Pith. sign in

REVIEW 4 major objections 6 minor 61 references

$\varphi$-Adapt: A Physics-Informed Adaptation Learning Approach to 2D Quantum Material Discovery

T0 review · 4 major / 6 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read φ-Adapt claims that physics-informed adaptation lets models trained on 600,000 synthetic images beat real-data baselines for 2D flake detection, classification, and thickness estimation.

desk verdict Synthetic data pipeline is a real idea, but the central physics equation is dimensionally inconsistent as written; the paper needs major revision before the SOTA claims can be credited. read the letter →

arxiv 2507.05184 v1 pith:PBN22CCQ submitted 2025-07-07 cs.CV cs.LG

Hoang-Quan Nguyen , Xuan Bac Nguyen , Sankalp Pandey , Tim Faltermeier , Nicholas Borys , Hugh Churchill , Khoa Luu This is my paper
classification cs.CVcs.LG
keywords physics-informeddomainadaptation2Dmaterialidentificationflakethicknessestimationsyntheticdatagenerationtransfermatrixmethodsource-freeentropyminimizationopticalmicroscopyquantumflakes
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 finding and measuring 2D material flakes under an optical microscope can be done by training on synthetic images alone, then adapting to real images using the physics of thin-film reflection. It introduces φ-Adapt, a pipeline that trains task models on a 600,000-image synthetic dataset and then, without real labels, adjusts real microscope images with learned color-normalization and spectral-inversion modules so they match the synthetic distribution. The reported results include 34.1% AP for MoS2 flake detection, layer-classification accuracies up to 93.9%, and a thickness estimation error of 5.8 nm, all claimed to be state of the art. If correct, the method would reduce the need for large manually labeled real datasets in automated quantum flake discovery.

What carries the argument

The key machinery is the transfer-matrix model of multilayer thin-film reflection, Eqn. (1), together with the image formation equation $x = S^\top(I \circ R)$, Eqn. (2). The synthetic dataset is generated by choosing material, substrate, flake shape, and thickness, computing reflectance $R$ with the transfer-matrix method, and rendering with the CIE 1931 color matching functions and D65 illuminant. The adaptation network then inverts that process: ColorNorm estimates the white-balance factor $G_t$, SpecInv estimates a reflectance map $R_t$ from the RGB image, and Source Transform re-renders using the known source factors $A_s$.

What would settle it

Measure a real flake's spectral reflectance with a spectrometer, pass the same flake's microscope RGB image through φ-Adapt, and compare SpecInv's output to the measured spectrum; if the recovered reflectance does not match the measured spectrum, the physics-informed mechanism described by Eqns. (3)-(6) is not what is driving the reported accuracy.

Watch

Extended reading notes

Core claim

The central claim is that the visual gap between synthetic and real flake images decomposes into a known source transform $A_s$, an unknown sensor-and-illumination factor $A_t$, and an unknown white-balance factor $G_t$, so a real image can be mapped into the source domain by $x_{t\to s}=A_sR_t=A_s(A_t^{-1}(G_t^{-1}(x_t)))$. The paper introduces learnable modules, ColorNorm for $G_t$ and SpecInv for the reflectance $R_t$, plus a Source Transform that re-renders with the known synthetic illumination and sensor functions. It further claims that source-free entropy minimization and a neighbor-wavelength regularization on the learned optical parameters make this adaptation work on unlabeled target images. On the Masubuchi et al., Uslu et al., and its own collected benchmarks, the paper reports state-of-the-art accuracy for flake detection, layer classification, and thickness regression.

Load-bearing premise

The load-bearing premise is that the learned SpecInv module can recover, from a single RGB image, the true spectral reflectance that the physics equations require, even though the module outputs only one channel per pixel while the equations describe a per-wavelength reflectance with many channels.

Editorial extensions

If this is right

  • A detector, classifier, or thickness regressor can be trained entirely on synthetic images and transferred to real microscope images without collecting new labels.
  • Physics-based adaptation should generalize across materials and imaging setups better than purely statistical domain adaptation, because the source of the shift is explicitly modeled.
  • The same pipeline can be applied to flake detection, layer classification, and thickness estimation, with the reported thickness error of 5.8 nm bringing optical screening closer to AFM-level measurement.
  • Source-free entropy minimization means adaptation happens at test time, so each new experimental setup can be handled without retraining or annotation.
  • A large, precisely labeled synthetic dataset provides control over optical parameters that would be impractical to collect manually.

Reading between the lines

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

  • If SpecInv genuinely recovers physical reflectance, the same inversion could be applied to other layered or thin-film specimens, such as different exfoliated materials or coated surfaces, by only swapping the source rendering parameters.
  • A direct test of whether the physics is doing the work would be to measure a real flake's spectral reflectance with a spectrometer and compare it to SpecInv's predicted $R_t$; a mismatch would indicate the reported gains come from learned adaptation rather than the stated optical model.
  • The paper's Eqn. (8) outputs a single-channel reflectance $R_t \in \mathbb{R}^{H\times W\times 1}$, while the physics equations use a $D$-dimensional spectral reflectance, so the implementation likely approximates a spectral average rather than the full spectrum—this dimension mismatch is the point most worth probing.
  • A follow-up ablation replacing SpecInv with a simple per-pixel color-affine transform could show how much of the accuracy depends on the specific optical inversion versus the entropy-minimization adaptation itself.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 6 minor

Summary. The paper introduces φ-Adapt, a physics-informed source-free domain adaptation method for the detection, layer classification, and thickness estimation of 2D material flakes in optical microscopy images. It first proposes a synthetic data generation framework that uses the transfer matrix method to render 600,000 training images across eight materials and forty thickness configurations. It then presents an adaptation network with three modules: ColorNorm to estimate white-balance factors, SpecInv to recover a reflectance map from the color-normalized target image, and a Source Transform that maps the target image into the source domain using the known source illumination and sensitivity matrix. The method is evaluated on the Masubuchi et al. and Uslu et al. benchmarks, reporting state-of-the-art detection AP of 34.1%, layer-classification accuracies up to 93.9%, and a 5.8 nm error on a private thickness-estimation benchmark. The central claim is that physics-based adaptation with synthesized data removes the need for large real labeled datasets.

Significance. If the proposed method were sound, it would be a valuable contribution: the synthetic data generation pipeline is well-motivated and could substantially reduce the cost of collecting and labeling real 2D-material images, and the attempt to ground domain adaptation in optical physics is an important direction. The paper also reports strong empirical numbers on established benchmarks and compares against several recent source-free domain adaptation methods. However, the core physics-informed transform is not defined in a mathematically consistent manner, the per-wavelength regularization has no architectural counterpart, and the thickness-estimation experiment uses real labels and a private dataset, so the source-free claim is not clean. The absence of code, data, and checkpoints further prevents independent verification of the reported results. As written, the central claim is therefore unsupported.

major comments (4)
  1. [§4.2, Eqns. (6) and (8)] The source transform in Eqn. (6) is not mathematically well-defined. Because A_t = S_t^T diag(I_t) is a 3×D matrix (with D=128 as stated in §5.1), the inverse A_t^{-1} does not exist. Moreover, Eqn. (8) defines R_t ∈ R^{H×W×1}, a single-channel map, whereas Eqns. (2)–(4) require a D-dimensional spectral reflectance per pixel; the multiplication A_s R_t in Eqn. (6) is therefore undefined for D>1. To make the physics-informed claim operational, the authors must specify an actual H×W×D spectral output and a principled inversion (e.g., a learned pseudo-inverse with a spectral reconstruction loss), and verify that the implemented modules match this description. As it stands, the reported results cannot be attributed to the physics-based transform.
  2. [§4.4, Eqn. (9)] The neighbor regularization τ_neighbor(θ_SpecInv) in Eqn. (9) uses per-wavelength parameters θ_SpecInv(λ), but the SpecInv architecture described in Eqn. (8) is an encoder-decoder that outputs a single channel and has no explicit wavelength indexing. The paper never defines how θ_SpecInv(λ) is extracted from the network parameters or the output tensor. Without such a definition, the regularization term is vacuous, and the claim that the optical parameters are 'well-structured' is unsubstantiated.
  3. [§5.2, Table 4] The thickness-estimation experiment is not consistent with the advertised source-free protocol. The text states that real flake instances were collected and measured, and then 'we train a new linear regression head upon the trained backbone for the thickness estimation.' This implies that real thickness labels are used to train the regression head, which means the 5.8 nm error is not achieved under the proposed source-free adaptation setting. In addition, the benchmark is private: the manuscript provides no details on sample size, flake thickness distribution, measurement uncertainty of the AFM or reference method, or how the error is aggregated. The reported result is therefore not independently assessable.
  4. [§5.4, Table 5] The ablation table is ambiguous and does not clearly support the claimed incremental contributions. The first row shows only '41.4%' with no checkmarks, and the subsequent rows list combinations of module toggles without a clear mapping to the described settings (e.g., 'without and with color normalization' versus the entries with and without the Source Transform). The reader cannot determine which configuration corresponds to each row, making it impossible to verify the value added by each component. A corrected table with explicit rows for each ablation setting is needed.
minor comments (6)
  1. [§5.2] The dataset name 'Masubichi' is spelled inconsistently; the correct reference is 'Masubuchi'.
  2. [§4.2] The word 'intractible' should be 'intractable'.
  3. [Table 5] The column header 'Lent τneightbor' appears to be a typo for 'Entropy τ_neighbor' and the Greek letter is misspelled.
  4. [§3.1, Eqn. (1)] The transfer matrix notation M and P is introduced only briefly; a one-sentence definition of each matrix before Eqn. (1) would improve readability.
  5. [§4.4] The entropy minimization objective is defined for classification (C classes). For the thickness-estimation task, the paper first quantizes thickness into classes before adaptation; this transition should be stated explicitly in the method section, not only in the experiments.
  6. [§1] The claim of being 'one of the first' physics-informed adaptation methods is vague; the authors should either cite the specific prior works that define this space or rephrase to describe the precise novelty.

Circularity Check

0 steps flagged · score 0.0 of 10

No material circularity: the physics-informed adaptation pipeline is benchmarked against external data and its equations are not defined in terms of the reported metrics.

full rationale

The paper's derivation chain is: synthetic images are produced through xs = AsRs (Eq. 3), real images are modeled as xt = diag(Gt)AtRt (Eq. 4), and the adapted image is xt->s = AsRt using learned ColorNorm Gt and SpecInv Rt modules. None of these relations is defined in terms of the reported accuracy, AP, or thickness error, and the empirical results are measured on external or independently collected benchmarks, so the central claims are not forced by construction. The source dataset is generated from a transfer-matrix optical model, and the adaptation modules are trained on unlabeled target images via entropy minimization, which is a standard transductive domain-adaptation paradigm rather than a statistical fit to the evaluation labels. The main deficiency is a correctness/executability problem, not circularity: Eq. (6) requires A_t^{-1} although A_t is a 3xD matrix, and Eq. (8) outputs a single-channel Rt while Eqs. (3)-(6) require a D-dimensional reflectance; this makes the physics claim unsupported but does not reduce the derivation to its inputs. Self-citations appear only in related work and are not load-bearing for the adaptation result. The paper also explicitly acknowledges its limitation regarding sensor noise and artifacts, which is an honest scope statement rather than a circular step.

Assumptions & free parameters 3 free parameters · 6 assumptions · 2 invented entities

The central claim rests on the validity of the transfer-matrix optical model, the assumption that ColorNorm and SpecInv implement inverses of the unknown acquisition factors, and the assumption that a single-channel representation can carry the spectral reflectance needed for the source transform. The most fragile entry is the dimensionality mismatch between Eqn (6) and Eqn (8), and the undefined per-wavelength parameters in Eqn (9).

free parameters (3)
  • Source synthetic data distributions (flake shapes, thicknesses, material configurations)
    The 600k-image dataset is generated from hand-chosen random configurations; the paper gives no evidence these distributions match real target setups, yet the source task training depends on them.
  • Loss weights for entropy minimization and neighbor regularization
    Loss weights are not reported anywhere; Table 5 shows these losses change accuracy by 10+ points, so the reported results depend on tuned weights not disclosed.
  • ColorNorm white-balance vector Gt = 3-dim per image, learned
    A small network predicts a per-image color factor with no constraint that it equals the microscope's actual white balance; it can absorb arbitrary color shift during entropy minimization.
assumptions (6)
  • domain assumption The transfer matrix method (Eqn (1)) with Fresnel coefficients accurately models the optical reflectance of thin 2D flakes on SiO2/Si substrates.
    Central to synthetic image generation; no validation against measured reflectance spectra is given, and details are deferred to Supplementary.
  • domain assumption The CIE 1931 color matching functions and D65 illuminant correctly represent the source-domain microscope sensor and illumination (Eqn (3)).
    Used to define As; real microscopes have device-specific sensors and light sources.
  • domain assumption The real-domain shift is fully described by a diagonal per-image white balance Gt and a linear device transform At (Eqn (4)); sensor noise and other artifacts are ignored.
    The paper's own Limitations section admits sensor noise and artifacts are not considered, so the model cannot capture them.
  • ad hoc to paper A single-channel map Rt estimated from an RGB image can substitute for the full spectral reflectance in the source transform (Eqn (8) and Eqn (6)).
    This is load-bearing and dimensionally inconsistent: Eqn (3) has D-dimensional reflectance, Eqn (8) outputs one channel.
  • domain assumption Entropy minimization on unlabeled target predictions drives target features into source class clusters without degenerate collapse (Section 4.4).
    Standard but unverified assumption for this domain; no analysis of target entropy distributions is provided.
  • domain assumption Target flakes have the same material reflectance as synthesized flakes of the same material, so aligning to the synthetic source is valid.
    The source transformation assumes shared material-dependent reflectance; substrate, doping, or fabrication variations could violate this.
invented entities (2)
  • Single-channel reflectance map Rt from SpecInv
    purpose: Claimed to recover the material's reflection factors so the Source Transform can convert real images to synthetic style via Eqn (6).
    No comparison to measured spectral reflectance is provided, and the map's single channel contradicts the D-dimensional R in Eqn (3), so its physical meaning is unverified.
  • Per-wavelength inverse optical parameters theta_SpecInv(lambda)
    purpose: Used in the neighbor smoothness regularizer Eqn (9) to encourage smooth variation over neighboring wavelengths.
    The SpecInv architecture in Eqn (8) outputs a single-channel spatial map and never defines per-wavelength parameters, so this entity is disconnected from the network.

how reviews work

0 comments
Cite this review

Pith. "Pith review of $\varphi$-Adapt: A Physics-Informed Adaptation Learning Approach to 2D Quantum Material Discovery." pith.science (2026). https://pith.science/paper/PBN22CCQ

@misc{pith2026250705184,
  author       = {Pith},
  title        = {Pith review of: $\varphi$-Adapt: A Physics-Informed Adaptation Learning Approach to 2D Quantum Material Discovery},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PBN22CCQ}},
  note         = {Machine review of arXiv:2507.05184}
}
abstract

Characterizing quantum flakes is a critical step in quantum hardware engineering because the quality of these flakes directly influences qubit performance. Although computer vision methods for identifying two-dimensional quantum flakes have emerged, they still face significant challenges in estimating flake thickness. These challenges include limited data, poor generalization, sensitivity to domain shifts, and a lack of physical interpretability. In this paper, we introduce one of the first Physics-informed Adaptation Learning approaches to overcome these obstacles. We focus on two main issues, i.e., data scarcity and generalization. First, we propose a new synthetic data generation framework that produces diverse quantum flake samples across various materials and configurations, reducing the need for time-consuming manual collection. Second, we present $\varphi$-Adapt, a physics-informed adaptation method that bridges the performance gap between models trained on synthetic data and those deployed in real-world settings. Experimental results show that our approach achieves state-of-the-art performance on multiple benchmarks, outperforming existing methods. Our proposed approach advances the integration of physics-based modeling and domain adaptation. It also addresses a critical gap in leveraging synthesized data for real-world 2D material analysis, offering impactful tools for deep learning and materials science communities.

Figures

Figures reproduced from arXiv: 2507.05184 by the authors.

Figure 1
Figure 1. (a) The 2D material identification tasks. (b) Due to the differences between experimental [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. The optical model of a 2D material system. [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. The proposed Physics-Informed Adaptation framework. [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: (a) The process of generating the synthesized 2D material dataset. (b) The synthesized [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

61 extracted references · 47 canonical work pages

  1. [1]

    A. F. Agarap. Deep learning using rectified linear units (relu). arXiv preprint arXiv:1803.08375, 2018

    arXiv 2018

  2. [2]

    L. Chen, W. Li, and D. Xu. Recognizing rgb images by learning from rgb-d data. InProceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 1418–1425, 2014

    work page 2014

  3. [3]

    Y . Chen, W. Li, and L. Van Gool. Road: Reality oriented adaptation for semantic segmentation of urban scenes. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 7892–7901, 2018

    work page 2018

  4. [4]

    Chen, W.-Y

    Y .-H. Chen, W.-Y . Chen, Y .-T. Chen, B.-C. Tsai, Y .-C. Frank Wang, and M. Sun. No more discrimination: Cross city adaptation of road scene segmenters. In Proceedings of the IEEE international conference on computer vision, pages 1992–2001, 2017

    work page 1992

  5. [5]

    Dendukuri, B

    A. Dendukuri, B. Keeling, A. Fereidouni, J. Burbridge, K. Luu, and H. Churchill. Defining quantum neural networks via quantum time evolution. arXiv preprint arXiv:1905.10912, 2019

    arXiv 1905

  6. [6]

    Dosovitskiy, L

    A. Dosovitskiy, L. Beyer, A. Kolesnikov, D. Weissenborn, X. Zhai, T. Unterthiner, M. Dehghani, M. Minderer, G. Heigold, S. Gelly, et al. An image is worth 16x16 words: Transformers for image recognition at scale. arXiv preprint arXiv:2010.11929, 2020

    arXiv 2010

  7. [7]

    E. Fix. Discriminatory analysis: nonparametric discrimination, consistency properties, vol- ume 1. USAF school of Aviation Medicine, 1985

    work page 1985

  8. [8]

    Ganin and V

    Y . Ganin and V . Lempitsky. Unsupervised domain adaptation by backpropagation. InInterna- tional conference on machine learning, pages 1180–1189. PMLR, 2015

    work page 2015

Show all 61 references
  1. [9]

    Grandvalet and Y

    Y . Grandvalet and Y . Bengio. Semi-supervised learning by entropy minimization.Advances in neural information processing systems, 17, 2004

    2004

  2. [10]

    B. Han, Y . Lin, Y . Yang, N. Mao, W. Li, H. Wang, K. Yasuda, X. Wang, V . Fatemi, L. Zhou, et al. Deep-learning-enabled fast optical identification and characterization of 2d materials. Advanced Materials, 32(29):2000953, 2020

    2020

  3. [11]

    Y . Hao, F. Forest, and O. Fink. Simplifying source-free domain adaptation for object detec- tion: Effective self-training strategies and performance insights. In European Conference on Computer Vision, pages 196–213. Springer, 2024

    2024

  4. [12]

    K. He, G. Gkioxari, P. Dollár, and R. Girshick. Mask r-cnn. In Proceedings of the IEEE international conference on computer vision, pages 2961–2969, 2017

    2017

  5. [13]

    K. He, X. Zhang, S. Ren, and J. Sun. Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 770– 778, 2016

    2016

  6. [14]

    M. A. Hearst, S. T. Dumais, E. Osuna, J. Platt, and B. Scholkopf. Support vector machines. IEEE Intelligent Systems and their applications, 13(4):18–28, 1998

    1998

  7. [15]

    Hoffman, S

    J. Hoffman, S. Gupta, and T. Darrell. Learning with side information through modality hallu- cination. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 826–834, 2016

    2016

  8. [16]

    Hoffman, E

    J. Hoffman, E. Tzeng, T. Park, J.-Y . Zhu, P. Isola, K. Saenko, A. Efros, and T. Darrell. Cycada: Cycle-consistent adversarial domain adaptation. In International conference on machine learning, pages 1989–1998. Pmlr, 2018. 10

    1989

  9. [17]

    Hoffman, D

    J. Hoffman, D. Wang, F. Yu, and T. Darrell. Fcns in the wild: Pixel-level adversarial and constraint-based adaptation. arXiv preprint arXiv:1612.02649, 2016

    2016 arXiv

  10. [18]

    W. Hong, Z. Wang, M. Yang, and J. Yuan. Conditional generative adversarial network for structured domain adaptation. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 1335–1344, 2018

    2018

  11. [19]

    Hwang, J

    U. Hwang, J. Lee, J. Shin, and S. Yoon. Sf(da)2: Source-free domain adaptation through the lens of data augmentation. arXiv preprint arXiv:2403.10834, 2024

    2024 arXiv

  12. [20]

    Ioffe and C

    S. Ioffe and C. Szegedy. Batch normalization: Accelerating deep network training by reducing internal covariate shift. In International conference on machine learning, pages 448–456. pmlr, 2015

    2015

  13. [21]

    James Singh, T

    K. James Singh, T. Ahmed, P. Gautam, A. S. Sadhu, D.-H. Lien, S.-C. Chen, Y .-L. Chueh, and H.-C. Kuo. Recent advances in two-dimensional quantum dots and their applications. Nanomaterials, 11(6):1549, 2021

    2021

  14. [22]

    D. P. Kingma and J. Ba. Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980, 2014

    2014 arXiv

  15. [23]

    Kirillov, E

    A. Kirillov, E. Mintun, N. Ravi, H. Mao, C. Rolland, L. Gustafson, T. Xiao, S. Whitehead, A. C. Berg, W.-Y . Lo, et al. Segment anything. InProceedings of the IEEE/CVF international conference on computer vision, pages 4015–4026, 2023

    2023

  16. [24]

    M. C. Lemme, D. Akinwande, C. Huyghebaert, and C. Stampfer. 2d materials for future heterogeneous electronics. Nature communications, 13(1):1392, 2022

    2022

  17. [25]

    W. Li, L. Niu, and D. Xu. Exploiting privileged information from web data for image catego- rization. In European Conference on Computer Vision, pages 437–452. Springer, 2014

    2014

  18. [26]

    X. Li, Q. Sun, Y . Liu, Q. Zhou, S. Zheng, T.-S. Chua, and B. Schiele. Learning to self-train for semi-supervised few-shot classification. Advances in neural information processing systems, 32, 2019

    2019

  19. [27]

    Y . Li, H. Mao, R. Girshick, and K. He. Exploring plain vision transformer backbones for object detection. In European conference on computer vision, pages 280–296. Springer, 2022

    2022

  20. [28]

    J. Liu, Y . Ding, M. Zeng, and L. Fu. Chemical insights into two-dimensional quantum materials. Matter, 5(7):2168–2189, 2022

    2022

  21. [29]

    M. Long, Y . Cao, J. Wang, and M. Jordan. Learning transferable features with deep adaptation networks. In International conference on machine learning, pages 97–105. PMLR, 2015

    2015

  22. [30]

    Lopez-Paz, L

    D. Lopez-Paz, L. Bottou, B. Schölkopf, and V . Vapnik. Unifying distillation and privileged information. arXiv preprint arXiv:1511.03643, 2015

    2015 arXiv

  23. [31]

    K. Luu, X. B. Nguyen, and H. Churchill. Automatically detecting false negative objects in 2d material detection data sets, Aug. 22 2024. US Patent App. 18/443,058

    2024

  24. [32]

    Masubuchi, E

    S. Masubuchi, E. Watanabe, Y . Seo, S. Okazaki, T. Sasagawa, K. Watanabe, T. Taniguchi, and T. Machida. Deep-learning-based image segmentation integrated with optical microscopy for automatically searching for two-dimensional materials. npj 2D Materials and Applications, 4(1):3, 2020

    2020

  25. [33]

    Mordan, N

    T. Mordan, N. Thome, G. Henaff, and M. Cord. Revisiting multi-task learning with rock: a deep residual auxiliary block for visual detection. Advances in neural information processing systems, 31, 2018

    2018

  26. [34]

    Murez, S

    Z. Murez, S. Kolouri, D. Kriegman, R. Ramamoorthi, and K. Kim. Image to image translation for domain adaptation. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 4500–4509, 2018. 11

    2018

  27. [35]

    Nguyen, T.-D

    P. Nguyen, T.-D. Truong, M. Huang, Y . Liang, N. Le, and K. Luu. Self-supervised domain adaptation in crowd counting. In 2022 IEEE international conference on image processing (ICIP), pages 2786–2790. IEEE, 2022

    2022

  28. [36]

    X. B. Nguyen, A. Bisht, B. Thompson, H. Churchill, K. Luu, and S. U. Khan. Two-dimensional quantum material identification via self-attention and soft-labeling in deep learning. IEEE Access, 2024

    2024

  29. [37]

    X. B. Nguyen, H. Churchill, K. Luu, and S. U. Khan. Quantum vision clustering. arXiv preprint arXiv:2309.09907, 2023

    2023 arXiv

  30. [38]

    I. C. on Illumination (CIE). Cie 1931 color matching functions. https://www.cie.co.at/, 1931

    1931

  31. [39]

    I. C. on Illumination (CIE). Cie standard illuminant d65. https://www.cie.co.at/, 1967

    1967

  32. [40]

    F. Pan, I. Shin, F. Rameau, S. Lee, and I. S. Kweon. Unsupervised intra-domain adaptation for semantic segmentation through self-supervision. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 3764–3773, 2020

    2020

  33. [41]

    Z. Qiu, Y . Zhang, H. Lin, S. Niu, Y . Liu, Q. Du, and M. Tan. Source-free domain adaptation via avatar prototype generation and adaptation. arXiv preprint arXiv:2106.15326, 2021

    2021 arXiv

  34. [42]

    N. Ravi, V . Gabeur, Y .-T. Hu, R. Hu, C. Ryali, T. Ma, H. Khedr, R. Rädle, C. Rolland, L. Gustafson, et al. Sam 2: Segment anything in images and videos. arXiv preprint arXiv:2408.00714, 2024

    2024 arXiv

  35. [43]

    Ronneberger, P

    O. Ronneberger, P. Fischer, and T. Brox. U-net: Convolutional networks for biomedical image segmentation. In Medical image computing and computer-assisted intervention–MICCAI 2015: 18th international conference, Munich, Germany, October 5-9, 2015, proceedings, part III 18, pag...

    2015

  36. [44]

    Sarafianos, M

    N. Sarafianos, M. Vrigkas, and I. A. Kakadiaris. Adaptive svm+: Learning with privileged information for domain adaptation. In Proceedings of the IEEE international conference on computer vision workshops, pages 2637–2644, 2017

    2017

  37. [45]

    Sharmanska, N

    V . Sharmanska, N. Quadrianto, and C. H. Lampert. Learning to rank using privileged informa- tion. In Proceedings of the IEEE international conference on computer vision, pages 825–832, 2013

    2013

  38. [46]

    Simonyan and A

    K. Simonyan and A. Zisserman. Very deep convolutional networks for large-scale image recognition. arXiv preprint arXiv:1409.1556, 2014

    2014 arXiv

  39. [47]

    J. T. Springenberg. Unsupervised and semi-supervised learning with categorical generative adversarial networks. arXiv preprint arXiv:1511.06390, 2015

    2015 arXiv

  40. [48]

    T. Sun, C. Lu, and H. Ling. Domain adaptation with adversarial training on penultimate activations. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 37, pages 9935–9943, 2023

    2023

  41. [49]

    Truong, C

    T.-D. Truong, C. N. Duong, P. Helton, A. Dowling, X. Li, and K. Luu. Comal: Conditional maximum likelihood approach to self-supervised domain adaptation in long-tail semantic segmentation. arXiv preprint arXiv:2304.07372, 2023

    2023 arXiv

  42. [50]

    Truong, C

    T.-D. Truong, C. N. Duong, N. Le, S. L. Phung, C. Rainwater, and K. Luu. Bimal: Bijective maximum likelihood approach to domain adaptation in semantic scene segmentation. In Proceedings of the ieee/cvf international conference on computer vision , pages 8548–8557, 2021

    2021

  43. [51]

    Truong, P

    T.-D. Truong, P. Helton, A. Moustafa, J. D. Cothren, and K. Luu. Conda: Continual unsupervised domain adaptation learning in visual perception for self-driving cars. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 5642–5650, 2024. 12

    2024

  44. [52]

    Truong, N

    T.-D. Truong, N. Le, B. Raj, J. Cothren, and K. Luu. Fredom: Fairness domain adaptation approach to semantic scene understanding. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 19988–19997, 2023

    2023

  45. [53]

    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 Proceedings of the IEEE conference on computer vision and pattern recognition, pages 7472–7481, 2018

    2018

  46. [54]

    Tzeng, J

    E. Tzeng, J. Hoffman, K. Saenko, and T. Darrell. Adversarial discriminative domain adaptation. In Proceedings of the IEEE conference on computer vision and pattern recognition , pages 7167–7176, 2017

    2017

  47. [55]

    J.-L. Uslu, T. Ouaj, D. Tebbe, A. Nekrasov, J. H. Bertram, M. Schütte, K. Watanabe, T. Taniguchi, B. Beschoten, L. Waldecker, et al. An open-source robust machine learning platform for real-time detection and classification of 2d material flakes. Machine Learning: Science and ...

    2024

  48. [56]

    Vapnik and A

    V . Vapnik and A. Vashist. A new learning paradigm: Learning using privileged information. Neural networks, 22(5-6):544–557, 2009

    2009

  49. [57]

    V . VS, P. Oza, and V . M. Patel. Instance relation graph guided source-free domain adaptive object detection. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 3520–3530, 2023

    2023

  50. [58]

    T.-H. Vu, H. Jain, M. Bucher, M. Cord, and P. Pérez. Advent: Adversarial entropy minimization for domain adaptation in semantic segmentation. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 2517–2526, 2019

    2019

  51. [59]

    X. Wei, T. Bai, Z. Duan, M. Zhao, C. Zhao, Y . Lu, and D. Hu. Entropy-minimization mean teacher for source-free domain adaptive object detection. In International Conference on Neural Information Processing, pages 513–524. Springer, 2022

    2022

  52. [60]

    J.-Y . Zhu, T. Park, P. Isola, and A. A. Efros. Unpaired image-to-image translation using cycle-consistent adversarial networks. In Proceedings of the IEEE international conference on computer vision, pages 2223–2232, 2017

    2017

  53. [61]

    Y . Zou, Z. Yu, B. Kumar, and J. Wang. Unsupervised domain adaptation for semantic segmenta- tion via class-balanced self-training. In Proceedings of the European conference on computer vision (ECCV), pages 289–305, 2018. 13

    2018

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.