Cross-Domain Generalization in Optical Networks via Joint Contrastive and Classification Learning

Ali Al Housseini; Carlos Natalino; Omran Ayoub; Paolo Monti

REVIEW 4 major objections 6 minor 21 references

Cross-Domain Generalization in Optical Networks via Joint Contrastive and Classification Learning

T0 review · 4 major / 6 minor · reviewed 2026-08-01 · deepseek-v4-flash

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

Pith's one-line read A model trained with joint contrastive and classification objectives on one optical network keeps detecting failing lightpaths on unseen networks, and fine-tunes to near-target performance with just 1–2% labeled samples.

desk verdict The paper has a real idea, but the zero-shot gains are confounded by balanced sampling, so the central attribution to contrastive learning doesn't yet stand. read the letter →

arxiv 2607.20666 v1 pith:QRBIQ3KO submitted 2026-07-22 cs.NI cs.LG

Ali Al Housseini , Carlos Natalino , Paolo Monti , Omran Ayoub This is my paper · ORCID

classification cs.NIcs.LG

keywords contrastivelearningdeepmetricmulti-similaritylosscross-domaingeneralizationqualityoftransmissionestimationopticalnetworkstransferfew-shotadaptation

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

This paper tries to establish that a model trained on one optical network can keep working on a different, unseen network if its feature space is shaped by contrastive and classification objectives at the same time, rather than by classification alone or by a frozen representation followed by a classifier. The proposed method—joint Multi-Similarity Loss plus binary cross-entropy on a projection head—learns embeddings organized by relative QoT similarity, which the authors argue are less sensitive to domain-specific amplifier profiles and link-length distributions. On lightpath quality-of-transmission (QoT) estimation, the method improves zero-shot Macro-F1 on the two hardest cross-topology transfers (D1→D2: 0.534 to 0.644; D1→D3: 0.487 to 0.564) and establishes a worst-case PR-AUC floor of 0.656 across six transfer directions, higher than the best tree-ensemble baseline's 0.542. With 1–2% labeled target lightpaths, fine-tuning the joint model closes most of the cross-domain gap, whereas retraining baselines from scratch can get worse before they get better. If correct, the result gives network operators a deployable zero-shot QoT pre-screening block and a data-efficient adaptation path.

What carries the argument

The central object is the jointly optimized latent space: a backbone deep network maps lightpath features to embeddings, and two losses act on them. The Multi-Similarity Loss (MS-Loss) is a contrastive loss that mines hard positive and negative pairs inside each mini-batch using multiple pairwise cosine similarities, pulling same-class lightpaths together and pushing different-class lightpaths apart. A projection network applies binary cross-entropy on the same embeddings, keeping the representation aligned with the QoT decision boundary. A class-conditional sampler balances each mini-batch so the contrastive signal is not dominated by healthy lightpaths. The MS-Loss encodes the domain-invar

What would settle it

Train the same architecture with the MS-Loss term removed (γ=0) on D1 and evaluate zero-shot on D2 and D3; if Macro-F1 does not drop, the contrastive objective is not what carries the generalization, and the paper's central claim is unsupported.

Watch

Extended reading notes

Core claim

Central claim: jointly training a contrastive loss and a classification head in one loop produces a latent representation that stays aligned with the QoT decision boundary when the network changes. The contrastive loss acts only on relative pairwise cosine similarities, so lightpaths with different absolute amplifier gains or link lengths but comparable SNR margin cluster together, and the classifier reads the boundary from that class-relative geometry. On six cross-topology transfers, the hardest shifts (D1→D2, D1→D3) improve Macro-F1 from 0.534 to 0.644 and from 0.487 to 0.564, and worst-case PR-AUC is 0.656 versus 0.542 for the best tree-ensemble baseline. Warm-start fine-tuning with 1–2%

Load-bearing premise

The load-bearing premise is that the relative pairwise cosine-similarity structure learned on the source network remains meaningful in the target network—lightpaths with comparable QoT margin cluster together regardless of topology, amplifier gains, or class balance—so target decision boundaries coincide with source clusters.

Editorial extensions

If this is right

Zero-shot deployment: a model trained only on one network's labeled data can be dropped into an unseen network and still detect failing lightpaths at a usable rate, without any target labels.
Worst-case floor: across six transfer directions the method's PR-AUC never falls below 0.656, whereas the best tree-ensemble baseline drops to 0.542, so the method protects against catastrophic generalization failures.
Rapid adaptation: with only 1–2% labeled target lightpaths, Macro-F1 closes most of the cross-domain gap (e.g., D1→D2 rises from 0.644 to roughly 0.94 at 2% injection).
Joint optimization beats the sequential pipeline: freezing the representation first and then training a classifier collapses under shift (MF1 0.33–0.35 in some transfers), while joint training stays robust.
The resulting model can serve as a QoT pre-screening block fed by telemetry, with retraining triggered only when injected-sample PR-AUC drops below a deployment-defined threshold.

Reading between the lines

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

If pairwise cosine-similarity geometry is the true carry-over, the same joint contrastive-plus-classification objective should transfer to other optical-network tasks with a natural similarity notion, such as OSNR regression or modulation-format identification; this extension is not tested in the paper.
The observed non-monotonicity of retrain-from-scratch baselines under small target injections suggests that naively pooling source and target data can hurt early adaptation; the paper's warm-start fine-tuning is a safer default, but this deserves a direct comparison on more dataset pairs.
A testable refinement would be to measure how the claimed invariance degrades as class-prior shift increases; the paper's hardest cases combine prior shift with configuration changes, so isolating the two would clarify whether the robustness comes from the contrastive geometry or from the class-conditional sampler.

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.

Desk Editor's Note

read the letter

Quick take: this is a decent empirical study of joint contrastive + classification training for QoT estimation across three public datasets. The authors report meaningful worst-case MF1/PR-AUC gains on hard transfer directions. But the comparison is confounded: the proposed pipeline uses a class-conditional sampler that balances the mini-batches, while every baseline trains on the imbalanced data as-is. Since D1 has 24.2% failures vs ~7% in D2/D3, balanced sampling alone could explain much of the improvement. The paper nowhere isolates the contrastive objective from the sampling scheme, and explicitly credits the contrastive loss for the gains. That's the load-bearing issue.

What's genuinely useful: the paper clearly documents the domain shift across the three topologies, reports six transfer directions, and shows the method doesn't collapse in easy transfers. Public code is promised, and the datasets are external. If the confound were fixed with an ablation (balanced sampling for baselines, or unbalanced training for the proposed method), the result would be worth serious attention.

Other soft spots are minor in comparison but worth noting. No error bars anywhere; all numbers are single runs, so the 0.03–0.11 improvements could be noise. The few-shot comparison gives the proposed method a warm-start fine-tuning advantage while baselines are retrained from scratch, so the rapid-adaptation claim is not an apples-to-apples comparison. Hyperparameters for MS-Loss and the loss weights are not fully specified.

On the citation pattern: it's honestly situated in the optical-network ML literature, and the self-citation to the Siamese zero-day anomaly detection work is relevant. No problem there.

Bottom line: the paper's core evidence does not yet support its central attribution. It's a promising direction and a fair referee could help the authors tighten the experimental design. I'd send it to review, but I would not cite the zero-shot claim as it stands.

Referee Report

4 major / 6 minor

Summary. The paper proposes a joint contrastive-and-classification training scheme for cross-domain QoT estimation in optical networks. The architecture combines a class-conditional mini-batch sampler, a backbone network trained with Multi-Similarity Loss on cosine similarities, and a projection network trained with binary cross-entropy; the two objectives are optimized jointly via a weighted sum. The authors evaluate on three public datasets (D1, D2, D3) from different topologies and operational modes, covering six zero-shot transfer directions and few-shot target-data injection at 1–4%. They report that the proposed method improves worst-case Macro-F1 and PR-AUC relative to tabular baselines, e.g., D1→D2 MF1 from 0.534 to 0.644 and PR-AUC from 0.558 to 0.656, and that fine-tuning with 1–2% target labels closes most of the cross-domain gap. The paper concludes that joint contrastive/classification learning learns task-relevant representations that are stable across network domains.

Significance. Cross-domain QoT estimation is practically important, and public datasets and open-source code strengthen reproducibility. The paper also targets a meaningful operational objective—improving worst-case, not just average, transfer performance—and compares against several standard baselines. If the reported gains are real and attributable to the proposed learning objective, the method would be a useful contribution to optical-network ML. However, the current experimental design contains a central confound: the proposed method uses a class-conditional balanced sampler, while the baselines are trained on the original imbalanced data. This alone can explain part of the Macro-F1 gains under class-prior shift. The few-shot comparison also gives the proposed method a warm-start advantage over baselines that are retrained from scratch. These issues must be resolved with additional ablations before the main claims can be accepted.

major comments (4)

[§III-B (Data Module) and §IV-B/Table I] The zero-shot comparison is confounded by the class-conditional sampler. The proposed architecture trains on mini-batches with equal numbers of positive and negative samples, while the tabular baselines (RF, XGB, CatBoost, LR) and the supervised MLP are trained on the original imbalanced data. Since D1 has a 24.2% failure rate and D2/D3 have ~7%, balanced sampling alone can substantially increase Macro-F1 by preventing majority-class collapse. The paper attributes the gains to the contrastive objective ('These gains stem from the contrastive objective...'), but it has not isolated the sampling scheme from the learning objective. Please report: (i) the proposed method without the class-conditional sampler, (ii) the baselines trained with the same class-balanced sampler or with class weights, and (iii) an ablation with reweighted cross-entropy only, to establish which component drives the
[§IV-C and Fig. 2] The few-shot comparison is not methodologically fair. For baselines, each injection ratio is obtained by retraining from scratch on the union of the full source dataset and i% target data; for the proposed method, the model is warm-started from a source-trained checkpoint and fine-tuned on only the injected target samples. This gives the proposed method two advantages: a better initialization and a different optimization protocol. The observed monotonic improvement and rapid adaptation may therefore reflect fine-tuning from a strong source model rather than the joint contrastive/classification design. Please compare under identical protocols—e.g., fine-tune a source-pretrained baseline as well, or retrain the proposed method from scratch on the same union—and additionally report results with balanced sampling for all methods.
[Table I and Fig. 2] No error bars, confidence intervals, or repeated-run statistics are reported. All cross-dataset comparisons appear to be single-point estimates, so it is unclear whether differences such as Ours vs. XGB in D2→D3 (MF1 0.974 vs. 0.992) or Ours vs. CatB in D3→D1 (MF1 0.674 vs. 0.523) are statistically reliable. Given that the paper's central claim is about robustness and stability, the authors should report mean ± std over multiple random seeds and, where possible, paired significance tests across transfer directions.
[§IV-B, Separate vs. Joint] The Separate baseline is used to motivate joint optimization, but its behavior is not adequately explained. In several settings (e.g., D3 as source) it reports MF1 near 0.333, which is essentially the value of a constant classifier that never predicts one class. The paper does not state whether Separate uses the same class-conditional sampler, the same backbone, or how the classifier is trained on frozen embeddings. Without these details, and without an explanation for the chance-level performance, the comparison does not cleanly support the claim that joint training is superior to a sequentially trained representation.

minor comments (6)

[Table I] The table formatting is broken in several cells, e.g., '0.9710.994' and '0.9930.990 0.993' in the D2-source D3-target column. Please fix the spacing so all values are readable.
[§IV-A / §III-B] Hyperparameters are not reported: the MS-Loss margin m, weighting exponents α and β, loss weights γ and η, the class-conditional sampler counts M and N, and the backbone/projection network architectures, optimizer, learning rate, and epochs. This information is necessary to reproduce the results and to assess sensitivity.
[§IV-B] The statement that the representation is 'largely insensitive to dataset-specific amplifier gain profiles and link-length distributions' is an interpretation, not an experimental finding. No embedding visualization or quantitative invariance analysis is provided. Please soften the claim or add evidence.
[§IV-C / Conclusion] The conclusion's operational suggestion—retraining triggered by a PR-AUC threshold—is not evaluated in the paper. It would be better to present this as future work rather than as a demonstrated property.
[General] The terms 'Online Transceiver Mode' and 'Predefined Transceiver Mode' are used without definition. A one-sentence explanation of what these operational modes mean would help readers who are not specialists in optical transceivers.
[Fig. 2] The figure caption should specify the exact transfer directions shown in each panel, the meaning of the curves (which method is which line), and whether the plotted values are averages over multiple runs. Currently the reader must infer this from the text.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity; central claims rest on external benchmark evaluation.

full rationale

The paper's central claim—that joint contrastive and classification learning improves cross-domain QoT generalization—is supported by experiments on public external datasets (D1, D2, D3 from Fraunhofer HHI, refs. [20], [21]) against standard baselines. The zero-shot setting uses no target labels, so the reported MF1/PR-AUC numbers are not fitted quantities renamed as predictions. The loss functions are standard external definitions (MS-Loss [17], cross-entropy), and the class-conditional sampler is external (FaceNet [16]). Self-citations to co-authored Siamese-network papers ([12], [14]) appear only as related work and do not carry the derivation. No uniqueness theorem or prior result by the same authors is invoked to force the methodology. The only notable concern—that tabular baselines may not use the same class-balanced sampling as the proposed Data Module—is an experimental-design fairness issue, not a circularity: it does not make any result true by construction. Therefore no circular step meeting the required evidentiary standard can be identified.

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

The paper introduces no new physical or conceptual entity. It introduces an architecture combination of backbone plus projection network trained jointly with MS-Loss and cross-entropy, and several tunable hyperparameters. The main unstated premises are the existence of a transferable latent geometry and the equivalence of QoT labels across datasets.

free parameters (4)

MS-Loss margin (m)
Hyperparameter in Eq. (3) controlling the triplet similarity threshold; value not reported in the paper.
MS-Loss weighting exponents (alpha, beta)
Hyperparameters in Eq. (3) controlling positive and negative pair weighting; values not reported.
Loss balance weights (gamma, eta)
Weights in Eq. (4) balancing contrastive and cross-entropy losses; chosen by hand and not reported.
Class-conditional sampler class counts (M, N)
Batch composition B equals M plus N in Sec. III-B assumes a fixed balance of negative and positive classes; the specific counts are not given.

assumptions (4)

domain assumption Binary QoT label semantics are identical across D1-D3; only marginal distributions differ.
Stated in Sec. III-A: the underlying learning task is assumed to remain unchanged. If the label semantics differ, transfer is ill-defined.
domain assumption A shared latent geometry exists in which QoT classes are compact and separated across topologies.
This is the core inductive bias of the method; the paper does not prove it and only argues qualitatively from pairwise similarity invariance.
domain assumption The D1-D3 datasets from [20] are representative of real cross-domain shifts in optical networks.
External validity of the empirical claim rests on these datasets being a meaningful proxy for operational heterogeneity.
standard math Multi-Similarity Loss and pair mining behave as described in the cited source [17].
The paper relies on the published properties of MS-Loss without re-deriving them.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Cross-Domain Generalization in Optical Networks via Joint Contrastive and Classification Learning." pith.science (2026). https://pith.science/paper/QRBIQ3KO

@misc{pith2026260720666,
  author       = {Pith},
  title        = {Pith review of: Cross-Domain Generalization in Optical Networks via Joint Contrastive and Classification Learning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/QRBIQ3KO}},
  note         = {Machine review of arXiv:2607.20666}
}

read the original abstract

The robustness of machine learning techniques across heterogeneous network domains remains an open challenge in optical networks. Models trained on data from a specific topology or operational configuration often exhibit degraded performance when deployed in unseen networks. In this work, we address this challenge by proposing a representation learning technique aimed at capturing task-relevant relationships that remain stable across domains. The proposed technique is based on a novel joint contrastive and classification learning approach in which representation learning and task optimization are performed simultaneously, allowing both objectives to shape the latent space. Experimental results on a representative use case, namely, lightpath quality of transmission estimation, demonstrate the effectiveness of our approach compared to baseline approaches, and highlight its capacity for rapid adaptation, providing excellent performance even with limited fine-tuning.

Figures

Figures reproduced from arXiv: 2607.20666 by the authors.

**Figure 1.** Framework of representation learning for QoT lightpath. (1) Dataset is first sampled by a mini-batch sampler. (2) The backbone neural network [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗

**Figure 2.** Performance comparison in terms of MF1 and PR-AUC across (a) D1 [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

21 extracted references · 1 linked inside Pith

[1]

On the generalization of cognitive optical networking applications using composable machine learning,

H. Gaoet al., “On the generalization of cognitive optical networking applications using composable machine learning,”Journal of Optical Communications and Networking, vol. 16, no. 6, pp. 631–643, 2024

2024
[2]

Generalizability of ML-based classification of state of polarization signatures across different bands and links,

L. Sadighiet al., “Generalizability of ML-based classification of state of polarization signatures across different bands and links,” inEuropean Conference on Optical Communications (ECOC), 2025, p. Th.02.01.2

2025
[3]

A module to enhance the generalization ability of end- to-end deep learning systems in optical fiber communications,

Z. Liuet al., “A module to enhance the generalization ability of end- to-end deep learning systems in optical fiber communications,”Journal of Lightwave Technology, vol. 43, no. 2, pp. 596–601, 2025

2025
[4]

Transfer learning aided QoT computation in network operating with the 400ZR standard,

F. Usmaniet al., “Transfer learning aided QoT computation in network operating with the 400ZR standard,” inInternational Conference on Optical Network Design and Modeling (ONDM), 2022

2022
[5]

Integrating knowledge distillation and transfer learning for en- hanced QoT-estimation in optical networks,

——, “Integrating knowledge distillation and transfer learning for en- hanced QoT-estimation in optical networks,”IEEE Access, vol. 12, pp. 156 785–156 802, 2024

2024
[6]

Neuron-level transfer learning for ANN-based QoT estimation in optical networks,

Y . Zhouet al., “Neuron-level transfer learning for ANN-based QoT estimation in optical networks,” inAsia Communications and Photonics Conference / International Photonics and Optoelectronics Meetings (ACP/POEM), 2023

2023
[7]

Evolutionary neuron-level transfer learning for QoT estimation in optical networks,

——, “Evolutionary neuron-level transfer learning for QoT estimation in optical networks,”Journal of Optical Communications and Networking, vol. 16, no. 4, pp. 432–448, 2024

2024
[8]

QoT estimation with margin-driven transfer learn- ing in time-varying optical networks,

P. Lechowiczet al., “QoT estimation with margin-driven transfer learn- ing in time-varying optical networks,” inOptical Fiber Communication Conference (OFC), 2025, p. M1J.5

2025

Show all 21 references

[9]

Automated, interpretable and efficient ML models for real-world lightpaths’ quality of transmission estimation,

S. Aladinet al., “Automated, interpretable and efficient ML models for real-world lightpaths’ quality of transmission estimation,”IEEE Open Journal of the Communications Society, vol. 6, pp. 9785–9801, 2025

2025
[10]

Multi-span optical power spectrum prediction using ML-based EDFA models and cascaded learning,

Z. Wanget al., “Multi-span optical power spectrum prediction using ML-based EDFA models and cascaded learning,” inOptical Fiber Communication Conference (OFC), 2024, p. M1H.6

2024
[11]

Leveraging shared data and models for ML-based QoT estimation: Toward standardized and generalizable models,

H. Akbariet al., “Leveraging shared data and models for ML-based QoT estimation: Toward standardized and generalizable models,” inEuropean Conference on Optical Communications (ECOC), 2025, p. W.04.01.3

2025
[12]

One-shot learning for modulation format identifica- tion in evolving optical networks,

C. Natalinoet al., “One-shot learning for modulation format identifica- tion in evolving optical networks,” inOSA Advanced Photonics Congress (AP), 2019, p. JW4A.2

2019
[13]

Fault tracing based on Siamese neural network for optical networks,

Y . Gaoet al., “Fault tracing based on Siamese neural network for optical networks,” inIEEE Opto-Electronics and Communications Conference (OECC), 2024

2024
[14]

A unified Siamese learning framework for zero-day anomaly detection and classification in optical networks,

C. Natalinoet al., “A unified Siamese learning framework for zero-day anomaly detection and classification in optical networks,” inOptical Fiber Communication Conference (OFC), 2026, to appear. [Online]. Available: https://research.chalmers.se/en/publication/549889

2026
[15]

Deep metric learning using triplet network,

E. Hofferet al., “Deep metric learning using triplet network,” in International Workshop on Similarity-Based Pattern Recognition, 2015, pp. 84–92

2015
[16]

Facenet: A unified embedding for face recognition and clustering,

F. Schroffet al., “Facenet: A unified embedding for face recognition and clustering,” inProceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2015, pp. 815–823

2015
[17]

Multi-similarity loss with general pair weighting for deep metric learning,

X. Wanget al., “Multi-similarity loss with general pair weighting for deep metric learning,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2019, pp. 5022–5030

2019
[18]

Dimensionality reduction by learning an invariant mapping,

R. Hadsellet al., “Dimensionality reduction by learning an invariant mapping,” inIEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR’06), vol. 2, 2006, pp. 1735–1742

2006
[19]

PyTorch metric learning,

K. Musgraveet al., “PyTorch metric learning,”arXiv preprint arXiv:2008.09164, 2020

2008 arXiv
[20]

ML-assisted QoT estimation: a dataset collection and data visualization for dataset quality evaluation,

G. Bergket al., “ML-assisted QoT estimation: a dataset collection and data visualization for dataset quality evaluation,”Journal of Optical Communications and Networking, vol. 14, no. 3, pp. 43–55, 2022

2022
[21]

QoT dataset collection,

Fraunhofer Heinrich-Hertz-Institut, “QoT dataset collection,” https: //www.hhi.fraunhofer.de/en/pn-software/qot-dataset-collection.html, 2026, accessed: 2026-03-29

2026

Pith tools

Reviewed August 1, 2026 · model on record in the stance chip above.

[1] [1]

On the generalization of cognitive optical networking applications using composable machine learning,

H. Gaoet al., “On the generalization of cognitive optical networking applications using composable machine learning,”Journal of Optical Communications and Networking, vol. 16, no. 6, pp. 631–643, 2024

2024

[2] [2]

Generalizability of ML-based classification of state of polarization signatures across different bands and links,

L. Sadighiet al., “Generalizability of ML-based classification of state of polarization signatures across different bands and links,” inEuropean Conference on Optical Communications (ECOC), 2025, p. Th.02.01.2

2025

[3] [3]

A module to enhance the generalization ability of end- to-end deep learning systems in optical fiber communications,

Z. Liuet al., “A module to enhance the generalization ability of end- to-end deep learning systems in optical fiber communications,”Journal of Lightwave Technology, vol. 43, no. 2, pp. 596–601, 2025

2025

[4] [4]

Transfer learning aided QoT computation in network operating with the 400ZR standard,

F. Usmaniet al., “Transfer learning aided QoT computation in network operating with the 400ZR standard,” inInternational Conference on Optical Network Design and Modeling (ONDM), 2022

2022

[5] [5]

Integrating knowledge distillation and transfer learning for en- hanced QoT-estimation in optical networks,

——, “Integrating knowledge distillation and transfer learning for en- hanced QoT-estimation in optical networks,”IEEE Access, vol. 12, pp. 156 785–156 802, 2024

2024

[6] [6]

Neuron-level transfer learning for ANN-based QoT estimation in optical networks,

Y . Zhouet al., “Neuron-level transfer learning for ANN-based QoT estimation in optical networks,” inAsia Communications and Photonics Conference / International Photonics and Optoelectronics Meetings (ACP/POEM), 2023

2023

[7] [7]

Evolutionary neuron-level transfer learning for QoT estimation in optical networks,

——, “Evolutionary neuron-level transfer learning for QoT estimation in optical networks,”Journal of Optical Communications and Networking, vol. 16, no. 4, pp. 432–448, 2024

2024

[8] [8]

QoT estimation with margin-driven transfer learn- ing in time-varying optical networks,

P. Lechowiczet al., “QoT estimation with margin-driven transfer learn- ing in time-varying optical networks,” inOptical Fiber Communication Conference (OFC), 2025, p. M1J.5

2025