Pith. sign in

REVIEW 5 major objections 4 minor 55 references

Robust Multiple Description Neural Video Codec with Masked Transformer for Dynamic and Noisy Networks

T0 review · 5 major / 4 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read NeuralMDC claims a bidirectional masked transformer that predicts missing latent tokens from the two previous frames yields 2-8x loss-resilience gains over Grace while cutting bitrate by 76.88%.

desk verdict NeuralMDC is the first neural MDC video codec with a neat masked-transformer design, but the loss resilience claims break on an unsolved entropy-coding synchronization problem. read the letter →

arxiv 2412.07922 v1 pith:MKOGFUEQ submitted 2024-12-10 cs.CV cs.AI

Xinyue Hu , Wei Ye , Jiaxiang Tang , Eman Ramadan , Zhi-Li Zhang This is my paper · ORCID
classification cs.CVcs.AI
keywords multipledescriptioncodingneuralvideocodecmaskedtransformerlossresiliencepacketlatentrepresentationsplittingentropy5Gstreaming
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 sets out to show that a bidirectional masked transformer can replace the entire motion-compensation and residual-coding pipeline of a video codec, and in doing so make multiple description coding practical. NeuralMDC tokenizes each frame with a CNN autoencoder, splits the latent tokens into several independently decodable descriptions, and uses the transformer to predict each token's distribution from the two previous frames. The same predicted distribution serves both to entropy code each description cheaply and to infer tokens lost in transmission. The paper claims this yields the best loss resilience of any tested codec, outperforming the strongest existing loss-resilient neural codec by 2 to 8 times in PSNR and MS-SSIM while using 76.88% fewer bits, which would let video streaming over noisy multi-path 5G networks avoid retransmission delays.

What carries the argument

The central mechanism is the masked spatial-temporal transformer entropy model: a bidirectional transformer that predicts a per-token mixture-of-Gaussians distribution, conditioned on the masked current latent $y_{t,M}$ and the two previous full latents $y_{t-1}, y_{t-2}$. It extends the MaskGIT masked-token-prediction idea with a learned mask token and follows the QLDS iterative masking schedule so that encoding proceeds in multiple passes. This one model carries both halves of the codec: it grows the entropy coder (cheap bits for tokens whose distributions are well predicted) and the loss-concealment mechanism (tokens lost in transmission are sampled from the same distributions at inference).

What would settle it

Evaluate NeuralMDC on bursty, correlated loss patterns taken from real 5G traces in which entire descriptions are dropped for several consecutive frames, and feed the decoder's own reconstructed frames (with their inference errors) back as context during loss inference; if PSNR at a given loss rate falls well below the paper's random-mask evaluation, the training proxy fails to transfer.

Watch

Extended reading notes

Core claim

The paper's central claim is that masked token prediction with a bidirectional transformer unifies compression and loss concealment in a video codec, removing the need for motion vectors, warping, and residual coding. Each encoded frame's latent representation is split into descriptions along the channel axis, and the transformer models $P(y_t \mid y_{t-1}, y_{t-2})$: when a token is masked out (lost or not yet entropy coded), the network predicts a mixture-of-Gaussians distribution for it conditioned on the visible current and past tokens. Entropy coding each description against these predicted distributions makes each description independently decodable, and the same distributions let the decoder sample missing tokens during inference. On UVG and MCL-JCV, the paper reports that NeuralMDC outperforms the best existing loss-resilient neural video codec, Grace, by 2 to 8 times in PSNR and MS-SSIM under packet loss and reduces bitrate by 76.88%, while accepting a modest rate-distortion gap to the feature-propagation codec DCVC-DC when no loss occurs.

Load-bearing premise

The load-bearing premise is that randomly masking 0-100% of the latent tokens during training, while feeding clean previous frames as context, teaches the network the same behaviour it needs when whole descriptions are missing and the previously decoded frames already contain errors.

Editorial extensions

If this is right

  • Motion vector losses, which dominate the failure mode of residual-coding codecs like Grace, become irrelevant: NeuralMDC only transmits latent tokens, so no separate motion stream can be lost.
  • Because each description is independently entropy coded, any subset of received descriptions yields a valid reconstruction, enabling the codec to exploit multiple noisy paths without encoder-decoder state synchronization.
  • Error propagation is limited to about two frames, since conditioning only uses the two previous representations; this bounds the temporal impact of a lost description.
  • The bitrate overhead of increasing the number of descriptions has an upper limit, so the codec can generate many streams without unbounded cost.
  • Under equal bandwidth and transmission time, NeuralMDC without retransmission beats DCVC-DC with retransmission once RTT exceeds 10 ms, and the advantage grows with RTT.

Reading between the lines

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

  • We infer that the same masked transformer could be used for joint source-channel coding: if the expected per-path loss probability is known, the number of descriptions and the masking schedule could be chosen adaptively per frame, trading a little bitrate for a lot of resilience.
  • We infer that a direct ablation with longer context (e.g., conditioning on three or four past frames) would likely improve compression but weaken resilience, since a single corrupted past frame would then corrupt more context; the paper's two-frame choice is an implicit trade-off the authors do not test.
  • We infer that the train/inference mismatch the paper leaves open—random masks on clean frames versus whole-description losses on reconstructed frames—could be closed by fine-tuning with simulated description dropouts and fed-back reconstructions, which would either raise the reported resilience numbers or reveal their limit.
  • We infer that the loss-inference reuses the distribution prediction from entropy decoding, so in a real deployment the marginal cost of concealment is almost zero; the real cost is concentrated in training-time masking.
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

5 major / 4 minor

Summary. The paper proposes NeuralMDC, a multiple description neural video codec. It tokenizes each frame into a quantized latent representation, splits the tokens along the channel axis into multiple descriptions, and trains a masked bidirectional transformer to model the conditional distribution of the current latent representation given the two previous latent representations. This predicted distribution is used both for entropy coding each description independently and for inferring lost tokens at the decoder. The authors evaluate loss resilience on UVG and MCL-JCV under simulated packet loss, reporting large improvements over Grace and DCVC-DC, and also report rate-distortion performance and runtime. The core claim is that the masked transformer enables independently decodable, mutually refinable descriptions with state-of-the-art loss resilience.

Significance. If the claims are correct, NeuralMDC would be a significant contribution: it is the first neural video codec designed from the ground up for multiple description coding, and it offers an elegant way to make video delivery robust over noisy multi-path networks such as 5G. The paper builds on strong existing components (ELIC autoencoder, M2T/VCT transformers) and provides an extensive experimental comparison, including a retransmission-based baseline and real-world network traces. The main weakness is that the conditional entropy coding architecture appears to break when previous frames are lost, because the decoder cannot reproduce the encoder's conditioning context, which is a load-bearing issue for the entire loss-resilience evaluation.

major comments (5)
  1. [Method, 'Entropy Coding' and 'Inference of Lost Tokens' (Eq. 1 and Eq. 4)] The entropy coding cost is defined as -log p(y_i_t | y_t,M, y_{t-1}, y_{t-2}), where y_{t-1} and y_{t-2} are the clean encoder-side latents. After any loss in a previous frame, the decoder's reconstructed context ~y_{t-1} and ~y_{t-2} differs from the clean context. Arithmetic/range decoding requires the decoder to reproduce the exact probability intervals used by the encoder, so a bitstream for frame t cannot be inverted if the conditioning context is mismatched, even if all descriptions of frame t are received. The paper's claim that conditioning on only two previous frames 'confines temporal error propagation to a few local frames' addresses prediction quality, not code decodability. No synchronization mechanism (e.g., encoder-side simulation of the decoder's reconstructed context, leaky prediction, or intra-refresh) is described. This is a load-bearing correctness issue that undermines the loss-resilience results.
  2. [Experiments, 'Loss Resilience Performance'] The packet loss model for NeuralMDC is never specified. For H.264 the authors state that they 'randomly corrupt bitstreams,' but for NeuralMDC the mapping from latent tokens to packets (or descriptions) and the way that partial loss of descriptions is simulated are absent. Without this information, the loss-resilience results are not reproducible, and the comparison with baselines may not reflect the same channel conditions. This should be described in detail, including how many tokens are lost per loss event and how the decoder detects which tokens are missing.
  3. [Abstract and Results] The abstract claims that NeuralMDC outperforms Grace 'by 2 to 8 times in terms of PSNR and MS-SSIM' and achieves '76.88% bitrate savings over Grace,' while the Results section states '1.78 to 8.66 times' and does not derive the 76.88% figure from any table, equation, or curve. These quantitative claims need precise definitions and supporting evidence, including the specific loss ratios and bitrate points at which they are measured.
  4. [Appendix, 'Implementation Details'] The appendix states that 'All experiments are conducted on Nvidia A6000 GPUs and independently run three times,' yet no error bars, standard deviations, or confidence intervals are reported in any of the figures. Given the stochastic nature of both training and the random loss simulation, the reported gains need statistical significance measures, especially for the large claimed improvements.
  5. [Method, 'Loss and Training Process'] No ablation study is presented for the central design choices: the number of descriptions S, the number of QLDS iterations L, the masking ratio distribution during training, or the channel-splitting strategy. Since the paper introduces several interacting components, ablations are needed to determine which ones contribute to the reported loss resilience and rate-distortion trade-off.
minor comments (4)
  1. [Eq. (2) text] There is a typo in the text preceding Eq. (2): 'distribuion' should be 'distribution'.
  2. [Baselines with Retransmission] The text refers to 'DCVC-DV' in one sentence; this should be 'DCVC-DC' for consistency with the rest of the paper.
  3. [Figure 13 caption] The caption contains 'MS-SSM' in one instance; it should be 'MS-SSIM'.
  4. [Appendix, 'Implementation Details'] The relationship between the implementation and the M2T and VCT works is not clear; please specify which modules are reused from each and what modifications were made.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the codec's loss-resilience and rate-distortion results are externally measured, and the self-citations are motivational only.

full rationale

NeuralMDC's central claims are empirical. The masked transformer entropy model is trained on Vimeo-90K with the cross-entropy objective in Eq. (3), then evaluated on UVG and MCL-JCV against external baselines (Grace, DCVC-DC, VCT, etc.) using public checkpoints. The loss-resilience numbers are measured from corrupted bitstreams, not derived from fitted constants or from the paper's own premises. The predicted distribution p(y_t|y_{t-1}, y_{t-2}) is used both as the entropy model and as the lost-token inference model, but this is a legitimate reuse of one learned conditional model, not a case where the output is defined as the input: training masks are random (0-100%), while inference masks correspond to whole lost descriptions, and the model is tested on unseen datasets. Self-citations (Narayanan et al., Ramadan et al., Ye et al.) appear only in the motivation about 5G dynamics and are not load-bearing for any codec equation or result. No uniqueness theorem is imported from the authors' prior work, and no known result is renamed. The reviewer's decodability concern about encoder-decoder context mismatch after loss is a correctness and robustness risk, not a circularity, because it does not make the reported measurements equivalent to the paper's assumptions. Under the hard rules, this paper exhibits no significant circularity.

Assumptions & free parameters 6 free parameters · 5 assumptions · 0 invented entities

The paper introduces no new physical entities and no parameter-free derivation; it is an empirical neural codec with several hand-chosen hyperparameters and domain assumptions about latent redundancy and loss modeling.

free parameters (6)
  • Number of descriptions S = 4 in the diagram; varied from 2 to 12 in Fig. 9; value used in loss-resilience experiments not stated
    Controls redundancy and bpp overhead; the main results do not report which S was used.
  • Latent channel count c = 192
    Set throughout; determines token dimensionality and the amount of correlation available for inference.
  • GMM mixture count N_M = 3
    Chosen by hand for the entropy model; no ablation.
  • QLDS masking iterations L and alpha = L=12, alpha=2.2
    Chosen for iterative entropy coding; affects bitrate and runtime; no ablation.
  • Temporal context length = 2
    Design choice to limit error propagation; no ablation shown.
  • Rate-distortion trade-off lambdas = six values from 0.0001 to 1
    Used to generate rate-distortion curves; standard for neural codecs.
assumptions (5)
  • domain assumption Latent representations produced by the ELIC autoencoder retain sufficient spatial-temporal correlation and energy compaction for lost tokens to be inferred from received tokens and two previous frames.
    The entire MDC inference mechanism depends on this correlation; cited to He et al. 2022 and Li et al. 2023b but not directly measured.
  • domain assumption Randomly masking 0-100% of tokens during training simulates the description-level losses encountered at inference.
    Used in Eq. 3; no evaluation of train and inference mask mismatch is provided.
  • ad hoc to paper Conditioning on only the two previous frames confines temporal error propagation and is sufficient for accurate prediction.
    Contrasted with MIMT; no ablation on context length is shown.
  • domain assumption The model trained on Vimeo-90K generalizes to UVG and MCL-JCV.
    Standard ML generalization assumption; not unique to this paper.
  • domain assumption The bitstream corruption framework used for H.264 and the loss model used for neural codecs are comparable.
    The paper states H.264 corruption uses the framework but does not describe the neural codec packetization and loss model.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Robust Multiple Description Neural Video Codec with Masked Transformer for Dynamic and Noisy Networks." pith.science (2026). https://pith.science/paper/MKOGFUEQ

@misc{pith2026241207922,
  author       = {Pith},
  title        = {Pith review of: Robust Multiple Description Neural Video Codec with Masked Transformer for Dynamic and Noisy Networks},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/MKOGFUEQ}},
  note         = {Machine review of arXiv:2412.07922}
}
read the original abstract

Multiple Description Coding (MDC) is a promising error-resilient source coding method that is particularly suitable for dynamic networks with multiple (yet noisy and unreliable) paths. However, conventional MDC video codecs suffer from cumbersome architectures, poor scalability, limited loss resilience, and lower compression efficiency. As a result, MDC has never been widely adopted. Inspired by the potential of neural video codecs, this paper rethinks MDC design. We propose a novel MDC video codec, NeuralMDC, demonstrating how bidirectional transformers trained for masked token prediction can vastly simplify the design of MDC video codec. To compress a video, NeuralMDC starts by tokenizing each frame into its latent representation and then splits the latent tokens to create multiple descriptions containing correlated information. Instead of using motion prediction and warping operations, NeuralMDC trains a bidirectional masked transformer to model the spatial-temporal dependencies of latent representations and predict the distribution of the current representation based on the past. The predicted distribution is used to independently entropy code each description and infer any potentially lost tokens. Extensive experiments demonstrate NeuralMDC achieves state-of-the-art loss resilience with minimal sacrifices in compression efficiency, significantly outperforming the best existing residual-coding-based error-resilient neural video codec.

Figures

Figures reproduced from arXiv: 2412.07922 by the authors.

Figure 1
Figure 1. Overview of NeuralMDC codec: an example of generating 4 descriptions. [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 3
Figure 3. Latent representation separation example: the [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figure 4
Figure 4. Impact of losses of motion vs. residuals [PITH_FULL_IMAGE:figures/full_fig_p004_4.png] view at source ↗
Figures from the paper (5 more)
Figure 6
Figure 6. Figure 6: At the same bitrate and without retransmis [PITH_FULL_IMAGE:figures/full_fig_p006_6.png]
Figure 8
Figure 8. Figure 8: Rate-distortion performance on UVG and MCL-JCV datasets. [PITH_FULL_IMAGE:figures/full_fig_p007_8.png]
Figure 9
Figure 9. Figure 9: The average BPP overhead incurred by multiple descriptions. 0 200 400 Time (ms) Encoding Decoding AutoEncoder Split/Merge Transformer [PITH_FULL_IMAGE:figures/full_fig_p007_9.png]
Figure 12
Figure 12. Figure 12: A representative sample of network traces [PITH_FULL_IMAGE:figures/full_fig_p011_12.png]
Figure 13
Figure 13. Figure 13: Inference examples with 50% packet loss of NeuralMDC and DCVC-DC under the same network bandwidth [PITH_FULL_IMAGE:figures/full_fig_p012_13.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

55 extracted references · 45 canonical work pages

  1. [1]

    J.; and Toderici, G

    Agustsson, E.; Minnen, D.; Johnston, N.; Balle, J.; Hwang, S. J.; and Toderici, G. 2020. Scale-space flow for end-to-end optimized video compression. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 8503--8512

    work page 2020

  2. [2]

    Badr, A.; Khisti, A.; Tan, W.-t.; Zhu, X.; and Apostolopoulos, J. 2017. FEC for VoIP using dual-delay streaming codes. In IEEE INFOCOM 2017-IEEE Conference on Computer Communications, 1--9. IEEE

    work page 2017

  3. [3]

    Chang, H.; Zhang, H.; Jiang, L.; Liu, C.; and Freeman, W. T. 2022. Maskgit: Masked generative image transformer. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 11315--11325

    2022

  4. [4]

    N.; and Shrivastava, A

    Chen, H.; He, B.; Wang, H.; Ren, Y.; Lim, S. N.; and Shrivastava, A. 2021. Nerv: Neural representations for videos. Advances in Neural Information Processing Systems, 34: 21557--21568

    2021

  5. [5]

    Y.; et al

    Cheng, Y.; Zhang, Z.; Li, H.; Arapin, A.; Zhang, Y.; Zhang, Q.; Liu, Y.; Du, K.; Zhang, X.; Yan, F. Y.; et al. 2024. \ GRACE \ : \ Loss-Resilient \ \ Real-Time \ Video through Neural Codecs. In 21st USENIX Symposium on Networked Systems Design and Implementation (NSDI 24), 509--531

    2024

  6. [6]

    Conci, N.; and De Natale, F. G. 2007. Real-time multiple description intra-coding by sorting and interpolation of coefficients. Signal, Image and Video Processing, 1: 1--10

    work page 2007

  7. [7]

    Fleming, M.; and Effros, M. 1999. Generalized multiple description vector quantization. In Proceedings DCC'99 Data Compression Conference (Cat. No. PR00096), 3--12. IEEE

    work page 1999

  8. [8]

    Franchi, N.; Fumagalli, M.; Lancini, R.; and Tubaro, S. 2005. Multiple description video coding for scalable and robust transmission over IP. IEEE Transactions on circuits and systems for video technology, 15(3): 321--334

    work page 2005

Show all 55 references
  1. [9]

    He, D.; Yang, Z.; Peng, W.; Ma, R.; Qin, H.; and Wang, Y. 2022. Elic: Efficient learned image compression with unevenly grouped space-channel contextual adaptive coding. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 5718--5727

    2022

  2. [10]

    Hu, X.; Ghosh, A.; Liu, X.; Zhang, Z.-L.; and Shroff, N. 2023. COREL: Constrained Reinforcement Learning for Video Streaming ABR Algorithm Design Over mmWave 5G. In 2023 IEEE International Workshop Technical Committee on Communications Quality and Reliability (CQR), 1--6. IEEE

    2023

  3. [11]

    Hu, X.; Pan, Y.; Wang, Y.; Zhang, L.; and Shirmohammadi, S. 2021. Multiple description coding for best-effort delivery of light field video using GNN-based compression. IEEE Transactions on Multimedia, 25: 690--705

    2021

  4. [12]

    Hu, Z.; Lu, G.; Guo, J.; Liu, S.; Jiang, W.; and Xu, D. 2022. Coarse-to-fine deep video coding with hyperprior-guided mode prediction. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 5921--5930

    2022

  5. [13]

    Hu, Z.; Lu, G.; and Xu, D. 2021. FVC: A new framework towards deep video compression in feature space. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 1502--1511

    2021

  6. [14]

    Kazemi, M.; Shirmohammadi, S.; and Sadeghi, K. H. 2014. A review of multiple description coding techniques for error-resilient video delivery. Multimedia Systems, 20: 283--309

    2014

  7. [15]

    M.; Gao, G.; Zhang, F.; Gower, A.; and Bull, D

    Kwan, H. M.; Gao, G.; Zhang, F.; Gower, A.; and Bull, D. 2024. HiNeRV: Video Compression with Hierarchical Encoding-based Neural Representation. Advances in Neural Information Processing Systems, 36

    2024

  8. [16]

    H.; Antonini, M.; Lambert, M.; and Alioua, K

    Le, T. H.; Antonini, M.; Lambert, M.; and Alioua, K. 2023. Multiple description video coding for real-time applications using HEVC. In 2023 IEEE International Conference on Image Processing (ICIP), 2580--2584. IEEE

    2023

  9. [17]

    H.; Pic, X.; and Antonini, M

    Le, T. H.; Pic, X.; and Antonini, M. 2023. INR-MDSQC: Implicit Neural Representation Multiple Description Scalar Quantization for robust image Coding. In 2023 IEEE 25th International Workshop on Multimedia Signal Processing (MMSP), 1--6. IEEE

    2023

  10. [18]

    Li, J.; Li, B.; and Lu, Y. 2021 a . Deep contextual video compression. Advances in Neural Information Processing Systems, 34: 18114--18125

    2021

  11. [19]

    Li, J.; Li, B.; and Lu, Y. 2021 b . Deep contextual video compression. Advances in Neural Information Processing Systems, 34: 18114--18125

    2021

  12. [20]

    Li, J.; Li, B.; and Lu, Y. 2023. Neural video compression with diverse contexts. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 22616--22626

    2023

  13. [21]

    Li, Q.; Zhang, Z.; Liu, Y.; Tan, Z.; Peng, C.; and Lu, S. 2023 a . CA++: Enhancing Carrier Aggregation Beyond 5G. In Proceedings of the 29th Annual International Conference on Mobile Computing and Networking, 1--14

    2023

  14. [22]

    Li, T.; Chang, H.; Mishra, S.; Zhang, H.; Katabi, D.; and Krishnan, D. 2023 b . Mage: Masked generative encoder to unify representation learning and image synthesis. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2142--2152

    2023

  15. [23]

    Lin, K.; Jia, C.; Zhang, X.; Wang, S.; Ma, S.; and Gao, W. 2022. DMVC: Decomposed motion modeling for learned video compression. IEEE Transactions on Circuits and Systems for Video Technology

    2022

  16. [24]

    Liu, T.; Wu, K.; Wang, Y.; Liu, W.; Yap, K.-H.; and Chau, L.-P. 2024. Bitstream-Corrupted Video Recovery: A Novel Benchmark Dataset and Method. Advances in Neural Information Processing Systems, 36

    2024

  17. [25]

    Lu, G.; Ouyang, W.; Xu, D.; Zhang, X.; Cai, C.; and Gao, Z. 2019. Dvc: An end-to-end deep video compression framework. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 11006--11015

    2019

  18. [26]

    Mentzer, F.; Agustson, E.; and Tschannen, M. 2023. M2t: Masking transformers twice for faster decoding. In Proceedings of the IEEE/CVF International Conference on Computer Vision, 5340--5349

    2023

  19. [27]

    Mentzer, F.; Toderici, G.; Minnen, D.; Hwang, S.-J.; Caelles, S.; Lucic, M.; and Agustsson, E. 2022. Vct: A video compression transformer. arXiv preprint arXiv:2206.07307

    2022 arXiv

  20. [28]

    Mercat, A.; Viitanen, M.; and Vanne, J. 2020. UVG dataset: 50/120fps 4K sequences for video codec analysis and development. In Proceedings of the 11th ACM Multimedia Systems Conference, 297--302

    2020

  21. [29]

    Minnen, D.; Ball \'e , J.; and Toderici, G. D. 2018. Joint autoregressive and hierarchical priors for learned image compression. Advances in neural information processing systems, 31

    2018

  22. [30]

    Minnen, D.; and Singh, S. 2020. Channel-wise autoregressive entropy models for learned image compression. In 2020 IEEE International Conference on Image Processing (ICIP), 3339--3343. IEEE

    2020

  23. [31]

    Narayanan, A.; Ramadan, E.; Carpenter, J.; Liu, Q.; Liu, Y.; Qian, F.; and Zhang, Z.-L. 2020 a . A First Look at Commercial 5G Performance on Smartphones. In Proceedings of The Web Conference 2020, WWW '20, 894–905. New York, NY, USA: Association for Computing Machinery. ISBN ...

    2020

  24. [32]

    A.; Dayalan, U

    Narayanan, A.; Ramadan, E.; Mehta, R.; Hu, X.; Liu, Q.; Fezeu, R. A.; Dayalan, U. K.; Verma, S.; Ji, P.; Li, T.; et al. 2020 b . Lumos5G: Mapping and predicting commercial mmWave 5G throughput. In Proceedings of the ACM Internet Measurement Conference, 176--193

    2020

  25. [33]

    M.; et al

    Narayanan, A.; Zhang, X.; Zhu, R.; Hassan, A.; Jin, S.; Zhu, X.; Zhang, X.; Rybkin, D.; Yang, Z.; Mao, Z. M.; et al. 2021. A variegated look at 5G in the wild: performance, power, and QoE implications. In Proceedings of the 2021 ACM SIGCOMM 2021 Conference, 610--625

    2021

  26. [34]

    M.; and Hallapuro, A

    Radulovic, I.; Frossard, P.; Wang, Y.-K.; Hannuksela, M. M.; and Hallapuro, A. 2009. Multiple description video coding with H. 264/AVC redundant pictures. IEEE Transactions on Circuits and Systems for Video Technology, 20(1): 144--148

    2009

  27. [35]

    K.; Fezeu, R

    Ramadan, E.; Narayanan, A.; Dayalan, U. K.; Fezeu, R. A. K.; Qian, F.; and Zhang, Z.-L. 2021. Case for 5G-Aware Video Streaming Applications. In Proceedings of the 1st Workshop on 5G Measurements, Modeling, and Use Cases, 5G-MeMU '21, 27–34. New York, NY, USA: Association for ...

    2021

  28. [36]

    G.; Tatwawadi, K.; Nair, S.; Lytle, C.; and Bourdev, L

    Rippel, O.; Anderson, A. G.; Tatwawadi, K.; Nair, S.; Lytle, C.; and Bourdev, L. 2021. Elf-vc: Efficient learned flexible-rate video coding. In Proceedings of the IEEE/CVF International Conference on Computer Vision, 14479--14488

    2021

  29. [37]

    I.; Ye, W.; Zhang, Z.-L.; and Ghosh, M

    Rochman, M. I.; Ye, W.; Zhang, Z.-L.; and Ghosh, M. 2023. A Comprehensive Real-World Evaluation of 5G Improvements over 4G in Low-and Mid-Bands. arXiv preprint arXiv:2312.00957

    2023 arXiv

  30. [38]

    Shirani, S. 2006. Content-based multiple description image coding. IEEE transactions on multimedia, 8(2): 411--419

    2006

  31. [39]

    Stevens, W. 1997. RFC2001: TCP slow start, congestion avoidance, fast retransmit, and fast recovery algorithms

    1997

  32. [40]

    Theis, L.; Shi, W.; Cunningham, A.; and Husz \'a r, F. 2022. Lossy image compression with compressive autoencoders. In International conference on learning representations

    2022

  33. [41]

    Tillo, T.; and Olmo, G. 2004. Low complexity pre postprocessing multiple description coding for video streaming. In Proceedings. 2004 International Conference on Information and Communication Technologies: From Theory to Applications, 2004., 519--520. IEEE

    2004

  34. [42]

    Y.; Jin, L.; Song, L.; Wang, P.; Katsavounidis, I.; Aaron, A.; and Kuo, C.-C

    Wang, H.; Gan, W.; Hu, S.; Lin, J. Y.; Jin, L.; Song, L.; Wang, P.; Katsavounidis, I.; Aaron, A.; and Kuo, C.-C. J. 2016. MCL-JCV: a JND-based H. 264/AVC video quality assessment dataset. In 2016 IEEE international conference on image processing (ICIP), 1509--1513. IEEE

    2016

  35. [43]

    T.; Vaishampayan, V.; and Reibman, A

    Wang, Y.; Orchard, M. T.; Vaishampayan, V.; and Reibman, A. R. 2001. Multiple description coding using pairwise correlating transforms. IEEE Transactions on Image Processing, 10(3): 351--366

    2001

  36. [44]

    P.; and Bovik, A

    Wang, Z.; Simoncelli, E. P.; and Bovik, A. C. 2003. Multiscale structural similarity for image quality assessment. In The Thrity-Seventh Asilomar Conference on Signals, Systems & Computers, 2003, volume 2, 1398--1402. Ieee

    2003

  37. [45]

    B.; and Bhargava, V

    Wicker, S. B.; and Bhargava, V. K. 1999. Reed-Solomon codes and their applications. John Wiley & Sons

    1999

  38. [46]

    Xiang, J.; Tian, K.; and Zhang, J. 2022. Mimt: Masked image modeling transformer for video compression. In The Eleventh International Conference on Learning Representations

    2022

  39. [47]

    Xue, T.; Chen, B.; Wu, J.; Wei, D.; and Freeman, W. T. 2019. Video enhancement with task-oriented flow. International Journal of Computer Vision, 127: 1106--1125

    2019

  40. [48]

    Yap c , Y.; Demir, B.; Ert \"u rk, S.; and Urhan, O. 2008. Downsampling-based multiple description image coding using optimal filtering. Journal of Electronic Imaging, 17(3): 033018--033018

    2008

  41. [49]

    K.; Hassan, A.; Fezeu, R

    Ye, W.; Hu, X.; Sleder, S.; Zhang, A.; Dayalan, U. K.; Hassan, A.; Fezeu, R. A.; Jajoo, A.; Lee, M.; Ramadan, E.; et al. 2024. Dissecting Carrier Aggregation in 5G Networks: Measurement, QoE Implications and Prediction. In Proceedings of the ACM SIGCOMM 2024 Conference, 340--357

    2024

  42. [50]

    G.; Yang, M.-H.; Hao, Y.; Essa, I.; et al

    Yu, L.; Cheng, Y.; Sohn, K.; Lezama, J.; Zhang, H.; Chang, H.; Hauptmann, A. G.; Yang, M.-H.; Hao, Y.; Essa, I.; et al. 2023. Magvit: Masked generative video transformer. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 10459--10469

    2023

  43. [51]

    Zhao, L.; Bai, H.; Wang, A.; and Zhao, Y. 2018. Multiple description convolutional neural networks for image compression. IEEE Transactions on Circuits and Systems for Video Technology, 29(8): 2494--2508

    2018

  44. [52]

    Zhao, L.; Zhang, J.; Bai, H.; Wang, A.; and Zhao, Y. 2022. LMDC: Learning a multiple description codec for deep learning-based image compression. Multimedia Tools and Applications, 81(10): 13889--13910

    2022

  45. [53]

    Zheng, Z.; Ma, Y.; Liu, Y.; Yang, F.; Li, Z.; Zhang, Y.; Zhang, J.; Shi, W.; Chen, W.; Li, D.; et al. 2021. Xlink: Qoe-driven multi-path quic transport in large-scale video services. In Proceedings of the 2021 ACM SIGCOMM 2021 Conference, 418--432

    2021

  46. [54]

    , " * write output.state after.block = add.period write newline

    ENTRY address archivePrefix author booktitle chapter edition editor eid eprint howpublished institution isbn journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all...

  47. [55]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 gl...

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.