Pith. sign in

REVIEW 4 major objections 5 minor 1 cited by

An online clustering framework interprets seafloor imagery in real time, reporting an average F1 of 0.68 across three field datasets.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · deepseek-v4-flash

2026-08-04 23:15 UTC pith:7IFQDZ6Z

load-bearing objection A useful engineering integration, but the abstract's 'highest F1' claim is contradicted by the paper's own numbers. the 4 major comments →

arxiv 2509.06678 v1 pith:7IFQDZ6Z submitted 2025-09-08 cs.CV cs.RO

Online Clustering of Seafloor Imagery for Interpretation during Long-Term AUV Operations

Cailei Liang , Adrian Bodenmann , Sam Fenton , Blair Thornton This is my paper
classification cs.CV cs.RO
keywords Autonomous underwater vehiclesSeafloor imagingOnline clusteringDirichlet process Gaussian mixtureRepresentative sample selectionCluster splitting and mergingReal-time image interpretation
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper aims to establish that a robot submarine can interpret seafloor images while the mission is still running, without waiting for the full dataset or human labels. It introduces an online clustering framework that merges and splits clusters continuously, guided by a small, fixed-size set of representative images that summarise the whole history of observations. Evaluated on three field surveys totalling about 90,000 images, the density-based variant reaches an average F1 of 0.678 (standard deviation 0.0147 across three survey trajectories), which the authors describe as comparable to offline methods and better than other online baselines; the paper also reports an HK-means-sampled variant at 0.726 with higher runtime. If the claim holds, long-endurance vehicles could produce habitat summaries, flag new patterns, and replan paths in real time instead of after recovery.

Core claim

The central claim is that dynamic split-merge clustering over a distilled representative subset can maintain accurate, self-consistent groupings of a growing image stream at constant cost. Each image is embedded in a fixed 16-dimensional space by a pretrained encoder; incoming batches are clustered, merged into the existing mixture model only if they pass Mahalanobis-distance and covariance-volume tests, and periodically re-examined for splits using AIC/BIC on a Gaussian mixture fit. Splitting decisions are made from a fixed-size (4,000-sample) density-stratified representative set, so the full history never has to be re-clustered. On the paper's three seafloor datasets, this online procedur

What carries the argument

The load-bearing object is the representative subset produced by density-based data distillation: a fixed-size queue of exemplars spaced by rank in local Manhattan-distance density, which lets split decisions see the whole historical feature distribution without storing or replaying all images. Around it, the framework uses a Dirichlet Process Gaussian Mixture Model backbone for stable Bayesian label inference, and merge criteria based on Mahalanobis distance and covariance-hyperellipsoid volume to keep the cluster count bounded.

Load-bearing premise

The pipeline assumes that the fixed, pretrained 16-dimensional feature space and a fixed 4,000-image representative subset are enough to represent all semantic categories seen during a long survey; if lighting, altitude, or new habitat types push features outside that space, the split/merge decisions will drift from the categories a human would name.

What would settle it

Compute the average F1 across the three datasets for every online variant using the paper's majority-voting protocol; if OC + HKmeans Sampling's mean (0.726) exceeds the density-based OCF's 0.678 by more than the reported 0.0147 standard deviation, then 'highest among online approaches' is not supported by the paper's own numbers. A field test: run the framework on imagery taken at a different altitude or light regime than the pretraining data and check whether the clusters still align with human labels.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

Share X Bluesky LinkedIn Reddit HN

If this is right

  • An AUV can have a continuously updated habitat map during the dive, since cluster assignments are produced after every 1,000-image batch.
  • Because the representative set is fixed at 4,000 samples, the per-batch cost of merging and splitting stays roughly constant as the survey grows, unlike full-history replay methods that scale linearly.
  • Clustering output is stable across raster and random survey trajectories (about 3% F1 variation), which is what an adaptive path-planning loop would need.
  • On class-imbalanced seafloors, choosing representatives by local density matters more than the backbone, while on balanced data even random sampling performs well.
  • The HK-means-sampled variant shows the framework can reach F1 0.726, but only by replaying historical images; the density-based variant trades a few points of F1 for bounded runtime.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • If the frozen encoder is the weakest link, a natural extension is to fine-tune or replace it periodically; the paper's own split/merge machinery would then operate on a shifting feature space, raising a question about how the representative queue would need to be re-sampled.
  • The reported gap between the abstract's 'highest among online approaches' and the paper's own Section IV.C.1 number (OC + HKmeans Sampling at 0.726 vs density-based OCF at 0.678) is a point a reader should track; it likely means 'the OCF' denotes the specific density-based configuration, not every variant.
  • A testable extension is to make the representative budget adaptive to the number of rare clusters, e.g., by allocating more of the 4,000 slots to low-density regions as imbalance is detected; this could improve minority-class recall without unbounded growth.
  • The entropy maps suggest residual errors concentrate in habitat transition zones, so a mission-level controller could use high-entropy clusters as trigger points for extra imaging or slower traversal in those areas.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. The paper proposes an online clustering framework (OCF) for real-time interpretation of seafloor imagery collected by AUVs. The method embeds images with a pretrained 16-dimensional encoder, processes data in batches, and maintains a dynamic set of clusters via Mahalanobis-distance/volume-based merging and AIC/BIC-based splitting. A fixed-size density-based representative subset (N_sub=4000) is retained to support splitting without reprocessing the full history. The framework is evaluated on three seafloor datasets (Darwin Mounds, Cawsand Bay, Southern Hydrate Ridge) under three synthetic survey trajectories, with comparisons against variants using different backbones, splitting strategies, and representative-sampling choices. The paper claims average F1=0.68, the highest among online clustering approaches, and bounded computational cost.

Significance. If the stated claims are correct, the work is a useful engineering contribution for long-duration AUV missions where in-situ interpretation is needed for adaptive path planning and communication minimisation. The evaluation uses external ground-truth labels rather than quantities defined by the method itself, so the reported F1 scores are not circular. The ablation study across backbones, splitting strategies, and representative-selection methods is informative and provides a clear efficiency/accuracy trade-off picture. However, the central numeric claim in the abstract and conclusion is contradicted by the paper's own results, and there are ambiguities about whether one of the best-performing 'online' variants is actually online. These issues must be resolved before the paper can be accepted.

major comments (4)
  1. [Abstract and §IV.C.1] The abstract states that the OCF achieves the highest average F1 score (0.68) among all comparative online clustering approaches, but §IV.C.1 reports that 'OC + HKmeans Sampling attains the best F1 (0.726)', while 'OC + Density-based Sampling attains an average F1 of 0.678'. Since OC + HKmeans Sampling is included as an online variant in §IV.B, the density-based OCF is not the highest among the online methods. The conclusion repeats this overstatement ('superior to other online clustering baselines'). The claim must be corrected, e.g., by restricting the comparison to the proposed density-based OCF configuration or by reporting the ranking that includes OC + HKmeans Sampling.
  2. [§IV.B and §IV.C.2] There is an inconsistency in how OC + HKmeans Sampling is treated. §IV.B lists it among online variants that 'update incrementally on the current batch only', but §IV.C.2 states that 'OC + HKmeans Sampling and Full History incur the highest costs because they replay all past images at each update.' If OC + HKmeans Sampling indeed replays the full history, it is not a fair online comparator under the framework's memory/compute constraints, and its higher F1 (0.726) should not be used to undercut the online comparison. Please clarify exactly what OC + HKmeans Sampling replays and whether it satisfies the online constraints defined in §III.
  3. [Abstract and §V] The reported uncertainty is numerically inconsistent. The abstract says 'a standard deviation of 3% across three distinct survey trajectories', while the conclusion reports an average F1 of '0.678 ± 0.0147', which is approximately 2.2% relative error. These numbers should be aligned, and the calculation (absolute vs relative standard deviation) should be stated explicitly.
  4. [§IV.B and §IV.C] The main results are obtained at a single operating point with manually chosen thresholds: epsilon_D=5, epsilon_V=1.1, batch size=1000, and N_sub=4000. No sensitivity analysis is reported, so the claimed robustness to trajectory variation and the stated accuracy/efficiency trade-off are demonstrated only for one configuration. A small sensitivity study around these thresholds, or a discussion of how they were calibrated and how strongly they affect the results, would materially strengthen the paper.
minor comments (5)
  1. [§IV.A] The dataset description states that AE2000f comes from 'The University of Tokyo, UK' — the location should be Japan, not the UK.
  2. [§IV.B] The phrase 'initialised without priors' is imprecise for a DPGMM, which always uses a concentration parameter alpha. Clarify whether alpha is fixed, estimated, or chosen empirically, and report its value in the experimental setup.
  3. [§IV.C] The F1 curves in Fig. 4 are smoothed with a 4th-degree polynomial fit. This can obscure the final values and make cross-method comparisons visually misleading. Consider reporting final mean/standard-deviation values in a table as well.
  4. [§III.B.3] The notation for the representative-set size is inconsistent: the text uses N_sub, then N_S in Eq. (21). Use a single symbol throughout.
  5. [General] No code or data accessibility statement is provided. Since the method involves several tunable components, making at least a reference implementation available would support reproducibility.

Circularity Check

0 steps flagged

No circularity found: the method is an empirical pipeline evaluated against external ground-truth labels; the abstract's own F1 claim is internally inconsistent, but that is a correctness issue, not circularity.

full rationale

The paper does not derive its clustering result from its own inputs by construction. The OCF combines known components—DPGMM backbone, CHV merging, BIC/AIC splitting, and density-based representative selection—and evaluates the resulting cluster assignments against human-annotated semantic labels using a majority-voting F1 score. No equation in the paper defines the predicted clusters in terms of the ground truth, and the representative subset, merging thresholds, and splitting criteria are computed from latent-space statistics rather than from the evaluation labels. The only self-citations that are load-bearing (refs [13], [50]) supply a pretrained encoder and an HK-means sampling baseline; these are external to the present paper's clustering loop and their quality is falsifiable through the reported F1 scores. The encoder is not defined in terms of the OCF's output, so this is lineage, not circularity. The hand-set hyperparameters (ε_D=5, ε_V=1.1, N_sub=4000) are chosen from heuristic ranges and hardware limits; there is no evidence they were fitted to the test-set F1 values, so the fitted-input-called-prediction pattern does not apply. The abstract's claim that OCF achieves the highest F1 among online approaches is contradicted by the paper's own results in Section IV.C.1, where OC+HKmeans Sampling achieves 0.726 and OC+Density-based Sampling achieves 0.678; this is an internal inconsistency in the headline claim, not a circular derivation. Likewise, the reported uncertainty differs between the abstract (3%) and the conclusion (±0.0147). These are correctness and reporting flaws, outside the scope of circularity. Overall, no circular step was identified.

Axiom & Free-Parameter Ledger

5 free parameters · 4 axioms · 0 invented entities

The framework introduces no new entities; all components are statistical models and algorithms from prior work. The main free parameters are the calibration thresholds and the representative set size, all hand-set without sensitivity analysis. The strongest assumptions are about the encoder's feature space and the representativeness of the fixed-size summary.

free parameters (5)
  • epsilon_D (Mahalanobis distance merging threshold) = 5
    Calibration parameter determining minimum overlap for cluster merging; set to 5 within the recommended range 3-5, no sensitivity analysis.
  • epsilon_V (volume ratio threshold) = 1.1
    Calibration parameter limiting volume increase on merging; 'typically set close to 1.0', chosen without sensitivity analysis.
  • N_sub (representative set size) = 4000
    Set from the rule 200d-300d with d=16; chosen based on CPU memory limits, affecting split quality and runtime.
  • Batch size (clustering trigger) = 1000
    Clustering is invoked after every 1,000 new images; affects adaptation latency and compute.
  • DPGMM concentration parameter alpha
    Prior over cluster weights from a Gamma distribution; not specified in the paper, yet it influences the number of clusters.
axioms (4)
  • domain assumption Latent features within each semantic class are approximately Gaussian
    DPGMM and GMM split/merge criteria assume Gaussian components; if features are non-Gaussian, AIC/BIC and Mahalanobis merging may misbehave. Invoked in Sections III.A and III.B.
  • domain assumption Pretrained encoder from ref [50] yields a latent space aligned with human semantic labels
    The evaluation uses F1 against human labels; if the encoder's features do not track semantic boundaries, clusters will not match ground truth. Used throughout the method, especially Section III.B.
  • domain assumption A fixed-size representative set captures the full historical distribution
    The method summarizes history with N_sub=4000 samples; if new modes appear or minority classes are lost, splits/merges will be biased. Stated as a risk in Section III.B.3.
  • standard math AIC and BIC are valid model selection criteria for split decisions
    The split operation accepts a split if AIC/BIC improves, assuming the likelihood in Eq. 16 is computable under the Gaussian model. Standard statistics, but validity depends on the Gaussian assumption.

pith-pipeline@v1.3.0-alltime-deepseek · 12118 in / 11652 out tokens · 110032 ms · 2026-08-04T23:15:38.754817+00:00 · methodology

0 comments
Cite this review

Pith. "Pith review of Online Clustering of Seafloor Imagery for Interpretation during Long-Term AUV Operations." pith.science (2026). https://pith.science/paper/7IFQDZ6Z

@misc{pith2026250906678,
  author       = {Pith},
  title        = {Pith review of: Online Clustering of Seafloor Imagery for Interpretation during Long-Term AUV Operations},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/7IFQDZ6Z}},
  note         = {Machine review of arXiv:2509.06678}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

As long-endurance and seafloor-resident AUVs become more capable, there is an increasing need for extended, real-time interpretation of seafloor imagery to enable adaptive missions and optimise communication efficiency. Although offline image analysis methods are well established, they rely on access to complete datasets and human-labelled examples to manage the strong influence of environmental and operational conditions on seafloor image appearance-requirements that cannot be met in real-time settings. To address this, we introduce an online clustering framework (OCF) capable of interpreting seafloor imagery without supervision, which is designed to operate in real-time on continuous data streams in a scalable, adaptive, and self-consistent manner. The method enables the efficient review and consolidation of common patterns across the entire data history in constant time by identifying and maintaining a set of representative samples that capture the evolving feature distribution, supporting dynamic cluster merging and splitting without reprocessing the full image history. We evaluate the framework on three diverse seafloor image datasets, analysing the impact of different representative sampling strategies on both clustering accuracy and computational cost. The OCF achieves the highest average F1 score of 0.68 across the three datasets among all comparative online clustering approaches, with a standard deviation of 3% across three distinct survey trajectories, demonstrating its superior clustering capability and robustness to trajectory variation. In addition, it maintains consistently lower and bounded computational time as the data volume increases. These properties are beneficial for generating survey data summaries and supporting informative path planning in long-term, persistent autonomous marine exploration.

Figures

Figures reproduced from arXiv: 2509.06678 by Adrian Bodenmann, Blair Thornton, Cailei Liang, Sam Fenton.

Figure 1
Figure 1. Figure 1: The OCF incrementally assigns images to a history-maintained cluster set [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Image examples and class proportions of three field survey datasets used in this paper. [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Class distributions in each cluster in the DM, CB and SH datasets. [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: F1 scores of different clustering methods across various seafloor datasets in DM, CB and SH from left to right. Results are smoothed using a [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Time cost of different clustering methods across various seafloor datasets. Results are smoothed using a 4th-degree polynomial fit for clearer [PITH_FULL_IMAGE:figures/full_fig_p008_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Comparison of the ground truth and the inferred clustering results [PITH_FULL_IMAGE:figures/full_fig_p008_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Comparison of the ground truth and the inferred clustering results [PITH_FULL_IMAGE:figures/full_fig_p009_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Comparison of the ground truth and the inferred clustering results [PITH_FULL_IMAGE:figures/full_fig_p009_8.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Long-Term Visual Localization in Dynamic Benthic Environments: A Dataset, Footprint-Based Ground Truth, and Visual Place Recognition Benchmark

    cs.CV 2026-03 conditional novelty 6.0

    A benchmark shows state-of-the-art visual place recognition performs poorly on a new multi-site, multi-year benthic AUV dataset, and that distance-based ground truth inflates recall.

Reference graph

Works this paper leans on

51 extracted references · 42 canonical work pages · cited by 1 Pith paper · 3 internal anchors

  1. [1]

    Monitoring of benthic reference sites: Using an autonomous underwater vehicle,

    S. B. Williams, O. R. Pizarro, M. V . Jakuba, C. R. Johnson, N. S. Barrett, R. C. Babcock, G. A. Kendrick, P. D. Steinberg, A. J. Heyward, P. J. Do- herty, I. Mahon, M. Johnson-Roberson, D. Steinberg, and A. Friedman, “Monitoring of benthic reference sites: Using an autonomous underwater vehicle,”IEEE Robotics & Automation Magazine, vol. 19, no. 1, pp. 73...

    work page 2012

  2. [2]

    Inspection and maintenance of indus- trial infrastructure with autonomous underwater robots,

    F. Nauert and P. Kampmann, “Inspection and maintenance of indus- trial infrastructure with autonomous underwater robots,”Frontiers in Robotics and AI, vol. 10, p. 1240276, 2023

    work page 2023

  3. [3]

    Visually mapping the rms titanic: Conservative covariance estimates for slam information filters,

    R. M. Eustice, H. Singh, J. J. Leonard, and M. R. Walter, “Visually mapping the rms titanic: Conservative covariance estimates for slam information filters,”The International Journal of Robotics Research, vol. 25, no. 12, pp. 1223–1242, 2006. [Online]. Available: https://doi.org/10.1177/0278364906072512

    work page doi:10.1177/0278364906072512 2006

  4. [4]

    Hierarchical bayesian models for unsupervised scene understanding,

    D. M. Steinberg, O. Pizarro, and S. B. Williams, “Hierarchical bayesian models for unsupervised scene understanding,”Computer Vision and Image Understanding, vol. 131, pp. 128–144, 2015

    work page 2015

  5. [5]

    GeoCLR: Georeference Contrastive Learning for Efficient Seafloor Image Interpretation

    T. Yamada, A. Pr ¨ugel-Bennett, S. B. Williams, O. Pizarro, and B. Thorn- ton, “Geoclr: Georeference contrastive learning for efficient seafloor image interpretation,”arXiv preprint arXiv:2108.06421, 2021

    work page internal anchor Pith review Pith/arXiv arXiv 2021

  6. [6]

    Generation of high-resolution three-dimensional reconstructions of the seafloor in color using a single camera and structured light,

    A. Bodenmann, B. Thornton, and T. Ura, “Generation of high-resolution three-dimensional reconstructions of the seafloor in color using a single camera and structured light,”Journal of Field Robotics, vol. 34, no. 5, pp. 833–851, 2017

    work page 2017

  7. [7]

    Assessing the repeatability of automated seafloor classification algorithms, with application in marine protected area monitoring,

    A. Zelada Leon, V . A. Huvenne, N. M. Benoist, M. Ferguson, B. J. Bett, and R. B. Wynn, “Assessing the repeatability of automated seafloor classification algorithms, with application in marine protected area monitoring,”Remote Sensing, vol. 12, no. 10, 2020. [Online]. Available: https://www.mdpi.com/2072-4292/12/10/1572

    work page 2020

  8. [8]

    Auv adaptive sampling methods: A review,

    J. Hwang, N. Bose, and S. Fan, “Auv adaptive sampling methods: A review,”Applied Sciences, vol. 9, no. 15, p. 3145, 2019

    work page 2019

  9. [9]

    Capture: A communications architecture for progressive transmission via underwater relays with eavesdropping,

    C. Murphy, J. M. Walls, T. Schneider, R. M. Eustice, M. Stojanovic, and H. Singh, “Capture: A communications architecture for progressive transmission via underwater relays with eavesdropping,”IEEE Journal of Oceanic Engineering, vol. 39, no. 1, pp. 120–130, 2014

    work page 2014

  10. [10]

    Dinov2: Learning robust visual features without supervision,

    M. Oquab, T. Darcet, T. Moutakanni, H. V o, M. Szafraniec, V . Khalidov, P. Fernandez, D. Haziza, F. Massa, A. El-Nouby,et al., “Dinov2: Learning robust visual features without supervision,”arXiv preprint arXiv:2304.07193, 2023

    Pith/arXiv arXiv 2023

  11. [11]

    A standardised vocabulary for identifying benthic biota and substrata from underwater imagery: The catami classification scheme,

    F. Althaus, N. Hill, R. Ferrari, L. Edwards, R. Przeslawski, C. H. L. Sch ¨onberg, R. Stuart-Smith, N. Barrett, G. Edgar, J. Colquhoun, M. Tran, A. Jordan, T. Rees, and K. Gowlett-Holmes, “A standardised vocabulary for identifying benthic biota and substrata from underwater imagery: The catami classification scheme,”PLOS ONE, vol. 10, no. 10, pp. 1–18, 10...

    doi:10.1371/journal.pone.0141039 2015

  12. [12]

    Gear-induced concept drift in marine images and its effect on deep learning classification,

    D. Langenk ¨amper, R. van Kevelaer, A. Purser, and T. W. Nattkemper, “Gear-induced concept drift in marine images and its effect on deep learning classification,”Frontiers in Marine Science, vol. V olume 7 - 2020, 2020. [Online]. Available: https://www.frontiersin.org/journals/ marine-science/articles/10.3389/fmars.2020.00506

    work page arXiv 2020

  13. [13]

    Guiding labelling effort for efficient learn- ing with georeferenced images,

    T. Yamada, M. Massot-Campos, A. Pr ¨ugel-Bennett, O. Pizarro, S. B. Williams, and B. Thornton, “Guiding labelling effort for efficient learn- ing with georeferenced images,”IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 45, no. 1, pp. 593–607, 2022

    work page 2022

  14. [14]

    K-means clustering algorithms: A comprehensive review, variants analysis, and advances in the era of big data,

    A. M. Ikotun, A. E. Ezugwu, L. Abualigah, B. Abuhaija, and J. Hem- ing, “K-means clustering algorithms: A comprehensive review, variants analysis, and advances in the era of big data,”Information Sciences, vol. 622, pp. 178–210, 2023

    2023

  15. [15]

    A survey of density based clustering algorithms,

    P. Bhattacharjee and P. Mitra, “A survey of density based clustering algorithms,”Frontiers of Computer Science, vol. 15, pp. 1–27, 2021

    work page 2021

  16. [16]

    Modeling unsupervised empirical adaptation by dpgmm and dpgmm-rnn hybrid model to extract perceptual features for low-resource asr,

    B. Wu, S. Sakti, J. Zhang, and S. Nakamura, “Modeling unsupervised empirical adaptation by dpgmm and dpgmm-rnn hybrid model to extract perceptual features for low-resource asr,”IEEE/ACM Transactions on Audio, Speech, and Language Processing, vol. 30, pp. 901–916, 2022

    work page 2022

  17. [17]

    Emerging properties in self-supervised vision transformers,

    M. Caron, H. Touvron, I. Misra, H. J ´egou, J. Mairal, P. Bojanowski, and A. Joulin, “Emerging properties in self-supervised vision transformers,” Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), pp. 9650–9660, 2021

    work page 2021

  18. [18]

    Learning features from georeferenced seafloor imagery with location guided autoencoders,

    T. Yamada, A. Pr ¨ugel-Bennett, and B. Thornton, “Learning features from georeferenced seafloor imagery with location guided autoencoders,” Journal of Field Robotics, vol. 38, no. 1, pp. 52–67, 2021

    work page 2021

  19. [19]

    Biological underpinnings for lifelong learning machines,

    D. Kudithipudi, M. Aguilar-Simon, J. Babb, M. Bazhenov, D. Black- iston, J. Bongard, A. P. Brna, S. Chakravarthi Raja, N. Cheney, J. Clune, et al., “Biological underpinnings for lifelong learning machines,”Nature Machine Intelligence, vol. 4, no. 3, pp. 196–210, 2022

    work page 2022

  20. [20]

    What and how: generalized lifelong spectral clustering via dual memory,

    G. Sun, Y . Cong, J. Dong, Y . Liu, Z. Ding, and H. Yu, “What and how: generalized lifelong spectral clustering via dual memory,”IEEE transactions on pattern analysis and machine intelligence, vol. 44, no. 7, pp. 3895–3908, 2021

    work page 2021

  21. [21]

    Large scale incremental learning,

    Y . Wu, Y . Chen, L. Wang, Y . Ye, Z. Liu, Y . Guo, and Y . Fu, “Large scale incremental learning,” inProceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2019, pp. 374–382

    2019

  22. [22]

    Continual unsupervised representation learning,

    D. Rao, F. Visin, A. Rusu, R. Pascanu, Y . W. Teh, and R. Hadsell, “Continual unsupervised representation learning,”Advances in neural information processing systems, vol. 32, 2019

    work page 2019

  23. [23]

    Experi- ence replay for continual learning,

    D. Rolnick, A. Ahuja, J. Schwarz, T. Lillicrap, and G. Wayne, “Experi- ence replay for continual learning,” inAdvances in Neural Information Processing Systems (NeurIPS), 2019, pp. 350–360

    work page 2019

  24. [24]

    Lifelong bayesian learning machines for streaming industrial big data,

    Z. Yang, J. Zheng, and Z. Ge, “Lifelong bayesian learning machines for streaming industrial big data,”IEEE Transactions on Systems, Man, and Cybernetics: Systems, vol. 53, no. 3, pp. 1554–1565, 2022

    work page 2022

  25. [25]

    An incremental dpmm-based method for trajectory clustering, modeling, and retrieval,

    W. Hu, X. Li, G. Tian, S. Maybank, and Z. Zhang, “An incremental dpmm-based method for trajectory clustering, modeling, and retrieval,” IEEE transactions on pattern analysis and machine intelligence, vol. 35, no. 5, pp. 1051–1065, 2013

    work page 2013

  26. [26]

    Tiny episodic memories in continual learning,

    A. Chaudhry, M. Rohrbach, M. Elhoseiny, T. Ajanthan, P. H. Torr, and P. K. Dokania, “Tiny episodic memories in continual learning,” inProceedings of the International Conference on Machine Learning (ICML), 2019, pp. 1954–1963

    work page 2019

  27. [27]

    Incremental Learning with Maximum Entropy Regularization: Rethinking Forgetting and Intransigence

    D. Kim, J. Bae, Y . Jo, and J. Choi, “Incremental learning with maximum entropy regularization: Rethinking forgetting and intransigence,”arXiv preprint arXiv:1902.00829, 2019

    work page internal anchor Pith review Pith/arXiv arXiv 1902

  28. [28]

    Knowledge distillation: A survey,

    J. Gou, B. Yu, S. J. Maybank, and D. Tao, “Knowledge distillation: A survey,”International Journal of Computer Vision, vol. 129, no. 6, pp. 1789–1819, 2021

    2021

  29. [29]

    Distilling the knowledge in a neural network,

    G. Hinton, O. Vinyals, and J. Dean, “Distilling the knowledge in a neural network,”arXiv preprint arXiv:1503.02531, 2015

    Pith/arXiv arXiv 2015

  30. [30]

    Ensemble Knowledge Distillation for Learning Improved and Efficient Networks

    U. Asif, J. Tang, and S. Harrer, “Ensemble knowledge distilla- tion for learning improved and efficient networks,”arXiv preprint arXiv:1909.08097, 2019

    work page internal anchor Pith review Pith/arXiv arXiv 1909

  31. [31]

    Feature fusion for online mutual knowledge distillation,

    J. Kim, M. Hyun, I. Chung, and N. Kwak, “Feature fusion for online mutual knowledge distillation,” in2020 25th International Conference on Pattern Recognition (ICPR). IEEE, 2021, pp. 4619–4625

    work page 2021

  32. [32]

    Incremental learning for semantic segmentation of large-scale remote sensing data,

    O. Tasar, Y . Tarabalka, and P. Alliez, “Incremental learning for semantic segmentation of large-scale remote sensing data,”IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, vol. 12, no. 9, pp. 3524–3537, 2019

    work page 2019

  33. [33]

    Autonomous data stream clustering implementing split-and-merge concepts–towards a plug-and- play approach,

    E. Lughofer and M. Sayed-Mouchaweh, “Autonomous data stream clustering implementing split-and-merge concepts–towards a plug-and- play approach,”Information Sciences, vol. 304, pp. 54–79, 2015

    work page 2015

  34. [34]

    Sampling in dirichlet process mixture models for clustering streaming data,

    O. Dinari and O. Freifeld, “Sampling in dirichlet process mixture models for clustering streaming data,” inInternational Conference on Artificial Intelligence and Statistics. PMLR, 2022, pp. 818–835

    work page 2022

  35. [35]

    Deepdpm: Deep clustering with an unknown number of clusters,

    M. Ronen, S. E. Finder, and O. Freifeld, “Deepdpm: Deep clustering with an unknown number of clusters,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2022, pp. 9861–9870

    work page 2022

  36. [36]

    Cluster merging and splitting in hierarchical clustering algorithms,

    C. Ding and X. He, “Cluster merging and splitting in hierarchical clustering algorithms,” in2002 IEEE International Conference on Data Mining, 2002. Proceedings.IEEE, 2002, pp. 139–146

    work page 2002

  37. [37]

    Aic model selection using akaike weights,

    E.-J. Wagenmakers and S. Farrell, “Aic model selection using akaike weights,”Psychonomic bulletin & review, vol. 11, pp. 192–196, 2004

    work page 2004

  38. [38]

    Generative adversarial networks: An overview,

    A. Creswell, T. White, V . Dumoulin, K. Arulkumaran, B. Sengupta, and A. A. Bharath, “Generative adversarial networks: An overview,”IEEE signal processing magazine, vol. 35, no. 1, pp. 53–65, 2018

    2018

  39. [39]

    Dataset distillation: A comprehensive re- view,

    R. Yu, S. Liu, and X. Wang, “Dataset distillation: A comprehensive re- view,”IEEE Transactions on Pattern Analysis and Machine Intelligence, 2023

    work page 2023

  40. [40]

    Anchor assisted experience replay for online class-incremental learning,

    H. Lin, S. Feng, X. Li, W. Li, and Y . Ye, “Anchor assisted experience replay for online class-incremental learning,”IEEE Transactions on Circuits and Systems for Video Technology, 2022

    work page 2022

  41. [41]

    Data distillation: A survey,

    N. Sachdeva and J. McAuley, “Data distillation: A survey,”arXiv preprint arXiv:2301.04272, 2023

    Pith/arXiv arXiv 2023

  42. [42]

    A comprehensive survey of dataset distillation,

    S. Lei and D. Tao, “A comprehensive survey of dataset distillation,” IEEE Transactions on Pattern Analysis and Machine Intelligence, 2023

    work page 2023

  43. [43]

    Spatially balanced sampling through the pivotal method,

    A. Grafstr ¨om, N. L. Lundstr ¨om, and L. Schelin, “Spatially balanced sampling through the pivotal method,”Biometrics, vol. 68, no. 2, pp. 514–520, 2012

    work page 2012

  44. [44]

    Spatial autocorrelation and neighborhood quality,

    R. A. Dubin, “Spatial autocorrelation and neighborhood quality,”Re- gional science and urban economics, vol. 22, no. 3, pp. 433–452, 1992

    work page 1992

  45. [45]

    Determining represen- tative sample size for validation of continuous, large continental remote sensing data,

    M. L. Blatchford, C. M. Mannaerts, and Y . Zeng, “Determining represen- tative sample size for validation of continuous, large continental remote sensing data,”International Journal of Applied Earth Observation and Geoinformation, vol. 94, p. 102235, 2021

    work page 2021

  46. [46]

    Representative subset selection,

    M. Daszykowski, B. Walczak, and D. Massart, “Representative subset selection,”Analytica chimica acta, vol. 468, no. 1, pp. 91–103, 2002

    work page 2002

  47. [47]

    Beta processes, stick- breaking and power laws,

    T. Broderick, M. I. Jordan, and J. Pitman, “Beta processes, stick- breaking and power laws,” 2012

    work page 2012

  48. [48]

    A dynamic split-and-merge approach for evolving cluster models,

    E. Lughofer, “A dynamic split-and-merge approach for evolving cluster models,”Evolving systems, vol. 3, no. 3, pp. 135–151, 2012

    work page 2012

  49. [49]

    Unsupervised learning of visual features by contrasting cluster assign- ments,

    M. Caron, I. Misra, J. Mairal, P. Goyal, P. Bojanowski, and A. Joulin, “Unsupervised learning of visual features by contrasting cluster assign- ments,”Advances in neural information processing systems, vol. 33, pp. 9912–9924, 2020

    2020

  50. [50]

    Self-supervised learning with multimodal remote sensed maps for seafloor visual class inference,

    C. Liang, J. Cappelletto, M. Massot-Campos, A. Bodenmann, V . A. Huvenne, C. Wardell, B. J. Bett, D. Newborough, and B. Thornton, “Self-supervised learning with multimodal remote sensed maps for seafloor visual class inference,”The International Journal of Robotics Research, p. 02783649251343640, 2025

    work page 2025

  51. [51]

    Investigating location-regularised self-supervised feature learning for seafloor visual imagery,

    C. Liang, A. Bodenmann, E. J. Curtis, S. Simmons, K. Nagano, S. Brown, A. Riese, and B. Thornton, “Investigating location-regularised self-supervised feature learning for seafloor visual imagery,” 2025, submitted to IEEE Journal of Oceanic Engineering

    work page 2025