, Vicki Cheung, Alec Radford, and Xi Chen. Improved techniques for training gans. In NIPS, 2016. 1

    2016

  42. [50]

    Learning from synthetic data: Addressing domain shift for semantic segmentation

    Swami Sankaranarayanan, Yogesh Balaji, Arpit Jain, Ser Nam Lim, and Rama Chellappa. Learning from synthetic data: Addressing domain shift for semantic segmentation. In CVPR, 2018. 4

    2018

  43. [51]

    Make3d: Learning 3d scene structure from a single still image

    Ashutosh Saxena, Min Sun, and Andrew Y Ng. Make3d: Learning 3d scene structure from a single still image. IEEE Transactions on Pattern Analysis and Machine Intelligence, 31(5):824–840, 2009. 6

    2009

  44. [52]

    Indoor segmentation and support inference from rgbd images

    Nathan Silberman, Derek Hoiem, Pushmeet Kohli, and Rob Fergus. Indoor segmentation and support inference from rgbd images. In ECCV. Springer, 2012. 6

    2012

  45. [53]

    Deep coral: Correlation alignment for deep domain adaptation

    Baochen Sun and Kate Saenko. Deep coral: Correlation alignment for deep domain adaptation. In ECCV Workshops,

  46. [54]

    Tsai, W.-C

    Y .-H. Tsai, W.-C. Hung, S. Schulter, K. Sohn, M.-H. Yang, and M. Chandraker. Learning to adapt structured output space for semantic segmentation. In CVPR, 2018. 5, 7

    2018

  47. [55]

    Simultaneous deep transfer across domains and tasks

    Eric Tzeng, Judy Hoffman, Trevor Darrell, and Kate Saenko. Simultaneous deep transfer across domains and tasks. In ICCV, 2015. 2, 7

    2015

  48. [56]

    Adversarial discriminative domain adaptation

    Eric Tzeng, Judy Hoffman, Trevor Darrell, and Kate Saenko. Adversarial discriminative domain adaptation. In CVPR,

  49. [57]

    Deep domain confusion: Maximizing for domain invariance

    Eric Tzeng, Judy Hoffman, Ning Zhang, Kate Saenko, and Trevor Darrell. Deep domain confusion: Maximizing for domain invariance. arXiv preprint arXiv:1412.3474, 2014. 1

    2014 arXiv

  50. [58]

    Towards unified depth and seman- tic prediction from a single image

    Peng Wang, Xiaohui Shen, Zhe Lin, Scott Cohen, Brian Price, and Alan L Yuille. Towards unified depth and seman- tic prediction from a single image. In CVPR, 2015. 6

    2015

  51. [59]

    Design- ing deep networks for surface normal estimation

    Xiaolong Wang, David Fouhey, and Abhinav Gupta. Design- ing deep networks for surface normal estimation. In CVPR,

  52. [60]

    Holistically-nested edge detection

    Saining ”Xie and Zhuowen” Tu. Holistically-nested edge detection. In ICCV, 2015. 4

    2015

  53. [61]

    Describing the scene as a whole: Joint object detection, scene classification and semantic segmentation

    Jian Yao, Sanja Fidler, and Raquel Urtasun. Describing the scene as a whole: Joint object detection, scene classification and semantic segmentation. In CVPR, 2012. 3

    2012

  54. [62]

    Curricu- lum domain adaptation for semantic segmentation of urban scenes

    Yang Zhang, Philip David, and Boqing Gong. Curricu- lum domain adaptation for semantic segmentation of urban scenes. In ICCV, 2017. 1, 7

    2017

  55. [63]

    Physically-based rendering for indoor scene understanding using convolutional neural networks

    Yinda Zhang, Shuran Song, Ersin Yumer, Manolis Savva, Joon-Young Lee, Hailin Jin, and Thomas Funkhouser. Physically-based rendering for indoor scene understanding using convolutional neural networks. In CVPR, 2017. 6

    2017

  56. [64]

    Energy- based generative adversarial network

    Junbo Zhao, Michael Mathieu, and Yann LeCun. Energy- based generative adversarial network. In ICLR, 2017. 2, 5

    2017

  57. [65]

    Gradnorm: Gradient normalization for adaptive loss balancing in deep multitask networks

    Chen-Yu Lee Zhao Chen, Vijay Badrinarayanan and Andrew Rabinovich. Gradnorm: Gradient normalization for adaptive loss balancing in deep multitask networks. In ICML, 2018. 3, 8

    2018

  58. [66]

    Unsupervised learning of depth and ego-motion from video

    Tinghui Zhou, Matthew Brown, Noah Snavely, and David G Lowe. Unsupervised learning of depth and ego-motion from video. In CVPR, 2017. 7

    2017

  59. [67]

    Unpaired image-to-image translation using cycle- consistent adversarial networks

    Jun-Yan Zhu, Taesung Park, Phillip Isola, and Alexei A Efros. Unpaired image-to-image translation using cycle- consistent adversarial networks. In ICCV, 2017. 2

    2017

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.