REVIEW 3 major objections 4 minor 295 references
NAE: Normalizing AutoEncoder
T0 review · 3 major / 4 minor · reviewed 2026-08-16 · deepseek-v4-flash
Pith's one-line read Flow autoencoders should follow reconstruction-loss gradients—NAE does this and sets new state-of-the-art results on molecular, tabular, and image benchmarks.
desk verdict Strong empirical paper with a heuristic theory; the conditional loss works, but the proof of alignment doesn't survive contact with off-diagonal terms. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The load-bearing object is the composite Jacobian $M = G_\phi J_\theta$ in data space and its latent-space counterpart $N = J_\theta G_\phi$. The reconstruction loss under isotropic noise is shown to approximately equal the squared deviation $\|I_d - N\|_F^2$, and its gradient is expanded in the scaled orthonormal Hutchinson probe basis $V$. The machinery is the scalar prefactor $c_{ii} = v_i^\top(JG - I_d)v_i$ for each probe: it measures whether the composite map is locally expansive ($c_{ii} > 0$) or contractive ($c_{ii} < 0$) along that probe direction, and it determines which of the two surrogate terms, $-v_i^\top J\,SG(Gv_i)$ or $SG(v_i^\top J)Gv_i$, is aligned with the reconstruction-loss gradient. Algorithm 1 uses this sign information to set a per-probe mask and build the conditional surrogate loss.
What would settle it
On a trained NAE, compute the full probe expansion of $\nabla\|I_d - JG\|_F^2$ on held-out data and compare the diagonal sum to the off-diagonal sum, or compute the cosine similarity between the conditional-loss gradient and the exact reconstruction-loss gradient on random batches; a frequently negative similarity or a large off-diagonal contribution would falsify the proposed mechanism.
Extended reading notes
Core claim
On the paper's own terms, the central discovery is that the reconstruction loss already encodes the correct training signal for both networks: under isotropic input perturbations, minimizing reconstruction loss is approximately minimizing $\|I_d - JG\|_F^2$, and the gradient of this term decomposes over probe vectors into the encoder and decoder surrogate gradients scaled by $c_{ii} = v_i^\top(JG - I_d)v_i$. When $c_{ii} > 0$ the decoder surrogate aligns with the reconstruction gradient and the encoder surrogate opposes it; when $c_{ii} < 0$ the roles reverse. Existing methods commit to one surrogate and therefore systematically fight the reconstruction objective in a subset of directions. NAE instead computes both surrogates per probe and, with the mask $\mathbb{1}[\mathrm{surr}_D > s_k]$, keeps only the term whose gradient is aligned. The paper reports that this conditional loss reaches the approximate-inverse regime at lower reconstruction weight $\beta$, stabilizes training, and achieves state-of-the-art negative log-likelihood on DW4, LJ13, LJ55, and QM9, best FID-like scores on Power, HEPMASS, and MiniBooNE, and best FID on CelebA in most settings.
Load-bearing premise
The conditional mask in Algorithm 1 is valid only if the diagonal terms of the probe expansion dominate and the sign of $\mathrm{surr}_D - s_k$ correctly identifies gradient alignment; otherwise the selected surrogate can oppose the true reconstruction gradient and the method's motivation collapses.
Editorial extensions
If this is right
- Training with the conditional loss instead of the encoder surrogate alone should lower negative log-likelihood on any flow-autoencoder setup, not just the benchmarks reported, because it removes directions in which the surrogate opposes reconstruction.
- The reconstruction weight $\beta$ needed to reach the approximate-inverse regime is smaller with NAE, reducing sensitivity to this hyperparameter.
- Because NAE and Free-Form Flows share architecture and inference procedure, the improved likelihoods on DW4, LJ13, LJ55, and QM9 come at no additional sampling cost.
- The near-equal FID between standard-normal and GMM samplers on CelebA suggests the latent space learned by NAE is regular enough to sample with a plain Gaussian prior.
- SiLU activations, whose second derivative is nonzero, preserve second-order surrogate gradients and further improve results; this points to activation choice as a first-class component of Jacobian-based training.
Reading between the lines
- The diagonal-dominance assumption in Section 3.4 could be checked directly by measuring the magnitude of off-diagonal terms $v_i^\top(N - I_d)v_j \cdot v_i^\top(\nabla N)v_j$; if they are not negligible, a weighted combination of both surrogates with coefficient $c_{ii}$ might outperform the hard mask.
- The same alignment argument transfers to any approximate-inverse training objective beyond flows, such as regularized autoencoders or learned forward/inverse maps in inverse problems, wherever a reconstruction loss and a surrogate likelihood-like term coexist.
- The empirical finding that larger latent dimensions need larger $\beta$ suggests an automatic scheduling rule: scale $\beta$ with the latent dimension or with the intrinsic dimensionality of the data instead of tuning it per dataset.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper studies 'flow autoencoders,' i.e., autoencoder architectures with separately parameterized encoder and decoder trained as approximate inverses, and proposes the Normalizing Autoencoder (NAE). The central contribution is a conditional surrogate loss that, per Hutchinson probe, selects either the encoder surrogate or the decoder surrogate based on a scalar mask intended to align the chosen surrogate gradient with the reconstruction-loss gradient. The authors first derive a relation between the reconstruction loss and the latent-space Jacobian residual, then use a probe-basis expansion to justify the mask, and finally report extensive experiments on molecular generation (DW4, LJ13, LJ55, QM9), tabular benchmarks (Power, Gas, HEPMASS, MiniBooNE), and image benchmarks (CelebA, MNIST, CIFAR-10). NAE achieves state-of-the-art or competitive results on most benchmarks, often with substantially better likelihoods or FID scores than prior flow-autoencoder baselines.
Significance. If the empirical results hold, NAE is a practically strong generative model: it improves over prior flow autoencoders on standard benchmarks while inheriting their architectural flexibility and sampling speed, and it does so with a minimal modification to the training objective. The paper is well positioned in the literature and includes a useful ablation study on the reconstruction weight and latent dimension, as well as honest discussion of limitations. The theoretical framing is less established: the central derivation in Section 3.4 rests on a diagonal truncation of a probe expansion and on a single-probe estimator whose justification is not supplied. Because the empirical gains are claimed to follow from this alignment mechanism, the gap is load-bearing rather than cosmetic. A revision that either proves the needed diagonal dominance, supplies empirical diagnostics for it, or appropriately reframes the theoretical claims would make the paper's contribution more solid.
major comments (3)
- [§3.4, Eq. (15) and Algorithm 1] The derivation expands the reconstruction-loss gradient as a double sum over probe indices i,j: Σ_i Σ_j (v_i^T(N−I_d)v_j)(v_i^T(∇N)v_j), then truncates to the equal-index terms with 'Finally, focusing on the term with equal indices' before defining c_ii and the conditional mask. This truncation is load-bearing: Algorithm 1 line 10 sets mask = 1[surrD > s_k], which with K=1 is exactly sign(c_kk), and the selection rule is only justified if the off-diagonal products are negligible. No argument or diagnostic for this diagonal dominance is provided; for d>1 the off-diagonal terms outnumber the diagonal ones. Moreover, all experiments use a single probe (K=1), so the completeness relation VV^T = d I_d used in the expansion does not hold for the realized estimator; a single random v satisfies E[vv^T] = I only in expectation, not for the actual sample. The central claim that the conditional loss aligns the surrogate gradient with the reconstruction gradient is therefore not established by the derivation as written.
- [Abstract and §3.4] The paper states in the abstract that it 'prove[s] that the proposed loss used by existing approaches is suboptimal,' but the argument in Section 3.4 is an approximate, heuristic derivation resting on the diagonal truncation and on the small-perturbation expansion of Eq. (6). No formal theorem is stated, and no verified set of conditions identifies when one surrogate is suboptimal. Please either replace 'prove' with a weaker claim clearly labeled as an analysis/heuristic, or supply a precise theorem with assumptions and proof.
- [§3.3, Eq. (13)] The equivalence between minimizing reconstruction loss in data space and minimizing ||I_d − N||_F^2 in latent space is approximate and inherits the bound in Eq. (11), which depends on δ = ||E_D||_F + ||E_d||_F. During training, before the approximate-inverse regime is reached, δ need not be small, and no empirical check of its magnitude is reported. Because this equivalence is the bridge to Eq. (14) and hence to the conditional-loss derivation, the paper should specify the regime in which the approximation is valid and, ideally, verify it experimentally (e.g., by reporting δ over the course of training).
minor comments (4)
- [Title page] The affiliation 'University of Hildeshiem' appears to be misspelled; it should be 'Hildesheim'.
- [Algorithm 1] The mask notation 1[condition] is used in line 10 but is not defined in the main text; please define it explicitly as an indicator function.
- [Figure 3 caption] The caption reads 'dim 1 and 2 corresponds to Two-moons'; the verb should agree with the plural subject ('correspond').
- [Tables 8 and 11] The SiLU-based results are marked as 'not considered for direct comparison,' which is clear, but the main text could state more explicitly which of these entries form the basis for the headline claims, since some tables include both ReLU and SiLU NAE results.
Circularity Check
One by-construction alignment step in the toy comparison; the headline SOTA results are externally validated and not load-bearing for circularity.
-
self definitional
[Section 3.4 (Eq. 15 and Algorithm 1, line 10) with Section 4.1 bullet]
"Since the proposed Conditional loss is aligned with the reconstruction objective by construction, compared to the other methods: – it requires a significantly smaller β to reach the approximate-inverse regime; – it achieves lower reconstruction error and nll within this regime."
Algorithm 1 sets mask←1[surrD > s_k], which is exactly sign(v_k^T(JG−I_d)v_k)=sign(c_kk), the scalar prefactor of the i=j term in the paper's expansion of ∇Lrec (Eq. 15). Thus the conditional loss's "alignment with the reconstruction objective" is true by definition of the mask, and the toy-dataset improvements (smaller β to enter the approximate-inverse regime, lower reconstruction error) are expected consequences of optimizing that same diagonal part of the objective, not an independent empirical confirmation. This is an explicitly acknowledged by-construction element and is not load-bearing for the external SOTA nll/FID results.
full rationale
Most of the derivation is not circular. Equations (14)–(15) derive a connection between the reconstruction-loss gradient and the surrogate terms from a Taylor expansion; this is an assumption-laden approximation, not a feedback of fitted values. Algorithm 1 implements the mask via sign(v_k^T(JG−I)v_k), and the abstract's 'proof of suboptimality' follows from the diagonal truncation, so it is better judged as a correctness risk: the off-diagonal i≠j terms are dropped without a diagnostic, and with K=1 the sign of a single diagonal coefficient need not determine alignment with the full ∇Lrec. The headline numbers on DW4, LJ13, LJ55, QM9, Power, HEPMASS, MiniBooNE, and CelebA are external metrics (nll/FID) that are not used to fit the method, and no target quantity is fed back into the derivation. The paper's citations to Sorrenson et al. and Draxler et al. are to prior work by other authors whose surrogate estimator is an input assumption, not a borrowed conclusion, so there is no load-bearing self-citation chain. The only mild circularity is the explicitly acknowledged 'by construction' alignment in the toy comparison, which does not drive the independently benchmarked SOTA claims.
Assumptions & free parameters
free parameters (3)
- Reconstruction weight beta =
DW4=10, LJ13=200, LJ55=500, QM9=2000, Power=10, Gas=10, HEPMASS=10, MiniBooNE=5, CelebA searched in {2,5,10,20,25}
- Perturbation noise scale sigma =
0.01 for tabular data (Appendix D.2); unspecified for other settings
- Latent dimension d =
Power=3, Gas=2, HEPMASS=10, MiniBooNE=21, image models 16/64/256 depending on dataset
assumptions (5)
- domain assumption Local first-order Taylor expansion of the autoencoder under small input perturbations (Eq. 6).
- domain assumption Isotropic Gaussian perturbation covariance Sigma_i = sigma^2 I_D (Eq. 8).
- domain assumption Approximate-inverse regime: N approx I_d and M approx rank-d orthogonal projector with small residual delta (Eq. 11, Eq. 13).
- ad hoc to paper Diagonal dominance: the probe expansion of the reconstruction gradient can be truncated to i=j terms.
- standard math Hutchinson trace estimator properties (E[vv^T]=I_d, v^T v=d) are standard.
Cite this review
Pith. "Pith review of NAE: Normalizing AutoEncoder." pith.science (2026). https://pith.science/paper/EPCK7RTN
@misc{pith2026260812084,
author = {Pith},
title = {Pith review of: NAE: Normalizing AutoEncoder},
year = {2026},
howpublished = {\url{https://pith.science/paper/EPCK7RTN}},
note = {Machine review of arXiv:2608.12084}
}
abstract
We consider the setting of Normalizing flows with approximate inverses, an established paradigm spanning both full-dimensional ($d=D$) and bottleneck ($d<D$) settings, and group these models under the term flow autoencoders. We present a theoretical investigation into their training dynamics and prove that the proposed loss used by existing approaches is suboptimal; specifically, both encoder and decoder surrogates must be optimized in alignment with reconstruction loss. Guided by these insights, we propose Normalizing Autoencoder (NAE), which employs a novel conditional loss that aligns the surrogate loss gradient with that of reconstruction loss, directly improving upon the current standard. Extensive experiments across molecule generation, tabular data, and image benchmarks demonstrate that NAE achieves state of the art performance. Our work highlights the importance of loss alignment in flow autoencoders and establishes NAE as a powerful generative framework.
Figures
Figures from the paper (5 more)
Reference graph
Works this paper leans on
-
[1]
LeCun, Yann , pages =. A
-
[2]
Bassey, Joshua and Qian, Lijun and Li, Xianfang , year = 2021, month = jan, number =. A. arXiv , langid =:2101.12249 , primaryclass =
arXiv 2021
-
[3]
Jaiswal, Ashish and Babu, Ashwin Ramesh and Zadeh, Mohammad Zaki and Banerjee, Debapriya and Makedon, Fillia , year = 2020, month = dec, journal =. A. doi:10.3390/technologies9010002 , urldate =
- [4]
-
[5]
Shi, Yuge and Siddharth, N. and Torr, Philip H. S. and Kosiorek, Adam R. , year = 2022, month = jul, number =. Adversarial. arXiv , langid =:2201.13100 , primaryclass =
arXiv 2022
-
[6]
Zhai, Shuangfei and Talbott, Walter and Srivastava, Nitish and Huang, Chen and Goh, Hanlin and Zhang, Ruixiang and Susskind, Josh , year = 2021, month = sep, number =. An. arXiv , langid =:2105.14103 , primaryclass =
arXiv 2021
-
[7]
Yin, Hongxu and Vahdat, Arash and Alvarez, Jose M and Mallya, Arun and Kautz, Jan and Molchanov, Pavlo , pages =. A-
-
[8]
Beyer, Lucas and Zhai, Xiaohua and Kolesnikov, Alexander , year = 2022, month = may, number =. Better Plain. arXiv , langid =:2205.01580 , primaryclass =
arXiv 2022
Show all 295 references
-
[9]
Wang, Zeyu and Bai, Yutong and Zhou, Yuyin and Xie, Cihang , pages =. Can
-
[10]
and Chandrasekaran, Shivkumar , year = 2022, month = jun, number =
Madala, Vamshi C. and Chandrasekaran, Shivkumar , year = 2022, month = jun, number =. arXiv , langid =:2205.10760 , primaryclass =
2022 arXiv
-
[11]
and Tan, Mingxing , year = 2021, month = sep, number =
Dai, Zihang and Liu, Hanxiao and Le, Quoc V. and Tan, Mingxing , year = 2021, month = sep, number =. arXiv , langid =:2106.04803 , primaryclass =
2021 arXiv
-
[12]
arXiv , langid =:2205.03892 , primaryclass =
Gao, Peng and Ma, Teli and Li, Hongsheng and Lin, Ziyi and Dai, Jifeng and Qiao, Yu , year = 2022, month = may, number =. arXiv , langid =:2205.03892 , primaryclass =
2022 arXiv
-
[13]
Data2vec:
Baevski, Alexei and Hsu, Wei-Ning and Xu, Qiantong and Babu, Arun and Gu, Jiatao and Auli, Michael , year = 2022, month = apr, number =. Data2vec:. arXiv , langid =:2202.03555 , primaryclass =
2022 arXiv
-
[14]
Ronen, Meitar and Finder, Shahaf E and Freifeld, Oren , pages =
-
[15]
Touvron, Hugo and Cord, Matthieu and Jegou, Herve , pages =
-
[16]
arXiv , langid =:2206.00386 , primaryclass =
Shi, Jie and Wu, Chenfei and Liang, Jian and Liu, Xiang and Duan, Nan , year = 2022, month = jun, number =. arXiv , langid =:2206.00386 , primaryclass =
2022 arXiv
-
[17]
Pan, Junting and Bulat, Adrian and Tan, Fuwen and Zhu, Xiatian and Dudziak, Lukasz and Li, Hongsheng and Tzimiropoulos, Georgios and Martinez, Brais , pages =
-
[18]
Efficient
Chen, Jun and Hu, Ming and Li, Boyang and Elhoseiny, Mohamed , year = 2022, month = jun, number =. Efficient. arXiv , langid =:2206.00790 , primaryclass =
2022 arXiv
-
[19]
Efficient
Liu, Yahui and Sangineto, Enver and Bi, Wei , pages =. Efficient
-
[20]
arXiv , langid =:2206.01191 , primaryclass =
Li, Yanyu and Yuan, Geng and Wen, Yang and Hu, Eric and Evangelidis, Georgios and Tulyakov, Sergey and Wang, Yanzhi and Ren, Jian , year = 2022, month = jul, number =. arXiv , langid =:2206.01191 , primaryclass =
2022 arXiv
-
[21]
Epistemic
Osband, Ian and Wen, Zheng and Asghari, Seyed Mohammad and Dwaracherla, Vikranth and Ibrahimi, Morteza and Lu, Xiuyuan and Van Roy, Benjamin , year = 2022, month = jul, number =. Epistemic. arXiv , langid =:2107.08924 , primaryclass =
2022 arXiv
-
[22]
Hong, Jiuk and Lee, Chaehyeon and Bang, Soyoun and Jung, Heechul , year = 2022, month = jun, number =. Fair. arXiv , langid =:2206.00244 , primaryclass =
2022 arXiv
-
[23]
Hatamizadeh, Ali and Yin, Hongxu and Kautz, Jan and Molchanov, Pavlo , year = 2022, month = jun, number =. Global. arXiv , langid =:2206.09959 , primaryclass =
2022 arXiv
-
[24]
Grokking:
Power, Alethea and Burda, Yuri and Edwards, Harri and Babuschkin, Igor and Misra, Vedant , year = 2022, month = jan, number =. Grokking:. arXiv , langid =:2201.02177 , primaryclass =
2022 arXiv
-
[25]
How to Train Your
Steiner, Andreas and Kolesnikov, Alexander and Zhai, Xiaohua and Wightman, Ross and Uszkoreit, Jakob and Beyer, Lucas , pages =. How to Train Your
-
[26]
Improved
Tang, Zhicong and Gu, Shuyang and Bao, Jianmin and Chen, Dong and Wen, Fang , year = 2022, month = may, number =. Improved. arXiv , langid =:2205.16007 , primaryclass =
2022 arXiv
-
[27]
and Xiong, Caiming , year = 2022, month = mar, number =
Liu, Zhiwei and Chen, Yongjun and Li, Jia and Luo, Man and Yu, Philip S. and Xiong, Caiming , year = 2022, month = mar, number =. Improving. arXiv , langid =:2203.15508 , primaryclass =
2022 arXiv
-
[28]
Improving the
Sariyildiz, Mert Bulent and Kalantidis, Yannis and Alahari, Karteek and Larlus, Diane , year = 2022, month = jun, number =. Improving the. arXiv , langid =:2206.15369 , primaryclass =
2022 arXiv
-
[29]
Information
Kimura, Masanari and Hino, Hideitsu , year = 2022, month = jun, number =. Information. arXiv , langid =:2206.10936 , primaryclass =
2022 arXiv
-
[30]
arXiv , langid =:2201.08239 , primaryclass =
Thoppilan, Romal and De Freitas, Daniel and Hall, Jamie and Shazeer, Noam and Kulshreshtha, Apoorv and Cheng, Heng-Tze and Jin, Alicia and Bos, Taylor and Baker, Leslie and Du, Yu and Li, YaGuang and Lee, Hongrae and Zheng, Huaixiu Steven and Ghafouri, Amin and Menegali, Marce...
-
[31]
arXiv , langid =:2206.11894 , primaryclass =
Gupta, Agrim and Tian, Stephen and Zhang, Yunzhi and Wu, Jiajun and. arXiv , langid =:2206.11894 , primaryclass =
-
[32]
Kothapalli, Vignesh and Rasromani, Ebrahim and Awatramani, Vasudev , year = 2022, month = jun, number =. Neural. arXiv , langid =:2206.04041 , primaryclass =
2022 arXiv
-
[33]
Ross, Brendan Leigh and. Neural. arXiv , langid =:2206.11267 , primaryclass =
-
[34]
, year = 2022, month = jan, number =
Li, Zengyi and Chen, Yubei and LeCun, Yann and Sommer, Friedrich T. , year = 2022, month = jan, number =. Neural. arXiv , langid =:2201.10000 , primaryclass =
2022 arXiv
-
[35]
Kidger, Patrick , year = 2022, month = feb, number =. On. arXiv , langid =:2202.02435 , primaryclass =
2022 arXiv
-
[36]
, year = 2022, month = jun, number =
Kaur, Simran and Cohen, Jeremy and Lipton, Zachary C. , year = 2022, month = jun, number =. On the. arXiv , langid =:2206.10654 , primaryclass =
2022 arXiv
-
[37]
Chollet, Fran. On the. arXiv , langid =:1911.01547 , primaryclass =
1911 arXiv
-
[38]
Chhabra, Sachin and Dutta, Prabal Bijoy and Venkateswara, Hemanth and Li, Baoxin , year = 2022, month = jul, pages =
2022
-
[39]
Sara and Lopes, Rapha Gontijo and Salimans, Tim and Ho, Jonathan and Fleet, David J
Saharia, Chitwan and Chan, William and Saxena, Saurabh and Li, Lala and Whang, Jay and Denton, Emily and Ghasemipour, Seyed Kamyar Seyed and Ayan, Burcu Karagol and Mahdavi, S. Sara and Lopes, Rapha Gontijo and Salimans, Tim and Ho, Jonathan and Fleet, David J. and Norouzi, Mo...
2022 arXiv
-
[40]
arXiv , langid =:2204.11167 , primaryclass =
Ma, Xiaojian and Nie, Weili and Yu, Zhiding and Jiang, Huaizu and Xiao, Chaowei and Zhu, Yuke and Zhu, Song-Chun and Anandkumar, Anima , year = 2022, month = jun, number =. arXiv , langid =:2204.11167 , primaryclass =
2022 arXiv
-
[41]
Revitalizing
Ge, Chongjian and Liang, Youwei and Song, Yibing and Jiao, Jianbo and Wang, Jue and Luo, Ping , pages =. Revitalizing
-
[42]
Azizi, Shekoofeh and Culp, Laura and Freyberg, Jan and Mustafa, Basil and Baur, Sebastien and Kornblith, Simon and Chen, Ting and MacWilliams, Patricia and Mahdavi, S. Sara and Wulczyn, Ellery and Babenko, Boris and Wilson, Megan and Loh, Aaron and Chen, Po-Hsuan Cameron and L...
2022 arXiv
-
[43]
Yu, Jiahui and Xu, Yuanzhong and Koh, Jing Yu and Luong, Thang and Baid, Gunjan and Wang, Zirui and Vasudevan, Vijay and Ku, Alexander and Yang, Yinfei and Ayan, Burcu Karagol and Hutchinson, Ben and Han, Wei and Parekh, Zarana and Li, Xin and Zhang, Han and Baldridge, Jason a...
2022 arXiv
-
[44]
Scaling-up
Su, Jiahao and Byeon, Wonmin and Huang, Furong , year = 2021, month = jun, number =. Scaling-up. arXiv , langid =:2106.09121 , primaryclass =
2021 arXiv
-
[45]
Self-Supervised
Liu, Xiao and Zhang, Fanjin and Hou, Zhenyu and Wang, Zhaoyu and Mian, Li and Zhang, Jing and Tang, Jie , year = 2021, journal =. Self-Supervised. doi:10.1109/TKDE.2021.3090866 , urldate =. arXiv , langid =:2006.08218 , primaryclass =
2021
-
[46]
Rabarisoa, Jaonary and Belissen, Valentin and Chabot, Florian and Pham, Quoc-Cuong , year = 2022, month = jun, number =. Self-. arXiv , langid =:2205.15173 , primaryclass =
2022 arXiv
-
[47]
arXiv , langid =:2104.03602 , primaryclass =
Atito, Sara and Awais, Muhammad and Kittler, Josef , year = 2021, month = nov, number =. arXiv , langid =:2104.03602 , primaryclass =
2021 arXiv
-
[48]
Ridnik, Tal and Lawen, Hussam and. Solving. arXiv , langid =:2204.03475 , primaryclass =
-
[49]
Survey on
Ali, Muhammad and Hashim, Sayed , year = 2022, month = feb, number =. Survey on. arXiv , langid =:2202.08514 , primaryclass =
2022 arXiv
-
[50]
Yang, Greg and Hu, Edward J. and Babuschkin, Igor and Sidor, Szymon and Liu, Xiaodong and Farhi, David and Ryder, Nick and Pachocki, Jakub and Chen, Weizhu and Gao, Jianfeng , year = 2022, month = mar, number =. Tensor. arXiv , langid =:2203.03466 , primaryclass =
2022 arXiv
-
[51]
Schwarzschild, Avi and Gupta, Arjun and Ghiasi, Amin and Goldblum, Micah and Goldstein, Tom , year = 2022, month = mar, number =. The. arXiv , langid =:2102.11011 , primaryclass =
2022 arXiv
-
[52]
and Karakulak, Serkan and Sobol, Vlad and Canziani, Alfredo and LeCun, Yann , year = 2022, month = jun, number =
Zhu, Jiachen and Moraes, Rafael M. and Karakulak, Serkan and Sobol, Vlad and Canziani, Alfredo and LeCun, Yann , year = 2022, month = jun, number =. arXiv , langid =:2206.10698 , primaryclass =
2022 arXiv
-
[53]
arXiv , langid =:2012.12877 , primaryclass =
Touvron, Hugo and Cord, Matthieu and Douze, Matthijs and Massa, Francisco and Sablayrolles, Alexandre and J. arXiv , langid =:2012.12877 , primaryclass =
2012 arXiv
-
[54]
arXiv , langid =:2205.02028 , primaryclass =
Duan, Haodong and Zhao, Nanxuan and Chen, Kai and Lin, Dahua , year = 2022, month = may, number =. arXiv , langid =:2205.02028 , primaryclass =
2022 arXiv
-
[55]
Goyal, Priya and Duval, Quentin and Seessel, Isaac and Caron, Mathilde and Misra, Ishan and Sagun, Levent and Joulin, Armand and Bojanowski, Piotr , year = 2022, month = feb, number =. Vision. arXiv , langid =:2202.08360 , primaryclass =
2022 arXiv
-
[56]
Where Are My
Camporese, Guglielmo and Izzo, Elena and Ballan, Lamberto , year = 2022, month = jun, number =. Where Are My. arXiv , langid =:2206.00481 , primaryclass =
2022 arXiv
-
[57]
Song, Jeonggeun and Lee, Heung-Chang , year = 2022, month = may, number =. X-. arXiv , langid =:2205.13805 , primaryclass =
2022 arXiv
-
[58]
Ramesh, Aditya and Pavlov, Mikhail and Goh, Gabriel and Gray, Scott and Voss, Chelsea and Radford, Alec and Chen, Mark and Sutskever, Ilya , year = 2021, month = feb, number =. Zero-. arXiv , langid =:2102.12092 , primaryclass =
2021 arXiv
- [59]
-
[60]
Paragraph-Based
Di Liello, Luca and Garg, Siddhant and Soldaini, Luca and Moschitti, Alessandro , year = 2022, month = jul, number =. Paragraph-Based. arXiv , file =:2205.01228 , primaryclass =
2022 arXiv
- [61]
-
[62]
Zhang, Xiao and Maire, Michael , year = 2020, month = dec, number =. Self-. arXiv , keywords =:2012.03044 , primaryclass =
2020 arXiv
- [63]
- [64]
- [65]
- [66]
-
[67]
International
Bao, Hangbo and Dong, Li and Piao, Songhao and Wei, Furu , year = 2022, month = may, urldate =. International
2022
-
[68]
Proceedings of the 37th
Chen, Mark and Radford, Alec and Child, Rewon and Wu, Jeff and Jun, Heewoo and Luan, David and Sutskever, Ilya , year = 2020, month = jul, series =. Proceedings of the 37th
2020
-
[69]
Conditional
Chu, Xiangxiang and Tian, Zhi and Zhang, Bo and Wang, Xinlong and Wei, Xiaolin and Xia, Huaxia and Shen, Chunhua , year = 2021, month = mar, number =. Conditional. arXiv , keywords =:2102.10882 , primaryclass =
2021 arXiv
- [70]
-
[71]
arXiv , keywords =:2106.09681 , primaryclass =
- [72]
- [73]
- [74]
- [75]
- [76]
- [77]
-
[78]
and Luong, Minh-Thang and Le, Quoc V
Trinh, Trieu H. and Luong, Minh-Thang and Le, Quoc V. , year = 2019, month = jul, number =. Selfie:. arXiv , keywords =:1906.02940 , primaryclass =
2019 arXiv
- [79]
- [80]
- [81]
- [82]
- [83]
- [84]
- [85]
-
[86]
doi:10.48550/arXiv.2106.05656 , urldate =
Li, Zhaowen and Chen, Zhiyang and Yang, Fan and Li, Wei and Zhu, Yousong and Zhao, Chaoyang and Deng, Rui and Wu, Liwei and Zhao, Rui and Tang, Ming and Wang, Jinqiao , year = 2021, month = oct, number =. doi:10.48550/arXiv.2106.05656 , urldate =. arXiv , keywords =:2106.05656...
- [87]
- [88]
-
[89]
doi:10.48550/arXiv.2111.12710 , urldate =
Dong, Xiaoyi and Bao, Jianmin and Zhang, Ting and Chen, Dongdong and Zhang, Weiming and Yuan, Lu and Chen, Dong and Wen, Fang and Yu, Nenghai , year = 2022, month = jan, number =. doi:10.48550/arXiv.2111.12710 , urldate =. arXiv , keywords =:2111.12710 , primaryclass =
-
[90]
and Yogatama, Dani and Metzler, Donald , year = 2022, month = jul, number =
Tay, Yi and Dehghani, Mostafa and Abnar, Samira and Chung, Hyung Won and Fedus, William and Rao, Jinfeng and Narang, Sharan and Tran, Vinh Q. and Yogatama, Dani and Metzler, Donald , year = 2022, month = jul, number =. Scaling. doi:10.48550/arXiv.2207.10551 , urldate =. arXiv ...
- [91]
- [92]
- [93]
-
[94]
Chen, Ting and Kornblith, Simon and Norouzi, Mohammad and Hinton, Geoffrey , year = 2020, month = jun, number =. A. arXiv , keywords =:2002.05709 , primaryclass =
2020 arXiv
-
[95]
and McAuley, Julian , year = 2020, month = jun, number =
Bachlechner, Thomas and Majumder, Bodhisattwa Prasad and Mao, Huanru Henry and Cottrell, Garrison W. and McAuley, Julian , year = 2020, month = jun, number =. doi:10.48550/arXiv.2003.04887 , urldate =. arXiv , keywords =:2003.04887 , primaryclass =
- [96]
- [97]
-
[98]
and Srinivas, Aravind and Lin, Tsung-Yi and Shlens, Jonathon and Zoph, Barret , year = 2021, month = mar, number =
Bello, Irwan and Fedus, William and Du, Xianzhi and Cubuk, Ekin D. and Srinivas, Aravind and Lin, Tsung-Yi and Shlens, Jonathon and Zoph, Barret , year = 2021, month = mar, number =. Revisiting. doi:10.48550/arXiv.2103.07579 , urldate =. arXiv , keywords =:2103.07579 , primaryclass =
-
[99]
Self-Supervised
Wu, Haiyan and Gao, Yuting and Li, Ke and Zhang, Yinqi and Lin, Shaohui and Xie, Yuan and Sun, Xing , year = 2021, month = nov, urldate =. Self-Supervised
2021
-
[100]
, year = 2022, month = apr, number =
Zhou, Daquan and Yu, Zhiding and Xie, Enze and Xiao, Chaowei and Anandkumar, Anima and Feng, Jiashi and Alvarez, Jose M. , year = 2022, month = apr, number =. Understanding. doi:10.48550/arXiv.2204.12451 , urldate =. arXiv , keywords =:2204.12451 , primaryclass =
- [101]
-
[102]
Learning
Sohn, Kihyuk and Lee, Honglak and Yan, Xinchen , year = 2015, volume =. Learning. Advances in
2015
-
[103]
Ruan, Bo-Kai and Shuai, Hong-Han and Cheng, Wen-Huang , year = 2022, month = jul, number =. Vision. arXiv , keywords =:2207.03041 , primaryclass =
2022 arXiv
-
[104]
Xiong, Ruibin and Yang, Yunchang and He, Di and Zheng, Kai and Zheng, Shuxin and Xing, Chen and Zhang, Huishuai and Lan, Yanyan and Wang, Liwei and Liu, Tie-Yan , year = 2020, month = jun, number =. On. arXiv , keywords =:2002.04745 , primaryclass =
2020 arXiv
-
[105]
Rethinking
Liu, Fenglin and Ren, Xuancheng and Zhang, Zhiyuan and Sun, Xu and Zou, Yuexian , year = 2021, month = may, number =. Rethinking. arXiv , keywords =:2105.07205 , primaryclass =
2021 arXiv
- [106]
- [107]
-
[108]
Advances in
Krizhevsky, Alex and Sutskever, Ilya and Hinton, Geoffrey E , year = 2012, volume =. Advances in
2012
-
[109]
and Kaiser, Lukasz and Polosukhin, Illia , year = 2017, month = dec, number =
Vaswani, Ashish and Shazeer, Noam and Parmar, Niki and Uszkoreit, Jakob and Jones, Llion and Gomez, Aidan N. and Kaiser, Lukasz and Polosukhin, Illia , year = 2017, month = dec, number =. Attention. doi:10.48550/arXiv.1706.03762 , urldate =. arXiv , keywords =:1706.03762 , pri...
-
[110]
Riesenhuber, Maximilian and Dayan, Peter , year = 1996, volume =. Neural. Advances in
1996
- [111]
- [112]
- [113]
- [114]
-
[115]
Li, Jiahao and Li, Bin and Lu, Yan , abstract =. Neural
- [116]
-
[117]
Xiang, Jinxi and Tian, Kuan and Zhang, Jun and Han, Xiao and Yang, Wei , year = 2023, abstract =
2023
-
[118]
Muckley, Matthew J. and. Improving. Proceedings of the 40th
- [119]
- [120]
-
[121]
and Salimans, Tim and Jozefowicz, Rafal and Chen, Xi and Sutskever, Ilya and Welling, Max , year = 2017, month = jan, number =
Kingma, Diederik P. and Salimans, Tim and Jozefowicz, Rafal and Chen, Xi and Sutskever, Ilya and Welling, Max , year = 2017, month = jan, number =. Improving. doi:10.48550/arXiv.1606.04934 , urldate =. arXiv , keywords =:1606.04934 , primaryclass =
- [122]
- [123]
- [124]
- [125]
-
[126]
Invertible
Finzi, Marc and Izmailov, Pavel and Maddox, Wesley and Kirichenko, Polina and Wilson, Andrew Gordon , year = 2019, abstract =. Invertible
2019
-
[127]
Invertible
Karami, Mahdi and Schuurmans, Dale and. Invertible. Advances in
- [128]
-
[129]
Kobyzev, Ivan and Prince, Simon J. D. and Brubaker, Marcus A. , year = 2021, month = nov, journal =. Normalizing. doi:10.1109/TPAMI.2020.2992934 , urldate =. arXiv , keywords =:1908.09257 , primaryclass =
2021
- [130]
- [131]
- [132]
-
[133]
Nature , volume =
Accurate Predictions on Small Data with a Tabular Foundation Model , author =. Nature , volume =. doi:10.1038/s41586-024-08328-6 , urldate =
- [134]
-
[135]
and Fan, Angela and Auli, Michael and Grangier, David , year = 2016, month = dec, journal =
Dauphin, Yann N. and Fan, Angela and Auli, Michael and Grangier, David , year = 2016, month = dec, journal =. Language
2016
- [136]
-
[137]
Everything
Nagel, Jory , year = 2024, month = sep, urldate =. Everything
2024
-
[138]
Survey on
Figueira, Alvaro and Vaz, Bruno , year = 2022, month = jan, journal =. Survey on. doi:10.3390/math10152733 , urldate =
2022 doi
- [139]
-
[140]
Bryce and Gagliano, Alexander and Yan, Ziang and Connolly, Andrew J
Crenshaw, John Franklin and Kalmbach, J. Bryce and Gagliano, Alexander and Yan, Ziang and Connolly, Andrew J. and Malz, Alex I. and Schmidt, Samuel J. and Collaboration, The LSST Dark Energy Science , year = 2024, month = aug, journal =. Probabilistic. doi:10.3847/1538-3881/ad...
2024 arXiv
-
[141]
Differentially
Lee, Jaewoo and Kim, Minjung and Jeong, Yonghyun and Ro, Youngmin , year = 2022, month = jun, journal =. Differentially. doi:10.1609/aaai.v36i7.20697 , urldate =
2022 doi
- [142]
- [143]
- [144]
- [145]
- [146]
- [147]
-
[148]
doi:10.48550/arXiv.2406.04619 , urldate =
Lin, Xiaofeng and Xu, Chenheng and Yang, Matthew and Cheng, Guang , year = 2024, month = jun, number =. doi:10.48550/arXiv.2406.04619 , urldate =. arXiv , keywords =:2406.04619 , primaryclass =
2024 doi
- [149]
-
[150]
Liu, Tennison and Qian, Zhaozhi and Berrevoets, Jeroen and van der Schaar, Mihaela , year = 2022, month = sep, urldate =. The
2022
- [151]
- [152]
- [153]
- [154]
- [155]
- [156]
-
[157]
Variational
Rezende, Danilo Jimenez and Mohamed, Shakir , year = 2015, month = may, journal =. Variational
2015
-
[158]
Normalizing
Zhai, Shuangfei and Zhang, Ruixiang and Nakkiran, Preetum and Berthelot, David and Gu, Jiatao and Zheng, Huangjie and Chen, Tianrong and Bautista, Miguel Angel and Jaitly, Navdeep and Susskind, Josh , year = 2024, month = dec, number =. Normalizing. doi:10.48550/arXiv.2412.063...
- [159]
- [160]
-
[161]
IEEE Transactions on Pattern Analysis and Machine Intelligence , volume =
Deep. IEEE Transactions on Pattern Analysis and Machine Intelligence , volume =. doi:10.1109/TPAMI.2021.3116668 , urldate =. arXiv , file =:2103.04922 , primaryclass =
2021
- [162]
- [163]
- [164]
-
[165]
Gu, Albert, Isaac Liao , year = 2025, month = mar, journal =
2025
-
[166]
Stoian, Mihaela C. A. doi:10.48550/arXiv.2503.05954 , urldate =. arXiv , keywords =:2503.05954 , primaryclass =
-
[167]
3rd International Conference on Learning Representations,
Dinh, Laurent and Krueger, David and Bengio, Yoshua , editor =. 3rd International Conference on Learning Representations,. arXiv , bibsource =:1410.8516 , primaryclass =
- [168]
- [169]
- [170]
- [171]
- [172]
- [173]
-
[174]
Generalized
Hagemann, Paul and Hertrich, Johannes and Steidl, Gabriele , year = 2023, month = feb, eprint =. Generalized. doi:10.1017/9781009331012 , urldate =
2023 doi
- [175]
-
[176]
doi:10.48550/arXiv.2506.06276 , urldate =
Gu, Jiatao and Chen, Tianrong and Berthelot, David and Zheng, Huangjie and Wang, Yuyang and Zhang, Ruixiang and Dinh, Laurent and Bautista, Miguel Angel and Susskind, Josh and Zhai, Shuangfei , year = 2025, month = jun, number =. doi:10.48550/arXiv.2506.06276 , urldate =. arXi...
- [177]
- [178]
- [179]
- [180]
- [181]
- [182]
- [183]
- [184]
- [185]
-
[186]
Raieli, Salvatore , urldate =. Tabular
- [187]
- [188]
-
[189]
Isomorphism,
Walton, Steven , year = 2023, url=. Isomorphism,
2023
- [190]
- [191]
- [192]
- [193]
- [194]
- [195]
- [196]
- [197]
- [198]
- [199]
- [200]
-
[201]
Free-Form
Draxler, Felix and Sorrenson, Peter and Zimmermann, Lea and Rousselot, Armand and K. Free-Form. Proceedings of
- [202]
- [203]
- [204]
- [205]
-
[206]
doi:10.48550/arXiv.2508.10133 , urldate =
Truong, Thanh-Dat and Bobda, Christophe and Agarwal, Nitin and Luu, Khoa , year = 2025, month = aug, number =. doi:10.48550/arXiv.2508.10133 , urldate =. arXiv , keywords =:2508.10133 , primaryclass =
2025 doi
-
[207]
doi:10.48550/arXiv.2508.16845 , urldate =
Tarasov, Denis and Nikulin, Alexander and Zisman, Ilya and Klepach, Albina and Lyubaykin, Nikita and Polubarov, Andrei and Derevyagin, Alexander and Kurenkov, Vladislav , year = 2025, month = aug, number =. doi:10.48550/arXiv.2508.16845 , urldate =. arXiv , keywords =:2508.168...
2025 doi
-
[208]
doi:10.48550/arXiv.2506.01158 , urldate =
Rehman, Danyal and Davis, Oscar and Lu, Jiarui and Tang, Jian and Bronstein, Michael and Bengio, Yoshua and Tong, Alexander and Bose, Avishek Joey , year = 2025, month = jun, number =. doi:10.48550/arXiv.2506.01158 , urldate =. arXiv , keywords =:2506.01158 , primaryclass =
2025 doi
- [209]
-
[210]
Manifold
Diepeveen, Willem and Needell, Deanna , year = 2025, month = jul, number =. Manifold. doi:10.48550/arXiv.2505.08087 , urldate =. arXiv , keywords =:2505.08087 , primaryclass =
2025 doi
-
[211]
, year = 2025, month = may, number =
Wang, Yihang and Chi, Chris and Dinner, Aaron R. , year = 2025, month = may, number =. Mitigating Mode Collapse in Normalizing Flows by Annealing with an Adaptive Schedule:. doi:10.48550/arXiv.2505.03652 , urldate =. arXiv , keywords =:2505.03652 , primaryclass =
- [212]
- [213]
-
[214]
Hutchinson
Hu, Zheyuan and Shi, Zekun and Karniadakis, George Em and Kawaguchi, Kenji , year = 2024, month = may, journal =. Hutchinson. doi:10.1016/j.cma.2024.116883 , urldate =. arXiv , keywords =:2312.14499 , primaryclass =
2024
-
[215]
Geometric
Krantz, Steven G and Parks, Harold R , year = 2008, month = dec, publisher =. Geometric. doi:10.1007/978-0-8176-4679-0 , isbn =
2008 doi
- [216]
- [217]
- [218]
- [219]
- [220]
- [221]
- [222]
- [223]
- [224]
- [225]
- [226]
-
[227]
Denoising
Horvat, Christian and Pfister, Jean-Pascal , year = 2021, volume =. Denoising. Advances in
2021
- [228]
-
[229]
Learning
Sorrenson, Peter and Draxler, Felix and Rousselot, Armand and Hummerich, Sander and K. Learning. Advances in. doi:10.52202/079017-2138 , urldate =
- [230]
- [231]
- [232]
-
[233]
and Deligiannidis, George and Doucet, Arnaud , year = 2021, month = apr, number =
Cornish, Rob and Caterini, Anthony L. and Deligiannidis, George and Doucet, Arnaud , year = 2021, month = apr, number =. Relaxing. doi:10.48550/arXiv.1909.13833 , urldate =. arXiv , keywords =:1909.13833 , primaryclass =
-
[234]
doi:10.48550/arXiv.2512.04084 , urldate =
Zhao, Qinyu and Zheng, Guangting and Yang, Tao and Zhu, Rui and Leng, Xingjian and Gould, Stephen and Zheng, Liang , year = 2025, month = dec, number =. doi:10.48550/arXiv.2512.04084 , urldate =. arXiv , keywords =:2512.04084 , primaryclass =
2025 doi
-
[235]
Bertrand, Quentin and Gagneux, Anne and Massias, Mathurin and Emonet, R. On the. doi:10.48550/arXiv.2506.03719 , urldate =. arXiv , keywords =:2506.03719 , primaryclass =
- [236]
-
[237]
and Zenke, Friedemann , year = 2026, month = may, urldate =
Mikulasch, Fabian A. and Zenke, Friedemann , year = 2026, month = may, urldate =. Understanding
2026
- [238]
- [239]
-
[240]
Marginal
Negri, Marcello Massimo and Aellen, Jonathan and Jahn, Manuel and Khorashadizadeh, AmirEhsan and Roth, Volker , year = 2025, month = sep, number =. Marginal. doi:10.48550/arXiv.2509.26221 , urldate =. arXiv , keywords =:2509.26221 , primaryclass =
2025 doi
-
[241]
Sorrenson, Peter , year = 2025, publisher =. Free-. doi:10.11588/heidok.00035980 , urldate =
2025 doi
- [242]
-
[243]
Equivariant
Hoogeboom, Emiel and Satorras, V. Equivariant. Proceedings of the 39th. arXiv , langid =:2203.17003 , primaryclass =
-
[244]
Adaptive
Bai, Zhiwei and Zhou, Zhangchen and Zhao, Jiajie and Li, Xiaolong and Li, Zhiyu and Xiong, Feiyu and Yang, Hongkang and Zhang, Yaoyu and Xu, Zhi-Qin John , year = 2026, month = may, number =. Adaptive. doi:10.48550/arXiv.2506.04805 , urldate =. arXiv , keywords =:2506.04805 , ...
- [245]
- [246]
- [247]
- [248]
- [249]
-
[250]
Analytical Chemistry , volume =
Skillful Writing of an Awful Research Paper , author =. Analytical Chemistry , volume =. doi:10.1021/ac2000169 , urldate =
- [251]
-
[252]
Scientific Data , volume =
Quantum Chemistry Structures and Properties of 134 Kilo Molecules , author =. Scientific Data , volume =. doi:10.1038/sdata.2014.22 , urldate =
2014 doi
-
[253]
Durkan, Conor and Bekasov, Artur and Murray, Iain and Papamakarios, George , year = 2019, volume =. Neural. Advances in
2019
-
[254]
Canonical Normalizing Flows for Manifold Learning , booktitle =
Flouris, Kyriakos and Konukoglu, Ender , year = 2023, volume =. Canonical Normalizing Flows for Manifold Learning , booktitle =. doi:10.52202/075280-1189 , urldate =
2023 doi
-
[255]
Flows for Simultaneous Manifold Learning and Density Estimation , booktitle =
Brehmer, Johann and Cranmer, Kyle , year = 2020, volume =. Flows for Simultaneous Manifold Learning and Density Estimation , booktitle =
2020
-
[256]
Invertible
Ahn, Byeongkeun and Kim, Chiyoon and Hong, Youngjoon and Kim, Hyunwoo , year = 2022, pages =. Invertible. Advances in. doi:10.52202/068431-1225 , urldate =
2022 doi
-
[257]
Behrmann, Jens and Grathwohl, Will and Chen, Ricky T. Q. and Duvenaud, David and Jacobsen, Joern-Henrik , year = 2019, month = may, pages =. Invertible. Proceedings of the 36th
2019
-
[258]
Chadebec, Cl. Pythae:. Advances in. doi:10.52202/068431-1568 , urldate =
-
[259]
Wildberger, Jonas and Dax, Maximilian and Buchholz, Simon and Green, Stephen and Macke, Jakob H and Sch. Flow. Advances in. doi:10.52202/075280-0737 , urldate =
-
[261]
Proceedings of the 18th
Kallappa, Aditya and Nagar, Sandeep and Varma, Girish , year = 2023, pages =. Proceedings of the 18th. doi:10.5220/0011876600003417 , urldate =
2023 doi
- [262]
-
[263]
Truong, Thanh-Dat and Duong, Chi Nhan and Tran, Minh-Triet and Le, Ngan and Luu, Khoa , year = 2021, month = jul, journal =. Fast. doi:10.3390/fi13070179 , urldate =
2021 doi
-
[264]
Rectangular
Caterini, Anthony L and. Rectangular. Advances in
-
[265]
Trumpets:
Kothari, Konik and Khorashadizadeh, AmirEhsan and de Hoop, Maarten and Dokmani. Trumpets:. Proceedings of the
-
[266]
Cao, Nicola De and Aziz, Wilker and Titov, Ivan , year = 2020, month = aug, pages =. Block. Proceedings of
2020
-
[267]
Papamakarios, George and Pavlakou, Theo and Murray, Iain , year = 2017, volume =. Masked. Advances in
2017
-
[268]
Invertible
Teng, Yunfei and Choromanska, Anna , year = 2019, month = mar, journal =. Invertible. doi:10.3390/computation7020020 , urldate =
2019 doi
-
[269]
Advances in
Midgley, Laurence and Stimper, Vincent and Antor. Advances in. doi:10.52202/075280-3466 , urldate =
-
[270]
Equivariant Flow Matching , booktitle =
Klein, Leon and Kr. Equivariant Flow Matching , booktitle =. doi:10.52202/075280-2617 , urldate =
-
[271]
Hong, Seongmin and Chun, Se Young , year = 2023, month = jun, journal =. Neural. doi:10.1609/aaai.v37i10.26441 , urldate =
2023 doi
-
[272]
Malnick, Shimon and Avidan, Shai and Fried, Ohad , year = 2024, month = jan, pages =. Taming. 2024. doi:10.1109/WACV57701.2024.00458 , urldate =
2024
-
[273]
Computer
Tschannen, Michael and Eastwood, Cian and Mentzer, Fabian , editor =. Computer. doi:10.1007/978-3-031-72998-0_17 , urldate =
-
[274]
Grover, Aditya and Dhar, Manik and Ermon, Stefano , year = 2018, month = apr, journal =. Flow-. doi:10.1609/aaai.v32i1.11829 , urldate =
2018 doi
-
[275]
doi:10.1109/ICASSP49357.2023.10095301 , urldate =
Mosi. doi:10.1109/ICASSP49357.2023.10095301 , urldate =
2023
-
[276]
Distilling
Walton, Steven and Klyukin, Valeriy and Artemev, Maksim and Derkach, Denis and Orlov, Nikita and Shi, Humphrey , year = 2025, month = jun, pages =. Distilling. 2025. doi:10.1109/CVPRW67362.2025.00316 , urldate =
2025
-
[277]
Feng, Shibo and Miao, Chunyan and Xu, Ke and Wu, Jiaxiang and Wu, Pengcheng and Zhang, Yang and Zhao, Peilin , year = 2024, month = may, journal =. Multi-. doi:10.1109/TKDE.2023.3319672 , urldate =
2024
-
[278]
Towards a
Hutter, Marcus , editor =. Towards a. Machine. doi:10.1007/3-540-44795-4_20 , urldate =
-
[279]
He, Kaiming and Zhang, Xiangyu and Ren, Shaoqing and Sun, Jian , year = 2016, month = jun, pages =. Deep. 2016. doi:10.1109/CVPR.2016.90 , urldate =
2016 doi
-
[280]
Hinton, Geoffrey , year = 2023, month = feb, journal =. How to. doi:10.1162/neco_a_01557 , urldate =
2023 doi
-
[281]
Long, Lin and Wang, Rui and Xiao, Ruixuan and Zhao, Junbo and Ding, Xiao and Chen, Gang and Wang, Haobo , year = 2024, pages =. On. Findings of the. doi:10.18653/v1/2024.findings-acl.658 , urldate =
2024 doi
-
[282]
Patch n'
Dehghani, Mostafa and Mustafa, Basil and Djolonga, Josip and Heek, Jonathan and Minderer, Matthias and Caron, Mathilde and Steiner, Andreas and Puigcerver, Joan and Geirhos, Robert and Alabdulmohsin, Ibrahim and Oliver, Avital and Padlewski, Piotr and Gritsenko, Alexey and Luc...
2023 doi
-
[283]
Yu, Weihao and Wang, Xinchao , year = 2025, month = jun, pages =. 2025. doi:10.1109/CVPR52734.2025.00423 , urldate =
2025
-
[284]
Computer
Zeng, Fangao and Dong, Bin and Zhang, Yuang and Wang, Tiancai and Zhang, Xiangyu and Wei, Yichen , editor =. Computer. doi:10.1007/978-3-031-19812-0_38 , urldate =
-
[285]
Xu, Mingle and Yoon, Sook and Fuentes, Alvaro and Park, Dong Sun , year = 2023, month = may, journal =. A. doi:10.1016/j.patcog.2023.109347 , urldate =
2023
-
[287]
doi:10.1088/1742-5468/ac9830 , urldate =
Journal of Statistical Mechanics: Theory and Experiment , volume =. doi:10.1088/1742-5468/ac9830 , urldate =
-
[289]
Russakovsky, Olga and Deng, Jia and Su, Hao and Krause, Jonathan and Satheesh, Sanjeev and Ma, Sean and Huang, Zhiheng and Karpathy, Andrej and Khosla, Aditya and Bernstein, Michael and Berg, Alexander C. and. International Journal of Computer Vision , volume =. doi:10.1007/s1...
-
[290]
Li, Yanghao and Wu, Chao-Yuan and Fan, Haoqi and Mangalam, Karttikeya and Xiong, Bo and Malik, Jitendra and Feichtenhofer, Christoph , year = 2022, month = jun, pages =. 2022. doi:10.1109/CVPR52688.2022.00476 , urldate =
2022
-
[292]
Squeeze-and-
Hu, Jie and Shen, Li and Sun, Gang , year = 2018, month = jun, pages =. Squeeze-and-. 2018. doi:10.1109/CVPR.2018.00745 , urldate =
2018
-
[293]
and Gal, Yarin , year = 2023, month = jun, pages =
Mukhoti, Jishnu and Kirsch, Andreas and Van Amersfoort, Joost and Torr, Philip H.S. and Gal, Yarin , year = 2023, month = jun, pages =. Deep. 2023. doi:10.1109/CVPR52729.2023.02336 , urldate =
2023
-
[294]
Learning
Gidaris, Spyros and Bursuc, Andrei and Komodakis, Nikos and Perez, Patrick and Cord, Matthieu , year = 2020, month = jun, pages =. Learning. 2020. doi:10.1109/CVPR42600.2020.00696 , urldate =
2020
-
[296]
Pinto, Francesco and Torr, Philip H. S. and K. Dokania, Puneet , editor =. An. Computer. doi:10.1007/978-3-031-19778-9_27 , urldate =
-
[297]
Hyperbolic
Peng, Wei and Varanka, Tuomas and Mostafa, Abdelrahman and Shi, Henglin and Zhao, Guoying , year = 2022, month = dec, journal =. Hyperbolic. doi:10.1109/TPAMI.2021.3136921 , urldate =
2022
-
[298]
, editor =
Liu, Nan and Li, Shuang and Du, Yilun and Torralba, Antonio and Tenenbaum, Joshua B. , editor =. Compositional. Computer. doi:10.1007/978-3-031-19790-1_26 , urldate =
-
[299]
Graham, Ben and. 2021. doi:10.1109/ICCV48922.2021.01204 , urldate =
2021
-
[300]
Inception
Si, Chenyang and Yu, Weihao and Zhou, Pan and Zhou, Yichen and Wang, Xinchao and Yan, Shuicheng , year = 2022, pages =. Inception. Advances in. doi:10.52202/068431-1707 , urldate =
2022 doi
Reviewed August 16, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.