Pith. sign in

REVIEW 3 major objections 5 minor 45 references

Using Data-Derived Priors to Guide CNN Architecture Design for NIR Chemometrics

T0 review · 3 major / 5 minor · reviewed 2026-08-01 · deepseek-v4-flash

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

Pith's one-line read This paper tries to establish that measurable spectral descriptors can act as empirical priors for CNN architecture choices, with receptive-field size and learning rate most strongly tied to dataset properties.

desk verdict The descriptor-to-kernel-size correlations are worth a look, but the central 'cheap rules beat HPO' claim collapses under the paper's own repeated runs. read the letter →

arxiv 2607.25636 v1 pith:6Z6S4YU2 submitted 2026-07-28 cs.LG physics.app-phphysics.comp-ph

Dário Passos This is my paper · ORCID
classification cs.LGphysics.app-phphysics.comp-ph
keywords NIRspectroscopyconvolutionalneuralnetworkshyperparameteroptimizationspectraldescriptorswarm-startpriorschemometricsreceptivefieldmeta-learning
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper asks whether a cheap set of numbers extracted from NIR spectra—how smooth, redundant, information-dense, and finely sampled they are—can replace generic architectural guesswork when designing shallow CNNs for chemometric regression. By running Bayesian hyperparameter optimization across 25 NIR/Vis-NIR regression tasks with two interpretable 1D-CNN scaffolds, the author finds that near-optimal kernel sizes are not random: the preferred convolutional receptive field shrinks as spectral entropy and intrinsic rank rise, and grows as the wavelet energy-support fraction grows, while learning rate tends to fall with training-set size. These relationships are converted into a small number of closed-form heuristics, and the resulting warm-start configurations match or slightly beat full hyperparameter optimization in median test RMSE (ratios 0.953 direct, 1.017 leave-one-dataset-out). The practical payoff would be that a new spectral dataset can be started in a sensible hyperparameter region from descriptor values alone, at a small fraction of the compute, before any dataset-specific tuning.

What carries the argument

The machinery is a descriptor vector computed from training spectra alone: gradient-based spectral entropy, PCA intrinsic rank (n95), wavelet energy-support fraction (C99, the fraction of detail coefficients needed to retain 99% of the detail energy), autocorrelation length, median wavelength spacing, and training-sample count. These are mapped to hyperparameters, above all the convolutional kernel fraction (kernel size normalized by input length) and learning rate, through power-law equations fitted to median near-optimal HPO values—for example, kernel fraction as exp(a − b·log spectral entropy) and learning rate as exp(c − d·log sample count). The mapping serves to translate dataset statis

What would settle it

Take a new collection of NIR regression datasets (not used here), compute the same descriptors, apply the fitted equations for kernel fraction and learning rate, and compare held-out test RMSE against a full Bayesian search on the same two scaffolds. If the heuristic's median RMSE ratio relative to HPO exceeds about 1.1, or if the central correlations (entropy/kernel fraction, sample size/learning rate) do not reappear when HPO is rerun with ten times more trials, the paper's central claim fails.

Watch

Extended reading notes

Core claim

The central claim is that descriptor-derived priors guide shallow CNN design for NIR chemometrics. Specifically, the paper reports that across 25 tasks, the median of the top-20 hyperparameter-search trials shows kernel fraction decreasing with spectral entropy (Spearman -0.85 in the minimal scaffold) and intrinsic rank (-0.63), increasing with wavelet energy-support fraction (0.72) and spectral step (0.53); learning rate decreases with sample count (-0.51). The extended scaffold shows similar but less transferable structure, distributing adaptation across branch usage, dilation, dropout, and filter counts. Heuristics built from these correlations achieve median test-RMSE ratios of 0.953 (di

Load-bearing premise

The meta-analysis assumes that the median of the top-20 out of roughly 500 hyperparameter-search trials is a reliable picture of each dataset's best CNN settings; if those trials are dominated by noise or the search is under-converged, the descriptor–hyperparameter relationships are built on noise.

Editorial extensions

If this is right

  • If the paper is correct, a new NIR dataset can be given a sensible CNN kernel size and learning rate from descriptor values alone, without running hundreds of hyperparameter-search trials.
  • Receptive-field quantities, rather than raw kernel-size integers, become the natural design axis for shallow spectral CNNs, making recommendations portable across instruments with different wavelength sampling.
  • The negative learning-rate–sample-size trend implies that larger spectral training sets should generally be paired with smaller learning rates in this model family.
  • The leave-one-dataset-out results suggest that descriptor-based priors transfer across NIR tasks within a benchmark family, though they do not replace local model selection.
  • Because the direct heuristic beat HPO on 17 of 25 datasets (though often by small margins), descriptor-based warm starts can serve as a strong initialization for a subsequent, shorter HPO run.

Reading between the lines

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

  • Going beyond the paper: if the receptive-field-centered interpretation holds, kernel-size advice for spectral CNNs should be expressed in physical wavelength units (nanometers) rather than integer tap counts—this would make the rules directly testable across instruments with different channel densities.
  • My inference: the learning-rate–sample-size relation, if real, suggests that efforts to scale NIR deep models to larger datasets should also scale down the optimizer step size; a direct experiment could compare training curves at fixed and descriptor-recommended learning rates.
  • A testable extension the author leaves implicit: use the descriptor-derived configuration as the starting point for a full Bayesian search rather than as a fixed warm start; if the prior is informative, the search should converge faster and to a better final solution than a cold start.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The paper proposes that cheap, measurable spectral descriptors of a new NIR dataset can be used to set CNN hyperparameters before task-specific tuning. Across 25 regression tasks, two shallow 1D-CNN scaffolds were optimized with roughly 500-trial TPE runs per task. Using per-dataset medians of the top-20 trials, the author reports descriptor–hyperparameter correlations (notably kernel fraction vs spectral entropy/rank/wavelet support, and learning rate vs sample size), converts the strongest relations into closed-form heuristics (Eqs. 2–4), and evaluates them both directly on the same 25 datasets and under LODO. The paper also includes a preprocessing-aware HPO extension and a 10-seed refit stability analysis. The main claimed results are median test-RMSE ratios of 0.953 (direct) and 1.017 (LODO) relative to HPO.

Significance. If the result holds, the paper would provide a useful, interpretable, and almost computationally free way to initialize shallow CNNs for NIR chemometrics, with an explicit benchmark, a LODO protocol, and public code/data as strengths. The descriptor correlations, especially the receptive-field relations in the minimal scaffold, are plausible and are partially supported by LODO. However, the headline direct-heuristic advantage is not supported by the paper's own repeated refits in Appendix C.1, so the contribution currently rests on parity-plus-lower-cost and on correlations that need a stronger out-of-sample check. This is a worthwhile direction, but the manuscript needs substantial revision before the central claim is reliable.

major comments (3)
  1. [§3.3; Appendix C.1; Abstract] The 0.953 median and 17/25 win count are based on a single unseeded test evaluation per configuration. Table C.11 reports 10 repeated refits for the same minimal-CNN configurations. Recomputing from Table C.11, the heuristic mean is lower than the HPO mean on only 11 of 25 datasets, and the median ratio is about 1.01, not 0.953. Individual rows reverse the Table 8 narrative: Wheat_flours_protein has HPO mean 0.516 vs heuristic 0.560, and Tecator_moisture has HPO mean 1.631 vs heuristic 1.438. Thus the abstract's direct-heuristic claim should be downgraded or replaced with repeated-run intervals; the appendix does not merely qualify the meta-analysis, it contradicts the direct-evaluation headline.
  2. [§2.6; Eqs. (2)–(4)] The direct warm-start evaluation is partly circular: the heuristic coefficients are fitted to the per-dataset top-20 HPO medians of the same 25 datasets on which Table 8 compares them. The manuscript itself labels this "warm-start assessment rather than true zero-shot transfer," but the abstract still says "direct (zero-shot)." The LODO protocol is the appropriate estimate, yet it is acknowledged to have same-source leakage for mango, tomato, pear, and cereal subsets. Please report LODO as the primary transfer evidence, with a sensitivity analysis that removes near-duplicate datasets, and present the direct result only as an in-sample calibration check.
  3. [§2.4, §2.6] The meta-analysis treats the median of the top-20 of ~500 TPE trials as a stable estimate of the preferred hyperparameter region. Only one HPO run per dataset was performed, and both the TPE sampler and the weight initialization are stochastic. Appendix C.1 repeats final refits but does not repeat the HPO search itself. If the top-20 pool is noisy, the correlations in Tables 5–7 and Eqs. (2)–(4) inherit that noise. A concrete test would be to repeat HPO with different random seeds on at least a subset of datasets and report the stability of the medians, Spearman coefficients, and LODO ratios; without this, the descriptor–hyperparameter relationships remain suggestive rather than established.
minor comments (5)
  1. [Abstract; §2.6] The abstract calls the direct evaluation "zero-shot," while §2.6 explicitly says it is a "warm-start assessment rather than true zero-shot transfer." Please harmonize the terminology.
  2. [Tables 5–7] Table 5 states that Benjamini–Hochberg correction was applied, but Tables 6 and 7 do not state whether FDR correction was used. Please report which extended-scaffold correlations survive multiple-testing correction, especially the branch-only columns with n=17.
  3. [§4] The sentence "Additional tests show that using the heuristic for D1 conducts to better model performance than using a fixed default value" makes a quantitative claim with no accompanying table or appendix reference. Add the supporting numbers or remove the claim.
  4. [§2.6; Eqs. (2)–(4)] Eq. (2) uses only spectral entropy even though Table 5 shows comparable correlations for intrinsic rank, autocorrelation length, and spectral step. Please justify the univariate choice and report the fit uncertainty or residual dispersion of the heuristic equations.
  5. [§3.3] The statement "no statistically evident paired difference (p=0.916, Wilcoxon test on log RMSE ratios)" would be more informative with the effect size and confidence interval; "not statistically significant" is not equivalent to "comparable."

Circularity Check

1 steps flagged · score 4.0 of 10

Direct heuristic is a partially in-sample fit; LODO provides independent parity.

  1. fitted input called prediction [Section 2.6 (meta-analysis and heuristic construction), Section 3.3 (Rule-Based Warm-Start and LODO Validation), Eqs. (2)-(4)]
    "The derived equations are not proposed as universal chemometric laws. They are compact empirical priors fitted from near-optimal HPO trials... The direct application of the heuristic rules to define the CNNs achieved a lower test RMSE than HPO on 17/25 datasets (although many by a small margin), whereas the stricter LODO rule did so on 12/25. Their median RMSE ratios relative to HPO were 0.953 and 1.017, respectively."

    Eqs. (2)-(4) are fitted to the per-dataset medians of the top-20 HPO trials from the same 25 target datasets, and the 'direct' warm-start evaluation then applies those same fitted rules back to the same 25 datasets. The headline 0.953 ratio and the 17/25 win count are therefore in-sample estimates of the fitted meta-model, not zero-shot predictions. The paper's caveat ('warm-start assessment rather than true zero-shot transfer') acknowledges only the epoch-calibration use of the target training split; it does not remove the fact that each target dataset's HPO output contributed to the rule being evaluated on it. The LODO result (1.017) is the out-of-sample check and carries the independent transfer evidence.

full rationale

The derivation chain is not globally circular. The descriptor–hyperparameter correlations in Section 3.2 (Tables 5-7) are genuine cross-dataset observations, and the LODO protocol in Section 2.6 removes the held-out dataset from rule construction and yields parity with HPO (median ratio 1.017), so the core transfer claim has independent content. The circular element is narrower and specific to the 'direct' evaluation: Eqs. (2)-(4) are fit to the per-dataset top-20 HPO medians of the same 25 datasets, and the direct warm-start comparison then reports wins and a median ratio against HPO on those same datasets. This is partially a fitted-input-called-prediction issue. The paper itself flags the direct procedure as a warm-start assessment rather than true zero-shot transfer, and Appendix C.1 shows the direct comparison is seed-sensitive under repeated refits (the 17/25 headline is not stable), which is more a correctness risk than a circularity. Self-citations to Passos (2026) are present but motivational/interpretive ('broadly aligned with', 'additional validation'), not load-bearing for the fitted equations or the LODO logic. A minor information-leakage caveat for mango-derived datasets is acknowledged and does not change the main argument. Overall, the central claim does not reduce to its inputs because LODO and the correlation analysis stand independently, but the abstract's 0.953 direct-heuristic advantage is partly in-sample and should be treated as such.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

No new physical entities, forces, or conserved quantities are introduced. The central claim rests mostly on treating HPO-derived medians as ground-truth preferred hyperparameters, on hand-chosen descriptor thresholds, and on fitted regression coefficients in the heuristic equations. The most honest measure of added value is the descriptor-to-hyperparameter mapping itself, which is empirical and not derived from spectral physics.

free parameters (4)
  • Kernel-fraction heuristic coefficients = 5.9701, -5.0126 in Eq. (2)
    Fitted by regression to per-dataset top-20 HPO medians of kernel fraction vs spectral entropy; they define the main warm-start rule for kernel size.
  • Learning-rate heuristic coefficients = -3.3204, -0.4681 in Eq. (4)
    Fitted to the observed learning-rate vs sample-count trend from HPO trials.
  • Pooled L2 regularization fallback = 4.2558e-6 for minimal direct warm-start; fold-specific values in LODO (3.7e-6 to 4.9e-6)
    Since no descriptor-based L2 rule was retained, the pooled near-optimal median is used as a default; it is a fitted constant from HPO output.
  • Descriptor calibration constants = n95=0.95, autocorr threshold=0.5, C99=0.99, entropy eps=1e-12, Daubechies-5, 50 sampled spectra, fixed seed
    Hand-chosen thresholds and sampling choices that affect all descriptor values and therefore all fitted heuristic rules.
assumptions (5)
  • ad hoc to paper The median of the top-20 HPO trials per dataset is a representative estimate of the near-optimal hyperparameter region.
    Used throughout §2.6 and §3.2 to compute descriptor–hyperparameter correlations and to fit heuristic equations. If the top-20 pool is noisy or unrepresentative, all downstream rules inherit the noise.
  • domain assumption The chosen spectral descriptors (gradient entropy, PCA rank, autocorrelation length, wavelet energy measures) capture spectral properties relevant to optimal CNN kernel and learning-rate choice.
    Invoked in §2.2 and §4 in the matched-filter-like interpretation; no independent evidence is given that these descriptors are the right or sufficient set.
  • domain assumption The 25 tasks form a transferable family such that descriptor–hyperparameter rules learned on other tasks apply to a held-out task.
    Required for the LODO validation in §2.6/§3.3. The paper acknowledges leakage between datasets derived from the same original databases (mango, pear, tomato).
  • domain assumption Approximately 500 TPE trials per dataset are sufficient to locate near-optimal configurations in the declared search spaces.
    Stated as a compute-budget cap in §2.4 and §4; the 10-run stability appendix checks seed sensitivity of final configs but not HPO convergence.
  • domain assumption The two shallow scaffold templates (minimalist and extended) are expressive enough to support the claim that descriptor-guided settings are 'good regions' of CNN hyperparameter space.
    The scaffolds deliberately exclude depth, skip connections, and other backbone choices; results may not transfer to deeper CNNs. This is stated in §2.3.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Using Data-Derived Priors to Guide CNN Architecture Design for NIR Chemometrics." pith.science (2026). https://pith.science/paper/6Z6S4YU2

@misc{pith2026260725636,
  author       = {Pith},
  title        = {Pith review of: Using Data-Derived Priors to Guide CNN Architecture Design for NIR Chemometrics},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/6Z6S4YU2}},
  note         = {Machine review of arXiv:2607.25636}
}
read the original abstract

Convolutional neural networks (CNN) for near-infrared (NIR) chemometrics are often designed using generic architectural rules, although spectral datasets differ in sampling, smoothness, redundancy, and sample size. We tested whether these properties can provide empirical priors for CNN design. Across 25 NIR regression tasks, we computed descriptors of dataset size, spectral length and spacing, entropy, intrinsic rank, autocorrelation, and wavelet-scale structure. Two interpretable 1D-CNN scaffolds (a minimal single-convolution model and an extended shallow model with optional branching, dilation, etc) were optimized using five-fold cross-validated Bayesian hyperparameter optimization (HPO). Relationships extracted from near-optimal trials were converted into warm-start heuristics and evaluated directly and through leave-one-dataset-out (LODO) validation. The clearest relationships involved convolutional receptive fields. In the minimal CNN, the preferred kernel fraction decreased with spectral entropy and intrinsic rank, increased with the wavelet energy-support fraction, and the learning rate tended to decrease with training-set size. Direct and LODO heuristics were competitive with HPO, with median test-RMSE ratios of 0.953 and 1.017, respectively. The extended CNN showed similar but less transferable structure across branch usage, dilation, dropout, filter counts, and receptive-field choices. Ten stochastic refits showed seed sensitivity comparable to that of HPO-selected configurations. In a separate experiment, joint preprocessing and CNN HPO outperformed standardized-spectra HPO in 19 of 25 tasks, although gains were dataset-dependent. These results show that spectral descriptors can provide practical CNN design priors, guiding shallow NIR models toward plausible hyperparameter regions before target-specific tuning

Figures

Figures reproduced from arXiv: 2607.25636 by the authors.

Figure 1
Figure 1. Schematic of the minimalist 1D-CNN used. The architecture is deliberately simple to allow direct attribution of performance to [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 2
Figure 2. Schematic of the extended 1D-CNN adopted in this work. The main convolutional branch uses [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Examples of minimal-scaffold meta-analysis correlations: kernel fraction [PITH_FULL_IMAGE:figures/full_fig_p011_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Selected extended-scaffold meta-analysis correlations: primary-branch kernel fraction [PITH_FULL_IMAGE:figures/full_fig_p013_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

45 extracted references · 14 canonical work pages

  1. [1]

    Acquarelli, T

    J. Acquarelli, T. van Laarhoven, J. Gerretzen, T.N. Tran, L.M.C. Buydens, E. Marchiori, Convolutional neural networks for vibrational spectroscopic data analysis, Analytica Chimica Acta 954 (2017) 22--31. https://doi.org/10.1016/j.aca.2016.12.010

    doi:10.1016/j.aca.2016.12.010 2017

  2. [2]

    B. K. Alsberg, A. M. Woodward, D. B. Kell, An introduction to wavelet transforms for chemometricians: A time-frequency approach, Chemometrics and Intelligent Laboratory Systems 37 (1997) 215--239. https://doi.org/10.1016/S0169-7439(97)00029-4

    work page doi:10.1016/s0169-7439(97)00029-4 1997

  3. [3]

    Anderson, K

    N. Anderson, K. Walsh, P. Subedi, J. Walsh, Mango DMC and NIR spectra, Mendeley Data 5 (2024). https://doi.org/10.17632/46htwnp833.5

    work page doi:10.17632/46htwnp833.5 2024

  4. [4]

    K. B. Be \'c , J. Grabska, C.W. Huck, Interpretability in near-infrared (NIR) spectroscopy: Current pathways to the long-standing challenge, TrAC Trends in Analytical Chemistry 189 (2025) 118254. https://doi.org/10.1016/j.trac.2025.118254

    arXiv 2025

  5. [5]

    Benjamini, Y

    Y. Benjamini, Y. Hochberg, 1995. Controlling the false discovery rate: A practical and powerful approach to multiple testing. Journal of the Royal Statistical Society: Series B (Methodological) 57, 289--300. https://www.jstor.org/stable/2346101

    arXiv 1995

  6. [6]

    S. Cruz, R. Guerra, A. Brazio, A.M. Cavaco, M.D. Antunes, D. Passos, Nondestructive simultaneous prediction of internal browning disorder and quality attributes in `Rocha' pear (Pyrus communis L.) using VIS-NIR spectroscopy, Postharvest Biology and Technology 179 (2021) 111562. https://doi.org/10.1016/j.postharvbio.2021.111562

    arXiv 2021

  7. [7]

    C. Cui, T. Fearn, Modern practical convolutional neural networks for multivariate regression: Applications to NIR calibration, Chemometrics and Intelligent Laboratory Systems 182 (2018) 9--20. https://doi.org/10.1016/j.chemolab.2018.07.008

    doi:10.1016/j.chemolab.2018.07.008 2018

  8. [8]

    J. A. Diaz-Olivares, A. van Nuenen, M.J. Gote, V. Fonseca Diaz, W. Saeys, I. Adriaens, B. Aernouts, Near-infrared spectra dataset of milk composition in transmittance mode, Zenodo dataset (2023). https://doi.org/10.5281/zenodo.8263430

    work page doi:10.5281/zenodo.8263430 2023

Show all 45 references
  1. [9]

    Dirks, D

    M. Dirks, D. Poole, Automatic neural network hyperparameter optimization for extrapolation: Lessons learned from visible and near-infrared spectroscopy of mango fruit, Chemometrics and Intelligent Laboratory Systems 231 (2022) 104685. https://doi.org/10.1016/j.chemolab.2022.104685

    2022

  2. [10]

    Elsken, J.H

    T. Elsken, J.H. Metzen, F. Hutter, Neural architecture search: A survey, Journal of Machine Learning Research 20 (2019) 1--21

    2019

  3. [11]

    Feurer, J.T

    M. Feurer, J.T. Springenberg, F. Hutter, Initializing Bayesian hyperparameter optimization via meta-learning, Proceedings of the Twenty-Ninth AAAI Conference on Artificial Intelligence (2015) 1128--1135. https://doi.org/10.1609/aaai.v29i1.9354

    2015 doi

  4. [12]

    F. Gan, J. Luo, Simple dilated convolutional neural network for quantitative modeling based on near infrared spectroscopy techniques, Chemometrics and Intelligent Laboratory Systems 232 (2023) 104710. https://doi.org/10.1016/j.chemolab.2022.104710

    2023

  5. [13]

    C. Guo, J. Zhang, W. Cai, X. Shao, Enhancing transferability of near-infrared spectral models for soluble solids content prediction across different fruits, Applied Sciences 13 (2023) 5417. https://doi.org/10.3390/app13095417

    2023 doi

  6. [14]

    Haffner, M

    F. Haffner, M. Lacoue-Negre, A. Pirayre, D. Gon c alves, J. Gornay, M. Moreaud, IPA: A deep CNN based on Inception for Petroleum Analysis, Fuel 379 (2025) 133016. https://doi.org/10.1016/j.fuel.2024.133016

    2025

  7. [15]

    Helin, U.G

    R. Helin, U.G. Indahl, O. Tomic, K.H. Liland, On the possible benefits of deep learning for spectral preprocessing, Journal of Chemometrics 36 (2022) e3374. https://doi.org/10.1002/cem.3374

    2022 doi

  8. [16]

    Iba \ n ez, J

    G. Iba \ n ez, J. Cebolla-Cornejo, R. Mart \'i , S. Rosell \'o , M. Valc \'a rcel, Non-destructive determination of taste-related compounds in tomato using NIR spectra, Zenodo dataset (2019). https://doi.org/10.5281/zenodo.10633732

    2019 doi

  9. [17]

    Kerschke, H.H

    P. Kerschke, H.H. Hoos, F. Neumann, H. Trautmann, Automated algorithm selection: Survey and perspectives, Evolutionary Computation 27 (2019) 3--45. https://doi.org/10.1162/evco\_a\_00242

    2019 doi

  10. [18]

    Kusumiyati, Y

    K. Kusumiyati, Y. Hadiwijaya, I.E. Putri, A.A. Munawar, Dataset of Vis-NIR spectra for intact cucurbitaceae fruits, Mendeley Data 2 (2021). https://doi.org/10.17632/k55b8mvs84.2

    2021 doi

  11. [19]

    J. Liu, Y. Duan, X. Zhou, and Z. Chen, A method for near-infrared spectral calibration of complex plant samples with wavelet transform and elimination of uninformative variables, Analytical and Bioanalytical Chemistry, vol. 378, pp. 1382--1387, 2004. https://doi.org/10.1007/s0...

    2004 doi

  12. [20]

    W. Luo, Y. Li, R. Urtasun, R. Zemel, Understanding the effective receptive field in deep convolutional neural networks, arXiv:1701.04128 (2017). https://doi.org/10.48550/arXiv.1701.04128

    2017 doi arXiv

  13. [21]

    N. Luo, D. Xu, B. Xing, X. Yang, C. Sun, Principles and applications of convolutional neural network for spectral analysis in food quality evaluation: A review, Journal of Food Composition and Analysis 128 (2024) 105996. https://doi.org/10.1016/j.jfca.2024.105996

    2024

  14. [23]

    Mishra, D.N

    P. Mishra, D.N. Rutledge, J.-M. Roger, K. Wali, H.A. Khan, Chemometric pre-processing can negatively affect the performance of near-infrared spectroscopy models for fruit quality prediction, Talanta 229 (2021) 122303. https://doi.org/10.1016/j.talanta.2021.122303

    2021

  15. [24]

    Mishra, D

    P. Mishra, D. Passos, Realizing transfer learning for updating deep learning models of spectral data to be used in new scenarios, Chemometrics and Intelligent Laboratory Systems 212 (2021) 104283. https://doi.org/10.1016/j.chemolab.2021.104283

    2021

  16. [25]

    Mishra, D

    P. Mishra, D. Passos, Deep multiblock predictive modelling using parallel input convolutional neural networks, Analytica Chimica Acta 1163 (2021) 338520. https://doi.org/10.1016/j.aca.2021.338520

    2021

  17. [26]

    Mishra, D

    P. Mishra, D. Passos, F. Marini, J. Xu, J.M. Amigo, A.A. Gowen, J.J. Jansen, A. Biancolillo, J.-M. Roger, D.N. Rutledge, A. Nordon, Deep learning for near-infrared spectral data modelling: Hypes and benefits, TrAC Trends in Analytical Chemistry 157 (2022) 116804. https://doi.o...

    2022

  18. [27]

    J. P. Nielsen, D.K. Pedersen, L. Munck, Development of nondestructive screening methods for single kernel characterization of wheat, Cereal Chemistry 80 (2003) 274--280. https://doi.org/10.1094/CCHEM.2003.80.3.274

    2003 doi

  19. [28]

    Pasquini, Near infrared spectroscopy: A mature analytical technique with new perspectives, Analytica Chimica Acta 1026 (2018) 8--36

    C. Pasquini, Near infrared spectroscopy: A mature analytical technique with new perspectives, Analytica Chimica Acta 1026 (2018) 8--36. https://doi.org/10.1016/j.aca.2018.04.004

    2018 doi

  20. [29]

    Passos, D

    D. Passos, D. Rodrigues, A.M. Cavaco, M.D. Antunes, R. Guerra, Non-Destructive Soluble Solids Content Determination for `Rocha' Pear Based on VIS-SWNIR Spectroscopy under `Real World' Sorting Facility Conditions, Sensors 19 (2019) 5165. https://doi.org/10.3390/s19235165

    2019 doi

  21. [30]

    Passos, P

    D. Passos, P. Mishra, A tutorial on automatic hyperparameter tuning of deep spectral modelling for regression and classification tasks, Chemometrics and Intelligent Laboratory Systems 223 (2022) 104520. https://doi.org/10.1016/j.chemolab.2022.104520

    2022

  22. [31]

    Passos, P

    D. Passos, P. Mishra, Deep Tutti Frutti: Exploring CNN architectures for dry matter prediction in fruit from multi-fruit near-infrared spectra, Chemometrics and Intelligent Laboratory Systems 243 (2023) 105023. https://doi.org/10.1016/j.chemolab.2023.105023

    2023

  23. [32]

    D. Passos, Deep tutti-frutti II: Explainability of CNN architectures for fruit dry matter predictions, Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 337 (2025) 126068. https://doi.org/10.1016/j.saa.2025.126068

    2025

  24. [33]

    Passos, Convolutional neural networks in Vis–NIR chemometrics: From contradiction to conditional design, TrAC Trends in Analytical Chemistry 203, 119015, (2026)

    D. Passos, Convolutional neural networks in Vis–NIR chemometrics: From contradiction to conditional design, TrAC Trends in Analytical Chemistry 203, 119015, (2026). https://doi.org/10.1016/j.trac.2026.119015

    2026

  25. [34]

    Perrone, R

    V. Perrone, R. Jenatton, M.W. Seeger, C. Archambeau, Scalable hyperparameter transfer learning, Advances in Neural Information Processing Systems 31 (2018) 6845--6855, ISBN: 9781510884472

    2018

  26. [35]

    M. S. Reis, P. M. Saraiva, and B. R. Bakshi, Denoising and signal-to-noise ratio enhancement: wavelet transform and Fourier transform, Comprehensive Chemometrics, 2009, pp. 25--55

    2009

  27. [36]

    J. R. Rice, The algorithm selection problem, Advances in Computers 15 (1976) 65--118. https://doi.org/10.1016/S0065-2458(08)60520-3

    1976 doi

  28. [37]

    https://doi.org/10.5281/zenodo.15838136

    sensAIfood (IG19145), sensAIfood cereal NIR data (wheat, corn, barley) and reference protein and moisture (Perten set), Zenodo dataset (2025). https://doi.org/10.5281/zenodo.15838136

    2025 doi

  29. [38]

    https://doi.org/10.5281/zenodo.15838272

    sensAIfood (IG19145), sensAIfood cereal NIR data (wheat, corn, barley) and reference protein and moisture (Grainit set), Zenodo dataset (2025). https://doi.org/10.5281/zenodo.15838272

    2025 doi

  30. [39]

    Vanschoren, Meta-learning, in: F

    J. Vanschoren, Meta-learning, in: F. Hutter, L. Kotthoff, J. Vanschoren (Eds.), Automated Machine Learning: Methods, Systems, Challenges, Springer, Cham, 2019, pp. 35--61. https://doi.org/10.1007/978-3-030-05318-5

    2019 doi

  31. [40]

    Workman, Interpretive Spectroscopy for Near Infrared, Applied Spectroscopy Reviews 31, 251-320, (1996) https://doi.org/10.1080/05704929608000571

    J. Workman, Interpretive Spectroscopy for Near Infrared, Applied Spectroscopy Reviews 31, 251-320, (1996) https://doi.org/10.1080/05704929608000571

    1996 doi

  32. [41]

    Walsh, A

    J. Walsh, A. Neupane, A. Koirala, M. Li, N. Anderson, Review: The evolution of chemometrics coupled with near infrared spectroscopy for fruit quality evaluation. II. The rise of convolutional neural networks, Journal of Near Infrared Spectroscopy 31 (2023) 109--125. https://do...

    2023 doi

  33. [42]

    Wistuba, N

    M. Wistuba, N. Schilling, L. Schmidt-Thieme, Scalable Gaussian process-based transfer surrogates for hyperparameter optimization, Machine Learning 107 (2017) 43--78. https://doi.org/10.1007/s10994-017-5684-y

    2017 doi

  34. [43]

    J. Yang, J. Xu, X. Zhang, C. Wu, T. Lin, Y. Ying, Deep learning for vibrational spectral analysis: Recent progress and a practical guide, Analytica Chimica Acta 1081 (2019) 6--17. https://doi.org/10.1016/j.aca.2019.06.012

    2019 doi

  35. [44]

    J. Yang, X. Luo, X. Zhang, D. Passos, L. Xie, X. Rao, H. Xu, K.C. Ting, T. Lin, Y. Ying, A deep learning approach to improving spectral analysis of fruit quality under interseason variation, Food Control 140 (2022) 109108. https://doi.org/10.1016/j.foodcont.2022.109108

    2022

  36. [45]

    F. Yu, V. Koltun, Multi-scale context aggregation by dilated convolutions, arXiv:1511.07122 (2015). https://doi.org/10.48550/arXiv.1511.07122

    2015 doi arXiv

  37. [46]

    Zhang, T

    X. Zhang, T. Lin, J. Xu, X. Zhang, Y. Ying, DeepSpectra: An end-to-end deep learning approach for quantitative spectral analysis, Analytica Chimica Acta 1058 (2019) 48--57. https://doi.org/10.1016/j.aca.2019.01.002

    2019 doi

Pith tools

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