pith. sign in

arxiv: 2606.21743 · v1 · pith:DNDC22IYnew · submitted 2026-06-19 · 💻 cs.MA · cs.AI

Cohort Organized Learning: Clustering Through Agreement

Finn Henry O'Shea , Maria Elena Monzani This is my paper

Pith reviewed 2026-06-26 12:23 UTC · model grok-4.3

classification 💻 cs.MA cs.AI
keywords clusteringneural networksexpectation maximizationunsupervised learningagreement objectivecohort organized learningvector dataimage clustering
0
0 comments X

The pith

CoOL clusters data by training neural networks to maximize agreement among cohort assignments without using distance or similarity measures.

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

The paper presents Cohort Organized Learning as a clustering technique that lets neural networks learn group assignments directly from data by optimizing an agreement objective. Gradients for this optimization come from expectation maximization, and the approach is shown to work on both vector data and images. A reader would care because this removes the need to define or compute explicit similarities, which often fail or require heavy preprocessing for complex inputs like images.

Core claim

CoOL performs clustering by feeding data into neural networks whose outputs define cohort memberships; these networks are trained by maximizing an agreement objective whose gradients are derived from the expectation-maximization algorithm, thereby producing cluster assignments for any data type compatible with the networks, illustrated on vectors and images.

What carries the argument

The agreement objective optimized by EM-derived gradients on neural networks, which produces stable cohort assignments by encouraging consistent predictions across data points.

If this is right

  • Clustering becomes possible for data where no natural distance function exists.
  • The same networks can be applied directly to images without first extracting features or similarities.
  • Training progress can be tracked by monitoring the agreement objective and convergence criteria described in the method.
  • Cluster quality can be assessed after training using the evaluation procedures outlined.

Where Pith is reading between the lines

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

  • The approach may combine with other neural unsupervised objectives to handle mixed data types in one pipeline.
  • It opens a route to clustering tasks where similarity is defined implicitly by the network architecture rather than by the user.
  • Extensions could test whether the same agreement mechanism improves semi-supervised settings with partial labels.

Load-bearing premise

That training neural networks on the agreement objective will yield non-degenerate, meaningful cluster assignments rather than arbitrary or collapsed groupings.

What would settle it

Run CoOL on a standard labeled dataset such as MNIST and check whether the resulting clusters align with the known digit labels better than random assignment.

Figures

Figures reproduced from arXiv: 2606.21743 by Finn Henry O'Shea, Maria Elena Monzani.

Figure 1
Figure 1. Figure 1: Diagram of data flow through CoOL. The solid lines show relationships that allow back propagation. number of experts. In doing so, they find estimates for the proportion of samples that belong to each class and the “error-rate matrix” (what we will call the reliability matrix) for each observer. CoOL modifies this approach, it uses EM and differentiable observers capable of learning to force the observers … view at source ↗
Figure 2
Figure 2. Figure 2: Data for clustering. The different colors and shapes indicate class membership of the samples. observers to agree on the cluster labels for each sample is to agree that all of the samples belong to a single cluster. This is both a strong attractor during optimization and not useful. This degenerate case has an easy-to-identify signature in that the reliability matrix for each of the observers has only a si… view at source ↗
Figure 3
Figure 3. Figure 3: Agreement rate (blue, solid line and circles) and number of unique labels (orange, dashed line) during training on the clustering problem shown in [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Results of clustering by CoOL. The different colors and shapes indicate cluster membership of the samples. The filled shapes show samples that were correctly clustered together while the unfilled shapes show samples that were incorrectly clustered together. For the latter, the color represents the cluster CoOL assigned the sample to while the shape represents the true cluster identity. scaling. Two example… view at source ↗
Figure 5
Figure 5. Figure 5: (top) Loss as function of epoch when learning the problem shown in [PITH_FULL_IMAGE:figures/full_fig_p008_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Plot of accuracy versus agreement rate for various hyperparameter settings on the MNIST digit problem. The different markers indicate the learning rate used: circles mean 10−5 , squares mean 10−4 , and triangles mean 10−3 . Filled markers indicate that a weight decay of 10−4 was used while empty markers mean no weight decay was used. For this hyperparameter scan we used 288 samples per class. where N is th… view at source ↗
Figure 7
Figure 7. Figure 7: Determinant of the π (k) matrix for all 5 observers during select runs of CoOL on a varying number of digits with a varying number of classes (shown in the legends). (left) shows a linear scale and (right) shows a log scale. For completeness, these plots show training through 5000 epochs regardless of when early stopping occurred. tends to be high when the agreement rate is high. The exception is when the … view at source ↗
Figure 8
Figure 8. Figure 8: The fraction of groups containing the given number of samples when the features of the MNIST digits are shuffled on a per-sample basis. The distribution is very close to random assignment of the samples to the groups. the algorithm, if some set of the samples are more similar to each other than they are to some other set of samples. Therefore, if CoOL is clustering in some sensible and consistent way, more… view at source ↗
Figure 9
Figure 9. Figure 9: (top) The agreement rate for all digits (and samples) CoOL was trained or validated on. (bottom) The agreement and disagreement rate for samples and digits not used during training. The digits to the left of the dashed vertical line were used for training. u t = ∆ − δt δu − ∆ . (13) If we assume, as seems reasonable from [PITH_FULL_IMAGE:figures/full_fig_p014_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Statistics for training time and memory consumption when training CoOL with convolutional layers. In the top row we show time to complete training and in the bottom row we show memory allocated by pytorch during training [60]. The left column shows batch training and the right column shows mini-batch training. The colored curves indicate the number of networks, K, used to learn 5 classes of the MNIST data… view at source ↗
Figure 11
Figure 11. Figure 11: Accuracy of the CoOL-trained convolutional networks on a hold-out set of 5 MINST digits as a function of the number of networks. Training is stopped when the agreement rate reaches 99.5% or 5000 epochs, which ever comes first. 7.3 Ablation Study Herein we investigate the learning properties of CoOL by making some critical changes to the algorithm. We vary the values of λ and α, defined in Section 4. We su… view at source ↗
Figure 12
Figure 12. Figure 12: Accuracy of the CoOL-trained convolutional networks on a hold-out set of MINST images as a function of the weight of the determinant term, λ. The box shows the settings used to produce [PITH_FULL_IMAGE:figures/full_fig_p018_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: Accuracy of the CoOL-trained convolutional networks on a hold-out set of MINST images as a function of the number of epochs. Training is stopped when the agreement rate reaches 99.5% or 5000 epochs, which ever comes first. The empty symbols represent models trained using Kullback–Leibler divergence without sampling and the full symbols are models trained using negative log-likelihood loss as in Section 4.… view at source ↗
Figure 14
Figure 14. Figure 14: Composite images for the 6 largest groups shown in [PITH_FULL_IMAGE:figures/full_fig_p020_14.png] view at source ↗
read the original abstract

In this article we describe Cohort Organized Learning (CoOL), a method for clustering data without explicit distance or similarity computations. Herein, we will describe CoOL, derive the gradients determined by expectation maximization to train the networks, show how to monitor convergence during training and evaluate the clusters after training, and discuss a series of examples and use cases. We also discuss CoOL's limitations and future prospects on related tasks. Because CoOL uses neural networks to estimate the clusters, it can be used to cluster any data that can be made compatible and we illustrate this on vector data and images.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The paper introduces Cohort Organized Learning (CoOL), a clustering approach that trains neural networks to assign data points to cohorts by optimizing an agreement objective whose gradients are derived via expectation maximization (EM). It claims this avoids explicit distance or similarity computations, applies to vector data and images, includes derivations of the gradients, methods for monitoring convergence and evaluating clusters post-training, example use cases, and discussion of limitations and future work.

Significance. If the central claim holds and the optimization yields non-degenerate, data-driven partitions, CoOL would offer a distance-free clustering method that exploits neural network flexibility for heterogeneous inputs. The explicit derivation of EM gradients and convergence monitoring would be strengths, as would any reproducible code or falsifiable predictions on standard benchmarks.

major comments (2)
  1. [Abstract / gradient derivation] Abstract and gradient derivation section: the central claim that CoOL produces meaningful clusters rests on the unexamined assumption that fixed points of the agreement objective correspond to semantically coherent partitions rather than trivial solutions (e.g., all points assigned to one cohort). No analysis, fixed-point characterization, or regularization term is supplied to exclude such attractors, which directly undermines the claim that the method clusters data without explicit distances.
  2. [Examples and evaluation] Examples and evaluation section: the manuscript states it will 'evaluate the clusters after training' and 'discuss a series of examples,' yet the provided abstract supplies neither quantitative recovery metrics on labeled benchmarks (e.g., adjusted Rand index on MNIST or UCI datasets) nor ablation showing that performance exceeds chance or collapsed assignments. Without such evidence the empirical support for non-degenerate behavior is absent.
minor comments (1)
  1. [Abstract] The abstract promises derivations, convergence monitoring, and use-case discussions but contains no equations, pseudocode, or numerical results; moving at least one key gradient expression or convergence criterion into the abstract would improve readability.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the detailed and constructive report. The two major comments identify important gaps in theoretical analysis and empirical validation. We respond to each below and commit to revisions where the manuscript is incomplete.

read point-by-point responses

  1. Referee: [Abstract / gradient derivation] Abstract and gradient derivation section: the central claim that CoOL produces meaningful clusters rests on the unexamined assumption that fixed points of the agreement objective correspond to semantically coherent partitions rather than trivial solutions (e.g., all points assigned to one cohort). No analysis, fixed-point characterization, or regularization term is supplied to exclude such attractors, which directly undermines the claim that the method clusters data without explicit distances.

    Authors: We agree that the manuscript lacks a fixed-point analysis or explicit safeguards against collapse. The current text derives the EM gradients and describes convergence monitoring but does not characterize the attractors of the agreement objective. In revision we will add a dedicated subsection that (i) identifies conditions under which the all-in-one-cohort assignment is or is not a fixed point and (ii) shows how the per-cohort network parameterization and the monitoring procedure reduce the likelihood of trivial solutions. We will also examine whether a simple entropy or diversity regularizer is needed. revision: yes

  2. Referee: [Examples and evaluation] Examples and evaluation section: the manuscript states it will 'evaluate the clusters after training' and 'discuss a series of examples,' yet the provided abstract supplies neither quantitative recovery metrics on labeled benchmarks (e.g., adjusted Rand index on MNIST or UCI datasets) nor ablation showing that performance exceeds chance or collapsed assignments. Without such evidence the empirical support for non-degenerate behavior is absent.

    Authors: The full manuscript contains an evaluation procedure and illustrative examples, but these are qualitative. We accept that quantitative recovery metrics (adjusted Rand index, normalized mutual information) on standard labeled benchmarks and explicit checks against chance or collapsed assignments are required to substantiate non-degenerate behavior. We will add these experiments, including ablations on network capacity and initialization, to the revised manuscript. revision: yes

Circularity Check

0 steps flagged

No circularity detected; derivation chain not reducible to inputs from available text

full rationale

The abstract outlines a derivation of EM-based gradients for training on an agreement objective and claims clustering without explicit distances, but supplies no equations, fitted parameters presented as predictions, or self-citations that could be inspected for self-definitional, load-bearing, or ansatz-smuggling patterns. No load-bearing step is visible that reduces by construction to its own inputs, and the central claim of producing meaningful clusters rests on the optimization behavior rather than a tautological renaming or fit. The provided text is therefore self-contained against external benchmarks for the purpose of this circularity check.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Only the abstract is available; no free parameters, axioms, or invented entities are described.

pith-pipeline@v0.9.1-grok · 5612 in / 864 out tokens · 22279 ms · 2026-06-26T12:23:35.699954+00:00 · methodology

discussion (0)

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

Reference graph

Works this paper leans on

66 extracted references · 17 canonical work pages · 8 internal anchors

  1. [1]

    DIII-D data management,

    B. McHarg, J. Burruss, J. Freeman, C. Parker, J. Schachter, and D. Schissel, “DIII-D data management,”Fusion Engineering and Design, vol. 60, pp. 305–309, June 2002

    2002

  2. [2]

    Iterdb—the data archiving system for iter,

    G. Abla, G. Heber, D. P. Schissel, D. Robinson, L. Abadie, A. Wallander, and S. M. Flanagan, “Iterdb—the data archiving system for iter,”Fusion Engineering and Design, vol. 89, no. 5, pp. 536–541, 2014. Proceedings of the 9th IAEA Technical Meeting on Control, Data Acquisition, and Remote Participation for Fusion Research

    2014

  3. [3]

    How to manage 2 petabytes of new data every day

    P. Brans, “How to manage 2 petabytes of new data every day.” https://www.iter.org/node/20687/how-manage-2-petabytes-new-data-every-day. Accessed: 2025-12-09

    2025

  4. [4]

    The Data Acquisition System of the LZ Dark Matter Detector: FADR,

    J. Aalbers, D. S. Akerib,et al., “The Data Acquisition System of the LZ Dark Matter Detector: FADR,” Aug. 2024. arXiv:2405.14732 [physics]. 20 IOP PublishingJournalvv(yyyy) aaaaaa Authoret al

    work page arXiv 2024

  5. [5]

    The xenon program for dark matter direct detection,

    E. Aprile, “The xenon program for dark matter direct detection,”Nuclear Physics B, vol. 1003, p. 116463, 2024. Special Issue of Nobel Symposium 182 on Dark Matter

    2024

  6. [6]

    Massive Scale Data Analytics at LCLS-II,

    J. Thayer, Z. Chen, R. Claus, D. Damiani, C. Ford, M. Dubrovin, V. Elmir, W. Kroeger, X. Li, S. Marchesini, V. Mariani, R. Melcchiori, S. Nelson, A. Peck, A. Perazzo, F. Poitevin, C. P. O’Grady, J. Otero, O. Quijano, M. Shankar, M. Uervirojnangkoorn, R. Veraldi, M. Weaver, C. Weninger, S. Yamajala, C. Wang, and C. H. Yoon, “Massive Scale Data Analytics at...

    2024

  7. [7]

    Data management at the advanced photon source,

    N. Schwarz, S. Veseli, and D. Jarosz, “Data management at the advanced photon source,” Synchrotron Radiation News, vol. 32, no. 3, pp. 13–18, 2019

    2019

  8. [8]

    LSST: from Science Drivers to Reference Design and Anticipated Data Products

    Z. Ivezic, S. M. Kahn,et al., “LSST: from Science Drivers to Reference Design and Anticipated Data Products,”The Astrophysical Journal, vol. 873, p. 111, Mar. 2019. arXiv:0805.2366 [astro-ph]

    work page internal anchor Pith review Pith/arXiv arXiv 2019

  9. [9]

    Prototyping a Bulk Data Management System for CTAO with Rucio,

    S. Hasan, M. Allaux, A. Biland, F. Gillardo, H. Li, M. Linhoff, E. Lyard, M. Pierre, V. Savchenko, and R. Walter, “Prototyping a Bulk Data Management System for CTAO with Rucio,” inProceedings of 39th International Cosmic Ray Conference, vol. 501, (Geneva, Switzerland), p. 672, July 2025

    2025

  10. [10]

    Expected performance of the atlas experiment - detector, trigger and physics,

    The ATLAS Collaboration, “Expected performance of the atlas experiment - detector, trigger and physics,” 2009

    2009

  11. [11]

    Scientific computing plan for the ECCE detector at the Electron Ion Collider,

    J. Bernauer, C. Dean,et al., “Scientific computing plan for the ECCE detector at the Electron Ion Collider,”Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, vol. 1047, p. 167859, Feb. 2023

    2023

  12. [12]

    James, D

    G. James, D. Witten, T. Hastie, and R. Tibshirani,An Introduction to Statistical Learning: with Applications in R. Springer, 2013

    2013

  13. [13]

    Data Clustering: A Review,

    A. K. Jain, M. Murty, and P. J. Flynn, “Data Clustering: A Review,”ACM Computing Surveys, vol. 31, pp. 264–323, Sept. 1999

    1999

  14. [14]

    A review of clustering techniques and developments,

    A. Saxena, M. Prasad, A. Gupta, N. Bharill, O. P. Patel, A. Tiwari, M. J. Er, W. Ding, and C.-T. Lin, “A review of clustering techniques and developments,”Neurocomputing, vol. 267, pp. 664–681, Dec. 2017

    2017

  15. [15]

    A Rapid Review of Clustering Algorithms,

    H. Yin, A. Aryani, S. Petrie, A. Nambissan, A. Astudillo, and S. Cao, “A Rapid Review of Clustering Algorithms,” Jan. 2024

    2024

  16. [16]

    Maximum Likelihood from Incomplete Data Via theEMAlgorithm,

    A. P. Dempster, N. M. Laird, and D. B. Rubin, “Maximum Likelihood from Incomplete Data Via theEMAlgorithm,”Journal of the Royal Statistical Society Series B: Statistical Methodology, vol. 39, pp. 1–22, Sept. 1977

    1977

  17. [17]

    The MNIST Database of Handwrit- ten Digits,

    Y. LeCun, C. Cortes, and C. J. C. Burges, “The MNIST Database of Handwrit- ten Digits,” 1998

    1998

  18. [18]

    Maximum likelihood estimation of observer error-rates using the em algorithm,

    A. P. Dawid and A. M. Skene, “Maximum likelihood estimation of observer error-rates using the em algorithm,”Applied Statistics, vol. 28, no. 1, p. 20, 1979

    1979

  19. [19]

    Empirical Evaluation of Rectified Activations in Convolutional Network

    B. Xu, N. Wang, T. Chen, and M. Li, “Empirical Evaluation of Rectified Activations in Convolutional Network,” Nov. 2015. arXiv:1505.00853 [cs]

    work page internal anchor Pith review Pith/arXiv arXiv 2015

  20. [20]

    DBSCAN Revisited, Revisited: Why and How You Should (Still) Use DBSCAN,

    E. Schubert, J. Sander, M. Ester, H. P. Kriegel, and X. Xu, “DBSCAN Revisited, Revisited: Why and How You Should (Still) Use DBSCAN,”ACM Transactions on Database Systems, vol. 42, pp. 1–21, Sept. 2017

    2017

  21. [21]

    OPTICS: Ordering Points to Identify the Clustering Structure,

    M. Ankerst, M. M. Breunig, H.-P. Kriegel, and J. Sander, “OPTICS: Ordering Points to Identify the Clustering Structure,”ACM SIGMOD Record, vol. 28, pp. 49–60, June 1999

    1999

  22. [22]

    A tutorial on spectral clustering,

    U. Von Luxburg, “A tutorial on spectral clustering,”Statistics and Computing, vol. 17, pp. 395–416, Dec. 2007

    2007

  23. [23]

    Hastie, R

    T. Hastie, R. Tibshirani, and J. Friedman,The Elements of Statistical Learning: Data Mining, Inference, and Prediction. Springer, 2017. 2nd edition. 21 IOP PublishingJournalvv(yyyy) aaaaaa Authoret al

    2017

  24. [24]

    Adam: A method for stochastic optimization,

    D. P. Kingma and J. Ba, “Adam: A method for stochastic optimization,” 2017

    2017

  25. [25]

    Support-vector networks,

    C. Cortes and V. Vapnik, “Support-vector networks,”Machine Learning, vol. 20, pp. 273–297, Sept. 1995

    1995

  26. [26]

    Sparse Subspace Clustering: Algorithm, Theory, and Applications

    E. Elhamifar and R. Vidal, “Sparse Subspace Clustering: Algorithm, Theory, and Applications,” Feb. 2013. arXiv:1203.1005 [cs]

    work page internal anchor Pith review Pith/arXiv arXiv 2013

  27. [27]

    A comprehensive survey on deep clustering: Taxonomy, challenges, and future directions,

    S. Zhou, H. Xu, Z. Zheng, J. Chen, Z. li, J. Bu, J. Wu, X. Wang, W. Zhu, and M. Ester, “A Comprehensive Survey on Deep Clustering: Taxonomy, Challenges, and Future Directions,” June 2022. arXiv:2206.07579 [cs]

    work page arXiv 2022

  28. [28]

    Improved Deep Embedded Clustering with Local Structure Preservation,

    X. Guo, L. Gao, X. Liu, and J. Yin, “Improved Deep Embedded Clustering with Local Structure Preservation,” inProceedings of the Twenty-Sixth International Joint Conference on Artificial Intelligence, (Melbourne, Australia), pp. 1753–1759, International Joint Conferences on Artificial Intelligence Organization, Aug. 2017

    2017

  29. [29]

    Deep clustering with concrete k-means,

    B. Gao, Y. Yang, H. Gouk, and T. M. Hospedales, “Deep clustering with concrete k-means,” Oct. 2019. arXiv:1910.08031 [cs]

    work page arXiv 2019

  30. [30]

    SPICE: Semantic Pseudo-labeling for Image Clustering,

    C. Niu, H. Shan, and G. Wang, “SPICE: Semantic Pseudo-labeling for Image Clustering,” IEEE Transactions on Image Processing, vol. 31, pp. 7264–7278, 2022. arXiv:2103.09382 [cs]

    work page arXiv 2022

  31. [31]

    arXiv preprint arXiv:2201.10000 , year=

    Z. Li, Y. Chen, Y. LeCun, and F. T. Sommer, “Neural Manifold Clustering and Embedding,” Jan. 2022. arXiv:2201.10000 [cs]

    work page arXiv 2022

  32. [32]

    XAI Beyond Classification: Interpretable Neural Clustering,

    X. Peng, Y. Li, I. W. Tsang, H. Zhu, J. Lv, and J. T. Zhou, “XAI Beyond Classification: Interpretable Neural Clustering,” Apr. 2022. arXiv:1808.07292 [cs]

    work page arXiv 2022

  33. [33]

    Interpretable Deep Clustering for Tabular Data,

    J. Svirsky and O. Lindenbaum, “Interpretable Deep Clustering for Tabular Data,” June 2024

    2024

  34. [34]

    Pseudo-supervised Deep Subspace Clustering,

    J. Lv, Z. Kang, X. Lu, and Z. Xu, “Pseudo-supervised Deep Subspace Clustering,”IEEE Transactions on Image Processing, vol. 30, pp. 5252–5263, 2021. arXiv:2104.03531 [cs]

    work page arXiv 2021

  35. [35]

    SpectralNet: Spectral Clustering using Deep Neural Networks,

    U. Shaham, K. Stanton, H. Li, B. Nadler, R. Basri, and Y. Kluger, “SpectralNet: Spectral Clustering using Deep Neural Networks,” Apr. 2018. arXiv:1801.01587 [stat]

    work page arXiv 2018

  36. [36]

    Joint Unsupervised Learning of Deep Representations and Image Clusters

    J. Yang, D. Parikh, and D. Batra, “Joint Unsupervised Learning of Deep Representations and Image Clusters,” June 2016. arXiv:1604.03628 [cs]

    work page internal anchor Pith review Pith/arXiv arXiv 2016

  37. [37]

    Efficient Deep Embedded Subspace Clustering,

    J. Cai, J. Fan, W. Guo, S. Wang, Y. Zhang, and Z. Zhang, “Efficient Deep Embedded Subspace Clustering,” in2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), (New Orleans, LA, USA), pp. 21–30, June 2022

    2022

  38. [38]

    UNSUPERVISED CLUSTERING USING PSEUDO-SEMI- SUPERVISED LEARNING,

    D. Gupta and R. Ramjee, “UNSUPERVISED CLUSTERING USING PSEUDO-SEMI- SUPERVISED LEARNING,” inProceedings of 8th International Conference on Learning Representations, (Addis Ababa, Ethiopia), 2020

    2020

  39. [39]

    A SURVEY OF CLUSTERING ENSEMBLE ALGORITHMS,

    S. Vega-Pons and J. Ruiz-Shulcloper, “A SURVEY OF CLUSTERING ENSEMBLE ALGORITHMS,”International Journal of Pattern Recognition and Artificial Intelligence, vol. 25, pp. 337–372, May 2011

    2011

  40. [40]

    Combining multiple weak clusterings,

    A. Topchy, A. Jain, and W. Punch, “Combining multiple weak clusterings,” inThird IEEE International Conference on Data Mining, (Melbourne, FL, USA), pp. 331–338, IEEE Comput. Soc, 2003

    2003

  41. [41]

    A Mixture Model for Clustering Ensembles,

    A. Topchy, A. K. Jain, and W. Punch, “A Mixture Model for Clustering Ensembles,” in Proceedings of the 2004 SIAM International Conference on Data Mining, pp. 379–390, Society for Industrial and Applied Mathematics, Apr. 2004

    2004

  42. [42]

    Clustering ensembles: models of consensus and weak partitions,

    A. Topchy, A. Jain, and W. Punch, “Clustering ensembles: models of consensus and weak partitions,”IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 27, pp. 1866–1881, Dec. 2005

    2005

  43. [43]

    Neural Expectation Maximization,

    K. Greff, S. v. Steenkiste, and J. Schmidhuber, “Neural Expectation Maximization,” Nov

  44. [44]

    Neural Expectation Maximization

    arXiv:1708.03498 [cs]. 22 IOP PublishingJournalvv(yyyy) aaaaaa Authoret al

    work page internal anchor Pith review Pith/arXiv arXiv

  45. [45]

    Using the EM Algorithm to Train Neural Networks: Misconceptions and a New Algorithm for Multiclass Classification,

    S.-K. Ng and G. McLachlan, “Using the EM Algorithm to Train Neural Networks: Misconceptions and a New Algorithm for Multiclass Classification,”IEEE Transactions on Neural Networks, vol. 15, pp. 738–749, May 2004

    2004

  46. [46]

    Neural mixture models with expectation-maximization for end-to-end deep clustering,

    D. Tissera, K. Vithanage, R. Wijesinghe, A. Xavier, S. Jayasena, S. Fernando, and R. Rodrigo, “Neural mixture models with expectation-maximization for end-to-end deep clustering,”Neurocomputing, vol. 505, pp. 249–262, Sept. 2022

    2022

  47. [47]

    Neural Network based Explicit Mixture Models and Expectation-maximization based Learning,

    D. Liu, M. T. Vu, S. Chatterjee, and L. K. Rasmussen, “Neural Network based Explicit Mixture Models and Expectation-maximization based Learning,” May 2020. arXiv:1907.13432 [cs]

    work page arXiv 2020

  48. [48]

    A Neural Expectation-Maximization Framework for Noisy Multi-Label Text Classification,

    J. Chen, R. Zhang, J. Xu, C. Hu, and Y. Mao, “A Neural Expectation-Maximization Framework for Noisy Multi-Label Text Classification,”IEEE Transactions on Knowledge and Data Engineering, vol. 35, pp. 10992–11003, Nov. 2023

    2023

  49. [49]

    Deep Canonical Correlation Analysis,

    G. Andrew, R. Arora, J. Bilmes, and K. Livescu, “Deep Canonical Correlation Analysis,” in Proceedings of the 30th International Conference on Machine Learning, vol. 28(3), pp. 1247–1255, 2013

    2013

  50. [50]

    An intercomparison of methods for finding coupled patterns in climate data,

    C. S. Bretherton, C. Smith, and J. M. Wallace, “An intercomparison of methods for finding coupled patterns in climate data,”Journal of Climate, vol. 5, no. 6, pp. 541 – 560, 1992

    1992

  51. [51]

    Vitac: Feature sharing between vision and tactile sensing for cloth texture recognition,

    S. Luo, W. Yuan, E. Adelson, A. G. Cohn, and R. Fuentes, “Vitac: Feature sharing between vision and tactile sensing for cloth texture recognition,” 2018

    2018

  52. [52]

    Coincident learning for unsupervised anomaly detection of scientific instruments,

    R. Humble, Z. Zhang, F. O’Shea, E. Darve, and D. Ratner, “Coincident learning for unsupervised anomaly detection of scientific instruments,”Machine Learning: Science and Technology, vol. 5, p. 035036, Sept. 2024

    2024

  53. [53]

    I. H. Witten and E. Frank,Data Mining: Practical Machine Learning Tools and Techniques. Morgan Kaufmann, 2nd ed., 2005

    2005

  54. [54]

    Comparing partitions,

    L. Hubert and P. Arabie, “Comparing partitions,”Journal of Classification, vol. 2, pp. 193–218, Dec. 1985

    1985

  55. [55]

    Scikit-learn: Machine learning in Python,

    F. Pedregosa, G. Varoquaux, A. Gramfort, V. Michel, B. Thirion, O. Grisel, M. Blondel, P. Prettenhofer, R. Weiss, V. Dubourg, J. Vanderplas, A. Passos, D. Cournapeau, M. Brucher, M. Perrot, and E. Duchesnay, “Scikit-learn: Machine learning in Python,”Journal of Machine Learning Research, vol. 12, pp. 2825–2830, 2011

    2011

  56. [56]

    Unsupervised Deep Embedding for Clustering Analysis

    J. Xie, R. Girshick, and A. Farhadi, “Unsupervised Deep Embedding for Clustering Analysis,” May 2016. arXiv:1511.06335 [cs]

    work page internal anchor Pith review Pith/arXiv arXiv 2016

  57. [57]

    Inductive Representation Learning on Large Graphs

    W. L. Hamilton, R. Ying, and J. Leskovec, “Inductive Representation Learning on Large Graphs,” Sept. 2018. arXiv:1706.02216 [cs]

    work page internal anchor Pith review Pith/arXiv arXiv 2018

  58. [58]

    Single-Model Uncertainties for Deep Learning,

    N. Tagasovska and D. Lopez-Paz, “Single-Model Uncertainties for Deep Learning,” in Proceedings of 33rd Conference on Neural Information Processing Systems, (Vancouver, Canada), Dec. 2019

    2019

  59. [59]

    Data-driven gradient optimization for field emission management in a superconducting radio-frequency linac,

    S. Goldenberg, K. Ahammed, A. Carpenter, J. Li, R. Suleiman, and C. Tennant, “Data-driven gradient optimization for field emission management in a superconducting radio-frequency linac,”Physical Review Accelerators and Beams, vol. 28, p. 044603, Apr. 2025

    2025

  60. [60]

    A mathematical theory of communication,

    C. E. Shannon, “A mathematical theory of communication,”The Bell System Technical Journal, vol. 27, no. 3, pp. 379–423, 1948

    1948

  61. [61]

    Automatic differentiation in pytorch,

    A. Paszke, S. Gross, S. Chintala, G. Chanan, E. Yang, Z. DeVito, Z. Lin, A. Desmaison, L. Antiga, and A. Lerer, “Automatic differentiation in pytorch,” inNIPS-W, 2017

    2017

  62. [62]

    Goodfellow, Y

    I. Goodfellow, Y. Bengio, and A. Courville,Deep Learning. MIT Press, 2016. http://www.deeplearningbook.org

    2016

  63. [63]

    Learning multiple layers of features from tiny images,

    A. Krizhevsky and G. Hinton, “Learning multiple layers of features from tiny images,” Tech. Rep. 0, University of Toronto, Toronto, Ontario, 2009. 23 IOP PublishingJournalvv(yyyy) aaaaaa Authoret al

    2009

  64. [64]

    Deep Residual Learning for Image Recognition

    K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for image recognition,”CoRR, vol. abs/1512.03385, 2015

    work page internal anchor Pith review Pith/arXiv arXiv 2015

  65. [65]

    Torchvision: Pytorch’s computer vision library

    TorchVision maintainers and contributors, “Torchvision: Pytorch’s computer vision library.” https://github.com/pytorch/vision, 2016

    2016

  66. [66]

    Krizhevsky, V

    A. Krizhevsky, V. Nair, and G. Hinton.https://www.cs.toronto.edu/ ~kriz/cifar.html, Accessed: 2025-12-16. 24

    2025