Pith. sign in

REVIEW 2 major objections 5 minor 1 cited by

Machine Learning Methods for Small Data and Upstream Bioprocessing Applications: A Comprehensive Review

T0 review · 2 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read This review claims that small-data machine learning is best understood not as a bag of techniques but as methods attached to three workflow stages—data collection and engineering, model development, and continuous monitoring and…

desk verdict A competent, well-written small-data ML review with a genuinely useful workflow-stage taxonomy, weaker where it over-claims crisp stage boundaries but still worth a serious referee. read the letter →

arxiv 2506.12322 v2 pith:UIXNMWH6 submitted 2025-06-14 cs.LG cs.AI

Johnny Peng , Thanh Tung Khuat , Katarzyna Musial , Bogdan Gabrys This is my paper
classification cs.LGcs.AI
keywords smalldatamachinelearningupstreambioprocessingbiopharmaceuticalsMLworkflowtaxonomyactivejust-in-timeRamanspectroscopy
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 review argues that the right way to organize machine learning for small data is by where a method intervenes in the ML workflow, not by its technical family. The authors propose a taxonomy with three stages—Data Collection & Engineering, Model Development, and Continuous Monitoring & Maintenance—and place a family of methods in each, from active learning and data augmentation through semi-supervised, Bayesian, and hybrid modelling to moving-window retraining and just-in-time learning. They define small data relative to the task rather than by a fixed threshold, and they ground the taxonomy in upstream bioprocessing, where expensive bioreactor runs and scarce offline labels make small data the norm. A sympathetic reader would care because the taxonomy turns method selection into a composition problem: instead of choosing among competing techniques, a practitioner can pick one intervention per workflow stage and combine them to squeeze value from limited data.

What carries the argument

The central object is the three-stage taxonomy of Figure 1, which partitions small-data ML practice into Data Collection & Engineering (active learning, data augmentation, dimensionality reduction), Model Development (semi-supervised learning, data- and knowledge-driven hybrid modelling, Bayesian methods, meta- and transfer learning), and Continuous Monitoring & Maintenance (moving-window retraining, just-in-time learning, recursive/instance reweighting, ensemble-based adaptation). The taxonomy does the argument's heavy lifting: it reframes the small-data problem as addressable at different workflow stages and makes cross-stage combination the natural design move, with the review's worked examples (such as data augmentation plus transfer learning plus ensemble adaptation) showing how a practitioner might stack methods. It is also the paper's stated reason for existing, since the authors argue no prior survey organises small-data methods this way.

What would settle it

Going through the application studies the review cites and recording, for each, which workflow stage the method was actually deployed in would settle the claim. The taxonomy predicts clean placement: one stage per method, with value from combining stages. If a large share of successful applications deployed methods at stages other than the taxonomy's assignment, or if studies combining methods across stages showed no benefit over single-stage pipelines, the taxonomy's organizing value would be undercut. A second decisive observation would be a substantial family of small-data methods that fits none of the three stages.

Watch

Extended reading notes

Core claim

The paper's central claim is that its workflow-oriented taxonomy (Figure 1) offers a more complete, holistic perspective on small-data ML methods than earlier technical taxonomies, which classify methods by technique type (augmentation-based, model-based, transfer-based) rather than by how they fit together in practice. The authors state this explicitly in Section 3: the taxonomy 'classifies these ML methods based on the three main ML steps of a typical ML workflow, including Data Collection & Engineering, Model Development, and Continuous Monitoring & Maintenance,' and 'emphasises the integration of techniques across the entire ML process.' Each of roughly eleven method families is assigned to one stage, and for each the paper reviews its core concepts, its suitability for small data, and its demonstrated results in upstream bioprocessing and related process-industry applications, with the aim of giving practitioners actionable guidance for data-constrained environments.

Load-bearing premise

The taxonomy only guides practice if the three workflow stages cleanly and completely partition small-data ML work, with each method belonging to exactly one stage; the review itself notes that methods such as data augmentation and active learning are used in more than one stage, so if those boundaries are porous, the categories could mislead rather than guide method selection.

Editorial extensions

If this is right

  • Practitioners can design small-data pipelines by combining one intervention per workflow stage (for example, data augmentation in the data engineering stage, transfer learning during model development, and ensemble-based adaptation in the update phase), rather than choosing between competing technical families.
  • The taxonomy exposes research gaps the review documents: active learning for Raman-based soft-sensor labelling, GAN- and VAE-based data augmentation, meta-learning for soft sensors, and federated learning across production sites are all flagged as promising but under-explored in upstream bioprocessing.
  • Within the monitoring stage, the review identifies just-in-time learning as the leading retraining strategy for Raman-based bioprocess monitoring, with recursive/instance reweighting and ensemble-based adaptation playing supporting roles for control and fault-tolerant state estimation.
  • The review's relative definition of small data—fewer points than the task typically requires, scaled by model complexity as in the Chinchilla law—implies that choosing a model whose complexity matches the available data is itself a small-data remedy.
  • Hybrid modelling that fuses mechanistic process knowledge with data-driven models emerges as a central remedy for bioprocessing, because it injects information into the model that does not come from labels.

Reading between the lines

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

  • If the taxonomy's composition logic is right, it should transfer beyond bioprocessing as a general checklist for data-scarce projects: a natural test is to take any small-data task, apply one intervention per stage, and measure each stage's marginal contribution through an ablation study.
  • The review assigns each method to exactly one stage, but the very applications it cites show porous boundaries (augmentation used during training, active learning used during monitoring); a more accurate model may be a method-by-stage incidence matrix, and collecting where methods actually get deployed in the literature would test this.
  • The Chinchilla-style scaling argument invites a scale-invariant reading: by the review's own definition, a billion-parameter language model trained on 300 billion tokens is still 'small data,' so the same three-stage remedies (curation, knowledge transfer, continual adaptation) plausibly apply at that scale.
  • A testable extension would be to treat the taxonomy as a recommender for method pairing and benchmark cross-stage combinations—such as dimensionality reduction plus hybrid modelling plus just-in-time learning—against single-stage baselines on bioprocess datasets.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 5 minor

Summary. This manuscript is a literature review of machine learning methods for settings with small datasets, with an application focus on upstream bioprocessing. It defines small data relative to the task at hand, proposes a taxonomy in Section 3 and Figure 1 that organizes methods according to stages of an ML workflow (data collection and engineering, model development, continuous monitoring and maintenance), and then surveys each method family—active learning, data augmentation, dimensionality reduction, semi-supervised learning, Bayesian methods, data- and knowledge-driven hybrid modelling, meta/transfer learning, and retraining/adaptation strategies—with applications in bioprocessing and selected results from other domains. The paper closes with research gaps and future directions, including federated learning, hybrid JITL/ensemble approaches, and meta-learning for soft sensors.

Significance. The review's main value is its broad compilation of applications and its attempt to provide a workflow-level view rather than a purely technical taxonomy. If the taxonomy were made precise, it could indeed help practitioners combine methods across the ML lifecycle, which is a real gap in prior surveys. Strengths include the detailed method descriptions, the explicit remarks on underexplored bioprocessing applications (e.g., GAN/VAE augmentation, meta-learning), and the concrete application results drawn from many domains. The contribution is organizational rather than technical: the paper introduces no new derivations, and the assessment of effectiveness is a narrative synthesis of heterogeneous reported results. The taxonomy-related concerns below are the main obstacles to accepting the paper in its current form.

major comments (2)
  1. [Section 3, Figure 1] Section 3 states that the proposed taxonomy is based on 'three main ML steps' (Data Collection & Engineering, Model Development, Continuous Monitoring & Maintenance), but Figure 1 depicts two additional workflow stages, 'Data & Problem Formulation & Context Understanding' and 'Deployment', and the figure also separates 'Data Engineering' from 'Data Collection'. The relationship between these additional boxes and the three-stage taxonomy is never explained. Since the paper's central claim is that the taxonomy maps methods onto the ML workflow to guide practitioners, this inconsistency needs to be resolved: either restrict the taxonomy to the three stated stages and relabel Figure 1 accordingly, or explicitly define how the additional stages integrate with the method categories.
  2. [Sections 4.1, 4.2, 5.2.1, Figure 1] The taxonomy assigns each method to a single workflow stage, but the described methods repeatedly cross stage boundaries. Active learning is classified under Data Collection & Engineering, yet Algorithm 1 trains and evaluates a model at every iteration, and the reviewed AL applications are mostly experimental design and model-guided optimisation. Data augmentation is placed under Data Engineering, but Section 4.2.3 discusses augmentation used to train and regularise ANNs. Bayesian methods are placed under Model Development, but Section 5.2.1 includes Bayesian optimisation for experimental design (data collection) and Bayesian inference for monitoring. No criterion is given for deciding the primary stage of a method or for representing methods that legitimately span several stages. This weakens the claimed ability of the taxonomy to provide systematic guidance; a revision should add an explicit placement rule and discuss cross-stage usage.
minor comments (5)
  1. [Section 3 (final paragraph)] The text promises 'a summary table highlighting key results' from applications in other fields, but no such table appears in Sections 4–6; please add the table or remove the promise.
  2. [Sections 4.1 and 5.4] There are typographical errors: 'activelly' and 'incrementaly' in Section 4.1 should be 'actively' and 'incrementally', and 'session' in Section 5.4 should be 'section'.
  3. [Section 2 and Table 1] The text refers to 'Table 2' for the common types of time series data, but the table is numbered Table 1; renumber or fix the cross-reference.
  4. [Section 6.3] The abbreviation 'AM' appears without definition in 'incorporate AM into ML models'; define the term or rephrase the sentence.
  5. [Algorithm 1] The stopping condition 'if performance satisfactory then stop' is not part of the standard active-learning objective and is not used in the subsequent query-strategy discussion; consider aligning the pseudocode with the query-budget formulation.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: this is a review/taxonomy paper with no fitted parameters, no predictions, and no derivation chain that reduces to its inputs.

full rationale

The paper is a comprehensive review and taxonomy of machine learning methods for small data in upstream bioprocessing. Its central claim is that the proposed workflow-oriented taxonomy (Figure 1) organizes methods by three ML workflow stages: Data Collection & Engineering, Model Development, and Continuous Monitoring & Maintenance. This is an organizational framework, not a derived empirical result. There are no equations whose outputs are recycled as inputs, no fitted parameters renamed as predictions, and no benchmark evaluation whose outcome is forced by construction. The authors do cite their own prior work (e.g., Kedziora et al., Khuat et al., Gabrys et al.) for background concepts such as ML workflows, soft sensors, and adaptive modelling, but these citations are not load-bearing in the sense of invoking a self-authored uniqueness theorem or smuggling in an ansatz; the workflow stages are independently familiar and the taxonomy's value does not depend on proving those citations. The skeptical concern about internal consistency between the three stated stages and the five workflow boxes in Figure 1, and about methods like active learning or data augmentation being used in more than one stage, is a legitimate correctness and clarity risk for the taxonomy, but it is not circularity: no claim is equivalent to its own input by definition. Accordingly, this is the common honest non-finding: no significant circularity, score 0.

Assumptions & free parameters 0 free parameters · 3 assumptions · 0 invented entities

The paper introduces no fitted parameters, physical entities, or mathematical axioms. Its conclusions rest on a relative definition of small data (Section 1.1), the completeness of the three-stage workflow (Section 3), and the transferability of methods from other process industries to bioprocessing.

assumptions (3)
  • domain assumption Small data is defined as 'fewer data points than what is typically required' for the task.
    Introduced in Section 1.1; the taxonomy's scope depends on this relative, subjective definition rather than a fixed threshold.
  • domain assumption The three-stage ML workflow (Data Collection & Engineering, Model Development, Continuous Monitoring & Maintenance) is a complete and non-overlapping partition of ML practice.
    Adopted in Section 3 and Figure 1; if stages overlap or are incomplete, the placement of methods in the taxonomy may be misleading.
  • domain assumption Methods shown to work in small-data settings in other process industries are transferable to upstream bioprocessing.
    Used throughout Sections 4-6 when citing chemical-process soft-sensor results as evidence for bioprocess applicability.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Machine Learning Methods for Small Data and Upstream Bioprocessing Applications: A Comprehensive Review." pith.science (2026). https://pith.science/paper/UIXNMWH6

@misc{pith2026250612322,
  author       = {Pith},
  title        = {Pith review of: Machine Learning Methods for Small Data and Upstream Bioprocessing Applications: A Comprehensive Review},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UIXNMWH6}},
  note         = {Machine review of arXiv:2506.12322}
}
read the original abstract

Data is crucial for machine learning (ML) applications, yet acquiring large datasets can be costly and time-consuming, especially in complex, resource-intensive fields like biopharmaceuticals. A key process in this industry is upstream bioprocessing, where living cells are cultivated and optimised to produce therapeutic proteins and biologics. The intricate nature of these processes, combined with high resource demands, often limits data collection, resulting in smaller datasets. This comprehensive review explores ML methods designed to address the challenges posed by small data and classifies them into a taxonomy to guide practical applications. Furthermore, each method in the taxonomy was thoroughly analysed, with a detailed discussion of its core concepts and an evaluation of its effectiveness in tackling small data challenges, as demonstrated by application results in the upstream bioprocessing and other related domains. By analysing how these methods tackle small data challenges from different perspectives, this review provides actionable insights, identifies current research gaps, and offers guidance for leveraging ML in data-constrained environments.

Figures

Figures reproduced from arXiv: 2506.12322 by the authors.

Figure 1
Figure 1. Machine learning methods for small data across ML development cycle [PITH_FULL_IMAGE:figures/full_fig_p008_1.png] view at source ↗
Figure 2
Figure 2. Major Branches and Approaches of Transfer Learning [PITH_FULL_IMAGE:figures/full_fig_p039_2.png] view at source ↗

Discussion (0). Sign in to comment.

Forward citations

Cited by 1 Pith paper

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

  1. Lessons Learned from Deploying Adaptive Machine Learning Agents with Limited Data for Real-time Cell Culture Process Monitoring

    q-bio.QM 2025-08 conditional novelty 4.0 of 10

    Adaptive ML models updated with live offline measurements outperform fixed pretrained models on new bioreactor conditions, with just-in-time learning most robust and averaging all models improving stability.

Reference graph

Works this paper leans on

68 extracted references · 64 canonical work pages · cited by 1 Pith paper

  1. [1]

    P., & Gomes, J

    Abaineh, B., Nayak, R., Ranjan, A. P., & Gomes, J. (2007). On-line control of a fed-batch fermentation by using som based multiple local linear models. InProceedings of the 10th International Conference on Engineering Applications of Neural Networks(pp. 368–375). Agharafeie, R., Ramos, J. R. C., Mendes, J. M., & Oliveira, R. (2023). From shallow to deep b...

    work page 2007

  2. [2]

    Liu, J. (2007). On-line soft sensor for polyethylene process with multiple production grades.Control Engi- neering Practice,15, 769–778. Liu, J., Hou, J., & Chen, J. (2021). Dual-layer feature extraction based soft sensor methods and applications to industrial polyethylene processes.Computers and Chemical Engineering,154. Liu, R., Bassalo, M. C., Zeitoun,...

    work page 2007

  3. [3]

    Comon, P. (1994). Independent component analysis, a new concept?Signal Processing,36, 287–314. Dai, Y., Yang, C., Liu, Y., & Yao, Y. (2023). Latent-enhanced variational adversarial active learning assisted soft sensor.IEEE Sensors Journal,23, 15762–15772. 57 Dai, Y., Zhang, Y., Yao, Y., & Liu, Y. (2022). Variational adversarial active learning assisted pr...

    work page 1994

  4. [9]

    Siegl, M., Brunner, V., Geier, D., & Becker, T. (2022). Ensemble-based adaptive soft sensor for fault-tolerant biomass monitoring.Engineering in Life Sciences,22, 229 –

    work page 2022

  5. [10]

    Parnami, A., & Lee, M. (2022). Learning from few examples: A summary of approaches to few-shot learning. ArXiv,abs/2203.04291. Pearson, K. (1901). Liii. on lines and planes of closest fit to systems of points in space.The London, Edinburgh, and Dublin Philosophical Magazine and Journal of Science,2, 559–572. Petitjean, F., Ketterlin, A., & Gan¸ carski, P....

    arXiv 2022

  6. [14]

    J., & Ben ´ ıtez, J

    Bergmeir, C., Hyndman, R. J., & Ben ´ ıtez, J. M. (2016). Bagging exponential smoothing methods using stl decomposition and box–cox transformation.International Journal of Forecasting,32, 303–312. Bernard, O., Chachuat, B., Helias, A., & Rodriguez, J. (2006). Can we assess the model complexity for a bioprocess: Theory and example of the anaerobic digestio...

    work page 2016

  7. [15]

    M., & Bernard, O

    Rodr ´ ıguez, J., Roca, E., Lema, J. M., & Bernard, O. (2008). Determination of the adequate minimum model complexity required in anaerobic bioprocesses using experimental data.Journal of Chemical Technology and Biotechnology,83, 1694 –

    work page 2008

  8. [20]

    B., & Yang, Y

    He, X. B., & Yang, Y. P. (2008). Variable mwpca for adaptive process monitoring.Industrial & engineering chemistry research,47, 419–427. He, Y., Luo, F., & Ranzi, G. (2022). Transferrable model-agnostic meta-learning for short-term household load forecasting with limited training data.IEEE Transactions on Power Systems,37, 3177–3180. Helleckes, L. M., Hem...

    arXiv 2008

Show all 68 references
  1. [21]

    I., Forestier, G., Weber, J., Idoumghar, L., & Muller, P.-A

    Fawaz, H. I., Forestier, G., Weber, J., Idoumghar, L., & Muller, P.-A. (2018). Data augmentation using synthetic data for time series classification with deep residual networks. InProceedings of the 3rd ECML/PKDD Workshop on Advanced Analytics and Learning on Temporal Data. Fe...

    2018

  2. [23]

    Simon, L., & Nazmul Karim, M. (2001). Probabilistic neural networks using bayesian decision strategies and a modified gompertz model for growth phase classification in the batch culture of bacillus subtilis. Biochemical Engineering Journal,7, 41 –

    2001

  3. [24]

    Ge, Z. (2014). Active learning strategy for smart soft sensor development under a small number of labeled data samples.Journal of Process Control,24, 1454 –

    2014

  4. [25]

    D., Adhikari, N., Wong, Y

    Yenice, Z. D., Adhikari, N., Wong, Y. K., Aksakalli, V., Gumus, A. T., & Abbasi, B. (2018). Spsa-fsr: Simultaneous perturbation stochastic approximation for feature selection and ranking.arXiv preprint arXiv:1804.05589, . 73 Yu, J. (2012). A bayesian inference based two-stage ...

    2018 arXiv

  5. [27]

    R., Budka, M., & Gabrys, B

    Ali, A. R., Budka, M., & Gabrys, B. (2020). A review of meta-level learning in the context of multi-component, multi-level evolving prediction systems.CoRR,abs/2007.10818. Alinaghi, M., Surowiec, I., Scholze, S., McCready, C., Zehe, C., Johansson, E., Trygg, J., & Cloarec, O. ...

    2020 arXiv

  6. [32]

    Kamycki, K., & Kapuscinski, T. (2020). Data augmentation with suboptimal warping for time-series classi- fication.Sensors,20,

    2020

  7. [39]

    Gysels, E., Renevey, P., & Celka, P. (2005). Svm-based recursive feature elimination to compare phase synchronization computed from broadband and narrowband eeg signals in brain–computer interfaces. Signal processing,85, 2178–2189. 60 Hamedi Rad, S., Chao, R., Weisberg, S., Li...

    2005

  8. [44]

    Liu, Y., Yang, C., Liu, K., Chen, B., & Yao, Y. (2019). Domain adaptation transfer learning soft sensor for product quality prediction.Chemometrics and Intelligent Laboratory Systems,192. Liu, Y., Yang, C., Zhang, M., Dai, Y., & Yao, Y. (2020). Development of adversarial trans...

    2019

  9. [48]

    M., Herwig, C., & Kager, J

    Sinner, P., Stiegler, M., Goldbeck, O., Seibold, G. M., Herwig, C., & Kager, J. (2022). Online estimation of changing metabolic capacities in continuous corynebacterium glutamicum cultivations growing on a complex sugar mixture.Biotechnology and Bioengineering,119, 575 –

    2022

  10. [50]

    Sagi, O., & Rokach, L. (2018). Ensemble learning: A survey.WIREs Data Mining and Knowledge Discovery, 8, e1249. Sakoe, H., & Chiba, S. (1978). Dynamic programming algorithm optimization for spoken word recognition. IEEE Transactions on Acoustics, Speech, and Signal Processing,...

    2018

  11. [51]

    B., de Silva, V., & Langford, J

    Tenenbaum, J. B., de Silva, V., & Langford, J. C. (2000). A global geometric framework for nonlinear dimensionality reduction.Science,290, 2319–2323. 71 Tharwat, A., & Schenck, W. (2023). A survey on active learning: State-of-the-art, practical challenges and research directio...

    2000

  12. [53]

    C., Soudier, P., & Faulon, J.-L

    Borkowski, O., Koch, M., Zettor, A., Pandi, A., Batista, A. C., Soudier, P., & Faulon, J.-L. (2020). Large scale active-learning-guided exploration for in vitro protein production optimization.Nature Communications, 11,

    2020

  13. [92]

    S., Seth, S., Mallick, K., Rahaman, Z., & Mallik, S

    Borah, K., Das, H. S., Seth, S., Mallick, K., Rahaman, Z., & Mallik, S. (2024). A review on advancements in feature selection and feature extraction for high-dimensional ngs data analysis.Functional & Integrative Genomics,24,

    2024

  14. [95]

    Shu, J., Xu, Z., & Meng, D. (2018). Small sample learning in big data era.arXiv preprint arXiv:1808.04572, . Sibley, M., Woodhams, A., Hoehse, M., & Zoro, B. (2020). Novel integrated raman spectroscopy technology for minibioreactors.BioProcess Int,18,

    2018 arXiv

  15. [98]

    Kaneko, H., & Funatsu, K. (2013). Classification of the degradation of soft sensor models and discussion on adaptive models.AIChE Journal,59, 2339–2347. Kano, M., & Fujiwara, K. (2013). Virtual sensing technology in process industries: trends and challenges revealed by recent ...

    2013

  16. [102]

    S., Escalante, S

    Baruch, I. S., Escalante, S. F., Mariaca-Gaspar, C. R., & Barrera-Cortes, J. (2007). Recurrent neural control of wastewater treatment bioprocess via marquardt learning.IF AC Proceedings Volumes,40, 289–294. 55 Baruch, I. S., & Mariaca-Gaspar, C. R. (2009). A levenberg-marquard...

    2007

  17. [114]

    Bakirov, R., Gabrys, B., & Fay, D

    All Open Access, Green Open Access, Hybrid Gold Open Access. Bakirov, R., Gabrys, B., & Fay, D. (2015). On sequences of different adaptive mechanisms in non-stationary regression problems. In2015 International Joint Conference on Neural Networks (IJCNN). IEEE. Bakirov, R., Gab...

    2015

  18. [119]

    A., Manirafasha, E., Zhang, D., Yue, Q., & Jing, K

    del Rio-Chanona, E. A., Manirafasha, E., Zhang, D., Yue, Q., & Jing, K. (2016). Dynamic modeling and optimization of cyanobacterial c-phycocyanin production process by artificial neural network.Algal Research,13, 7 –

    2016

  19. [138]

    Chen, K., & Liu, Y. (2013). Adaptive control of continuous time-varying bioprocesses using recursive kernel learning controller with polynomial form.IF AC Proceedings Volumes,46, 359–364. Chiu, K.-C., & Du, D. (2024). A neural ordinary differential equation model for predictin...

    2013

  20. [139]

    Borisyak, M., Born, S., Neubauer, P., & Cruz-Bournazou, M. N. (2024). Deep learning for fast inference of mechanistic models’ parameters. InComputer Aided Chemical Engineering(pp. 3043–3048). Elsevier volume

    2024

  21. [144]

    Yuan, X., Ge, Z., & Song, Z. (2014). Locally weighted kernel principal component regression model for soft sensing of nonlinear time-variant processes.Industrial & Engineering Chemistry Research,53, 13736– 13749. Zhang, L., Wang, M., Castan, A., Hjalmarsson, H., & Chotteau, V....

    2014

  22. [151]

    E., Shen, Z., Long, S

    Huang, N. E., Shen, Z., Long, S. R., Wu, M. C., Shih, H. H., Zheng, Q., Yen, N.-C., Tung, C. C., & Liu, H. H. (1998). The empirical mode decomposition and the hilbert spectrum for nonlinear and non-stationary time series analysis.Proceedings of the Royal Society. A, Mathematic...

    1998

  23. [152]

    Zhu, Q.-X., Hou, K.-R., Chen, Z.-S., Gao, Z.-S., Xu, Y., & He, Y.-L. (2021). Novel virtual sample generation using conditional gan for developing soft sensor with small data.Engineering Applications of Artificial Intelligence,106, 104497. Zhu, Q.-X., Xu, T.-X., Xu, Y., & He, Y...

    2021

  24. [211]

    Yang, Y., Farid, S

    All Open Access, Bronze Open Access. Yang, Y., Farid, S. S., & Thornhill, N. F. (2014). Data mining for rapid prediction of facility fit and debottlenecking of biomanufacturing facilities.Journal of Biotechnology,179, 17 –

    2014

  25. [214]

    S., & Banerjee, T

    68 Ray, P., Reddy, S. S., & Banerjee, T. (2021). Various dimension reduction techniques for high dimensional data analysis: a review.Artificial Intelligence Review,54, 3473 –

    2021

  26. [241]

    70 Siegl, M., K¨ ampf, M., Geier, D., Andreeßen, B., Max, S., Zavrel, M., & Becker, T. (2023). Generalizability of soft sensors for bioprocesses through similarity analysis and phase-dependent recalibration.Sensors,

    2023

  27. [263]

    Cioffi, R., Travaglioni, M., Piscitelli, G., Petrillo, A., & De Felice, F. (2020). Artificial intelligence and machine learning applications in smart production: Progress, trends, and directions.Sustainability,12. Cleveland, R. B., Cleveland, W. S., & Terpenning, I. (1990). St...

    2020

  28. [334]

    L., Amicarelli, A

    Quintero, O. L., Amicarelli, A. A., Scaglia, G., & di Sciascio, F. (2009). Control based on numerical methods and recursive bayesian estimation in a continuous alcoholic fermentation process.BioResources,4, 1372 –

    2009

  29. [416]

    T., Pfister, F

    Um, T. T., Pfister, F. M. J., Pichler, D., Endo, S., Lang, M., Hirche, S., Fietzek, U., & Kuli´ c, D. (2017). Data augmentation of wearable sensor data for parkinson’s disease monitoring using convolutional neural networks. InProceedings of the 19th ACM International Conferenc...

    2017

  30. [440]

    Esche, E., Talis, T., Weigert, J., Brand Rihm, G., You, B., Hoffmann, C., & Repke, J.-U. (2022). Semi- supervised learning for data-driven soft-sensing of biological and chemical processes.Chemical Engineer- ing Science,251, 117459. Esmonde-White, K. A., Cuellar, M., Uerpmann,...

    2022

  31. [483]

    Yamada, N., & Kaneko, H. (2022). Adaptive soft sensor based on transfer learning and ensemble learning for multiple process states.Analytical Science Advances,3, 205 –

    2022

  32. [508]

    T., Chen, H.-C., Prayitno, Ogiela, M

    Putra, K. T., Chen, H.-C., Prayitno, Ogiela, M. R., Chou, C.-L., Weng, C.-E., & Shae, Z.-Y. (2021). Federated compressed learning edge computing framework with ensuring data privacy for pm2.5 prediction in smart city sensing applications.Sensors,21. Qiu, K., Wang, J., Zhou, X....

    2021

  33. [511]

    Baruch, I., Mariaca-Gaspar, C.-R., Barrera-Cortes, J., & Castillo, O. (2010). Direct and indirect neural identification and control of a continuous bioprocess via marquardt learning.Studies in Computational Intelligence,318, 81 –

    2010

  34. [590]

    St ˆ ıng˘ a, F., & Petre, E. (2018). Estimation based control strategies for an aerobic bioprocess. In2018 22nd International Conference on System Theory, Control and Computing (ICSTCC)(pp. 218–223). IEEE. Stonier, A., Pain, D., Westlake, A., Hutchinson, N., Thornhill, N. F., ...

    2018

  35. [683]

    Feng, L., & Zhao, C. (2020). Adversarial sample based semi-supervised learning for industrial soft sensor. IF AC-PapersOnLine,53, 11644–11649. Fern´ andez, M. C., Pantano, M. N., Rodriguez, L., & Scaglia, G. (2021). State estimation and nonlinear tracking control simulation ap...

    2020

  36. [787]

    Mohanty, A., Sutherland, A., Bezbradica, M., & Javidnia, H. (2022). Skin disease analysis with limited data in particular rosacea: A review and recommended framework.IEEE Access,10, 39045 – 39068. Mumuni, A., & Mumuni, F. (2025). Data augmentation with automated machine learni...

    2022

  37. [892]

    Le Guennec, A., Malinowski, S., & Tavenard, R. (2016). Data Augmentation for Time Series Classification using Convolutional Neural Networks. InECML/PKDD Workshop on Advanced Analytics and Learning on Temporal Data. Riva Del Garda, Italy. 64 Lee, D. D., & Seung, H. S. (1999). L...

    2016

  38. [922]

    Al-Digeil, M., Grinberg, Y., Daniele, M., Dezfouli, M., Schmid, J., Cheben, P., Janz, S., & Xu, D. (2022). Pca-boosted autoencoders for nonlinear dimensionality reduction in low data regimes.arXiv preprint arXiv:2205.11673, . Al-Rawi, M., & Karajeh, H. (2007). Genetic algorith...

    2022 arXiv

  39. [1114]

    Bayer, B., Duerkop, M., Striedner, G., & Sissolak, B. (2021). Model transferability and reduced experimental burden in cell culture process development facilitated by hybrid modeling and intensified design of exper- iments.Frontiers in bioengineering and biotechnology,9, 74021...

    2021

  40. [1200]

    Arslan, M., Guzel, M., Demirci, M., & Ozdemir, S. (2019). Smote and gaussian noise based sensor data augmentation. In2019 4th International Conference on Computer Science and Engineering (UBMK) (pp. 1–5). Bader, J., Narayanan, H., Arosio, P., & Leroux, J.-C. (2023). Improving ...

    2019

  41. [1241]

    J., Musial, K., & Gabrys, B

    63 Kedziora, D. J., Musial, K., & Gabrys, B. (2024). Autonoml: Towards an integrated framework for au- tonomous machine learning.Foundations and Trends®in Machine Learning,17, 590–766. Khatibisepehr, S., Huang, B., & Khare, S. (2013). Design of inferential sensors in the proce...

    2024 arXiv

  42. [1359]

    Pan, T., Chen, J., Zhang, T., Liu, S., He, S., & Lv, H. (2022). Generative adversarial network in mechanical fault diagnosis under small sample: A systematic review on applications and future perspectives.ISA Transactions,128, 1 –

    2022

  43. [1360]

    Streefland, M., Martens, D., Beuvery, E., & Wijffels, R. (2013). Process analytical technology (pat) tools for the cultivation step in biopharmaceutical production.Engineering in Life Sciences,13. Sugiyama, M. (2006). Local fisher discriminant analysis for supervised dimension...

    2013

  44. [1395]

    Radivojevi´ c, T., Costello, Z., Workman, K., & Garcia Martin, H. (2020). A machine learning automated recommendation tool for synthetic biology.Nature Communications,11,

    2020

  45. [1461]

    Ge, Z. (2015). Mixture bayesian regularization of pcr model and soft sensing application.Industrial Elec- tronics, IEEE Transactions on,62, 4336–4343. Ge, Z., Huang, B., & Song, Z. (2014). Mixture semisupervised principal component regression model and soft sensor application....

    2015 arXiv

  46. [1702]

    W., Sathish, A., Wan, N., Cheng, Q., Wen, Z., Tang, Y

    Roell, G. W., Sathish, A., Wan, N., Cheng, Q., Wen, Z., Tang, Y. J., & Bao, F. S. (2022). A comparative evaluation of machine learning algorithms for predicting syngas fermentation outcomes.Biochemical Engineering Journal,186, 108578. Roger, V., Farinas, J., & Pinquier, J. (20...

    2022

  47. [1768]

    Fields, T., Hsieh, G., & Chenou, J. (2019). Mitigating drift in time series data with noise augmentation. In 2019 International Conference on Computational Science and Computational Intelligence (CSCI)(pp. 227–230). Fortescue, T., Kershenbaum, L., & Ydstie, B. (1981). Implemen...

    2019

  48. [1872]

    Bosch, J., Holmstr¨ om Olsson, H., & Crnkovic, I. (2022). Chapter 13 engineering ai systems. In J. Bosch, J. Carlson, H. Holmstr¨ om Olsson, K. Sandahl, & M. Staron (Eds.),Accelerating Digital Transformation: 10 Years of Software Center(pp. 407–425). Cham: Springer Internation...

    2022

  49. [1954]

    M., Soner, B., Coleri, S., Gunduz, D., & Bennis, M

    Elbir, A. M., Soner, B., Coleri, S., Gunduz, D., & Bennis, M. (2022). Federated learning in vehicular networks. Eltoft, T. (2002). Data augmentation using a combination of independent component analysis and non-linear time-series prediction. InProceedings of the 2002 Internati...

    2022

  50. [1996]

    Tran, L., & Choi, D. (2020). Data augmentation for inertial sensor-based gait deep neural network.IEEE Access,8, 12364–12378. Tulsyan, A., Garvin, C., & ¨Undey, C. (2018). Advances in industrial biopharmaceutical batch process monitoring: Machine-learning methods for small dat...

    2020

  51. [2175]

    W., Ong, Y

    Cao, X., Bu, W., Huang, S., Zhang, M., Tsang, I. W., Ong, Y. S., & Kwok, J. T. (2022). A survey of learning on small data: generalization, optimization, and challenge.arXiv preprint arXiv:2207.14443, . Chai, Z., Zhao, C., Huang, B., & Chen, H. (2022). A deep probabilistic tran...

    2022 arXiv

  52. [2586]

    Tulsyan, A., Wang, T., Schorner, G., Khodabandehlou, H., Coufal, M., & Undey, C. (2020). Automatic real-time calibration, assessment, and maintenance of generic raman models for online monitoring of cell culture processes.Biotechnology and Bioengineering,117, 406 –

    2020

  53. [3224]

    Hotelling, H. (1936). Relations between two sets of variates.Biometrika,28, 321–377. Hsiao, Y.-D., Kang, J.-L., & Wong, D. S.-H. (2021). Development of robust and physically interpretable soft sensor for industrial distillation column using transfer learning with small dataset...

    1936

  54. [3459]

    Zhao, C., Wang, F., & Jia, M. (2007). Dissimilarity analysis based batch process monitoring using moving windows.AIChE journal,53, 1267–1277. Zhou, H., Zhang, Y., Zhang, Y., & Liu, H. (2019). Feature selection based on conditional mutual information: minimum conditional releva...

    2007

  55. [3468]

    Syu, M.-J., & Hou, C.-L. (1997). A neural network study on the dynamic identification of a fermentation system.Bioprocess Engineering,17, 203–213. Tang, Q., Li, D., & Xi, Y. (2018). A new active learning strategy for soft sensor modeling based on feature reconstruction and unc...

    1997

  56. [3515]

    R., Albarqouni, S., Bakas, S., Galtier, M

    Rieke, N., Hancox, J., Li, W., Milletar ` ı, F., Roth, H. R., Albarqouni, S., Bakas, S., Galtier, M. N., Landman, B. A., Maier-Hein, K., Ourselin, S., Sheller, M., Summers, R. M., Trask, A., Xu, D., Baust, M., & Cardoso, M. J. (2020). The future of digital health with federate...

    2020

  57. [4541]

    Hutter, C., Stosch, M., Cruz Bournazou, M

    All Open Access, Green Open Access, Hybrid Gold Open Access. Hutter, C., Stosch, M., Cruz Bournazou, M. N., & Butt´ e, A. (2021). Knowledge transfer across cell lines using hybrid gaussian process models with entity embedding vectors.Biotechnology and bioengineering, 118, 4389...

    2021

  58. [4879]

    Raschka, S. (2018). Model evaluation, model selection, and algorithm selection in machine learning.ArXiv, abs/1811.12808. Rashedi, M., Khodabandehlou, H., Wang, T., Demers, M., Tulsyan, A., Garvin, C., & Undey, C. (2024). Integration of just-in-time learning with variational a...

    2018 arXiv

  59. [5150]

    Hashizume, T., Ozawa, Y., & Bei-Wen, Y. (2023). Employing active learning in the optimization of culture medium for mammalian cells.npj Systems Biology and Applications,9,

    2023

  60. [7609]

    (2010).Semi-supervised learning

    Chapelle, O., Scholkopf, B., & Zien, A. (2010).Semi-supervised learning. Adaptive computation and machine learning series. Cambridge, Massachusetts: MIT Press. Chawla, N. V., Bowyer, K. W., Hall, L. O., & Kegelmeyer, W. P. (2002). Smote: synthetic minority over- sampling techn...

    2010

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.