Pith. sign in

REVIEW 4 major objections 6 minor 28 references

Accurate and Uncertainty-Aware Multi-Task Prediction of HEA Properties Using Prior-Guided Deep Gaussian Processes

T0 review · 4 major / 6 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read This paper shows that a two-layer multi-task deep Gaussian process trained on residuals around an encoder-decoder prior, with auxiliary computational tasks, predicts experimental high-entropy alloy properties most accurately while also…

desk verdict A solid empirical benchmark with a real gap: the accuracy claim for HDGP P-All mostly holds up, but the 'uncertainty-aware' half of the title is not demonstrated. read the letter →

arxiv 2506.14828 v1 pith:DF4K2OGZ submitted 2025-06-13 cs.LG cond-mat.mtrl-sci

Sk Md Ahnaf Akif Alvi , Mrinalini Mulukutla , Nicolas Flores , Danial Khatamsaz , Jan Janssen , Danny Perez , Douglas Allaire , Vahid Attari
show 1 more author
Raymundo Arroyave
This is my paper · ORCID
classification cs.LGcond-mat.mtrl-sci
keywords deepGaussianprocessesmulti-tasklearninghigh-entropyalloysuncertaintyquantificationsurrogatemodelingheterotopicdatamaterialsinformaticsencoder-decoderprior
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 which surrogate model best predicts several correlated mechanical properties of high-entropy alloys from composition alone, when data are sparse, noisy, and incomplete. It compares conventional Gaussian processes (cGP), deep Gaussian processes (DGP), an encoder-decoder neural network, and XGBoost on a curated dataset of more than 100 Al-Co-Cr-Cu-Fe-Mn-Ni-V alloys that mixes experimental measurements with computational descriptors. Its central claim is that a two-layer variational multi-task DGP which models residuals around an encoder-decoder prior, and which also trains on auxiliary computational tasks, is the overall best: it gives the lowest errors and highest rank correlations for most experimental properties while producing uncertainty estimates. If this is right, materials scientists gain a practical recipe for uncertainty-aware surrogate modeling in alloy design, where experiments are costly and the composition space is vast.

What carries the argument

The load-bearing mechanism is a two-layer variational multi-task deep Gaussian process with residual prior injection. Each layer is a multi-task GP with 10 latent functions, and a reduction parameter controls the dimensionality of the latent space between layers; the prior is a regularized dense encoder-decoder network trained on the same tasks. The prior's predictions are subtracted from the training outputs so that the DGP models only residuals, then added back to form final predictions and variances. This multi-task structure lets the model share information across correlated properties, handle heterotopic (partially missing) data by evaluating the likelihood only on observed outputs, and avoid defaulting to a constant mean in regions far from training data.

What would settle it

Take a held-out set of compositions far from the training region where the encoder-decoder prior is known to be inaccurate, compute HDGP-P-All's predictive intervals, and check how often the true values fall within two standard deviations; if coverage is far below the nominal 95% and uncertainties do not grow outside the training domain, the claim that the model is uncertainty-aware in the stated sense fails.

Watch

Extended reading notes

Core claim

The paper's central discovery is that injecting a deterministic neural-network prior into a deep Gaussian process changes the quality of multi-task property prediction. Instead of learning raw outputs, the DGP learns residuals: the encoder-decoder's predictions are subtracted from the training outputs, the DGP is fit to those residuals, and the prior is added back at prediction time. In the best configuration, called HDGP-P-All, the DGP is trained on all tasks, including the six measured experimental properties plus auxiliary computed descriptors such as valence electron concentration, stacking fault energy, model-based yield strength, and simulated depth of penetration, and it outperforms the other six configurations on most experimental tasks, with substantial RMSE reductions relative to conventional GP and XGBoost on yield strength, ultimate tensile strength, hardness, modulus, and elongation. The paper states directly that HDGP-P-All emerges as the optimal modeling choice for predicting correlated material properties within HEA datasets, and it reports model-predicted standard deviations as error bars.

Load-bearing premise

The weakest load-bearing premise is that the DGP's predictive variance on the residuals is the full uncertainty of the final prediction: the encoder-decoder prior is treated as a deterministic function whose own error is never propagated, so if the prior is wrong far from the training data, the reported error bars will be too small.

Editorial extensions

If this is right

  • For sparse multi-task materials datasets, a prior-guided deep Gaussian process is a better default than conventional single-task Gaussian processes and XGBoost when both accuracy and uncertainty are needed.
  • Adding auxiliary computational descriptors as extra tasks improves prediction of the main experimental properties, so experimental and simulation data should be modeled jointly rather than separately.
  • The hybrid structure, deterministic neural prior plus probabilistic GP residual, offers a way to obtain error bars from deep models without full Bayesian neural networks.
  • Because the DGP handles heterotopic and incomplete data natively, no imputation is required before modeling, which simplifies practical workflows.
  • The best configuration's predictive error bars make it suitable for active learning and Bayesian optimization loops aimed at discovering new alloy compositions.

Reading between the lines

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

  • The same residual-prior recipe likely transfers to other small-data multi-task materials datasets, such as steels or oxides, whenever a cheap deterministic surrogate can be trained first; the expected gain should grow as the prior improves.
  • A direct testable extension would be to propagate uncertainty from the encoder-decoder prior into the DGP (for example by treating the prior as a random function or ensembling priors); if predictive interval coverage improves, the paper's uncertainty estimates are incomplete as currently reported.
  • For out-of-distribution compositions, bias in the encoder-decoder prior may dominate the prediction error, so acquisition functions in Bayesian optimization that rely on the reported variances could over-exploit if that prior is inaccurate.
  • The dataset's planned release of uncertainty estimates for experimental replicates and computed descriptors would allow direct calibration of the model-predicted standard deviations, something the present study does not perform.
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

4 major / 6 minor

Summary. The paper benchmarks four surrogate model families—conventional Gaussian processes (cGP), two-layer variational deep Gaussian processes (DGP), an encoder-decoder neural network, and XGBoost—on the BIRDSHOT high-entropy alloy dataset, which contains multiple experimental and computational property tasks. It introduces 'HDGP' configurations in which a DGP is trained on residuals after subtracting the predictions of a pre-trained encoder-decoder 'prior,' with variants using either all tasks or only main tasks. The central claim is that the prior-guided DGP using all tasks (HDGP P-All) gives the best predictive accuracy for most experimental properties while also providing predictive uncertainty estimates. Experiments on five 80/20 random splits report R², RMSE, MAE, SMAPE, and Spearman correlation; parity plots with model-predicted standard deviations are shown for one split. The paper concludes that HDGP P-All is the optimal modeling choice for correlated HEA properties.

Significance. If fully supported, the work would provide a useful benchmark and a practical recipe for combining deterministic neural priors with deep GPs in sparse multi-output materials data. A clear strength is the public release of code and data, as well as the use of multiple random splits with reported standard deviations. However, the 'uncertainty-aware' half of the claim is not established: the reported intervals do not include the prior's uncertainty, and no calibration or coverage analysis is provided. The accuracy claim is also overstated relative to the tables, which show that HDGP P-All is not the best model for several auxiliary tasks. The core idea is defensible, but the manuscript needs revision to align its claims with the evidence.

major comments (4)
  1. [Section 2.3, Figure 5] The predictive uncertainty shown in Figure 5 is described as 'the standard deviation as predicted by the model.' Because the encoder-decoder prior is re-added deterministically at test time and is not a probabilistic model, the reported variance is only the residual DGP's predictive variance; the prior's parameter and model-form uncertainty is never propagated. The paper itself states that in out-of-distribution regions the model 'converges to the prior' (Section 2.3), meaning the final point prediction is dominated by the prior exactly where the prior's uncertainty matters most, yet the reported interval excludes that uncertainty. The paper provides no calibration analysis, coverage test, or comparison of predictive intervals against empirical residuals. Since 'uncertainty-aware' appears in the title and abstract, this is a load-bearing gap. Please either propagate the prior's uncertainty (e.g., treating the prior as a GP or using an ensemble/Bayesian neural network) or provide calibration evidence showing that the residual-DGP variance alone is adequate.
  2. [Tables 2 and 3] The statement 'HDGP P-All emerges as the optimal modeling choice' (Section 2.3) is not supported by the tables for all tasks. For VarvYS, HDGP P-All has RMSE 133.2±25.1, which is worse than XGB (63.1±43.6) and cGP (85.3±60.2); for VEC, HDGP P-All (R²=0.676) is far worse than cGP (R²=1.000); for SFE, HDGP P-All (R²=0.19) is not better than HDGP NP-All (0.29). Thus the optimal-model claim should be restricted to the main experimental mechanical properties (yield strength, UTS, elongation, hardness, modulus, dynamic hardness) where the improvement is more consistent. Moreover, with only five random splits, many differences are within one standard deviation (e.g., UTS/YS R² is 0.803±0.131 for HDGP P-All vs 0.785±0.028 for cGP), and no significance tests are provided, so 'consistently superior' is not statistically established.
  3. [Section 4.3] The selection of the DGP reduction parameter is described only as choosing the value that gave the highest Spearman coefficient and lowest RMSE in 'cross-validation accuracy metrics,' without stating whether this cross-validation was performed on the training folds only, whether the same 80/20 splits were reused, or what the chosen reduction parameter values were. If the reduction parameter was selected using the same test splits that are later averaged, the reported metrics are optimistically biased. Please specify the inner validation scheme and report the selected reduction parameter for each DGP configuration.
  4. [Table 1 and Section 2.3] There is an inconsistency about which tasks are included in the prior: Table 1 says HDGP P-All uses 'priors derived from an encoder-decoder model trained on main tasks,' while the text says HDGP P-All 'had encoder-decoder priors for all tasks except Varveillien yield strength, SFE and VEC.' Clarify the encoder-decoder's output space and how the auxiliary computationally predicted properties (VarvYS, SFE, VEC, depth of penetration) enter the DGP model and the prior injection.
minor comments (6)
  1. [Section 4.4] Please provide the encoder-decoder architecture and training hyperparameters (number of hidden layers and units, activation, regularization strength, optimizer, epochs, learning rate); the code link is useful but the manuscript should be self-contained.
  2. [Section 2.1] The dataset description would benefit from the exact number of compositions, the number of samples per task, and the missingness pattern across tasks; these are needed to assess the claimed heterotopic data challenge.
  3. [Tables 2 and 3] The '-' entries for HDGP NP-Main and HDGP P-Main on the first three tasks are not explained; state that these models were only trained on main tasks and therefore do not predict auxiliary tasks.
  4. [Section 4.2, Eq. (9)] The ELBO expression is incomplete: the variables Q, p(Y,H,F) are undefined and the equation as written is not a usable description; either define it properly or omit it.
  5. [Introduction] There are typographical errors such as 'quantificdaation' and 'ans' in Section 1; please proofread the manuscript.
  6. [Figure 5] The caption lists panels (a)-(k) but the text does not describe which train-test split is shown; state whether this is a representative split and note that the error bars are the predictive standard deviations (see also the major comment about uncertainty).

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the central benchmark is out-of-sample residual learning, and the uncertainty caveat is a calibration limitation, not a circular step.

full rationale

The paper's central derivation is an empirical comparison of seven surrogate models on an 80/20 train-test split of the BIRDSHOT dataset, with results averaged over five random splits. For the prior-guided DGP, the encoder-decoder prior is trained on the training split, the DGP is trained on the residuals of that same training split, and test predictions are formed by adding the prior prediction to the DGP residual prediction. The authors explicitly prevent leakage: "We fixed the train-test split between the two models (DGP and encoder-decoder) for the cases where DGPs use encoder-decoder as prior. This prevents data leakage. By preventing data leakage, we ensure our test set is totally unseen by the DGP models during training, as it does not exist in the training set of encoder-decoder used by DGP as prior." This is residual learning, not circular reasoning: the held-out test targets never enter either model's fitting procedure, and the reported accuracy metrics in Tables 2 and 3 are genuine out-of-sample evaluations. The claim that HDGP P-All is the best model is a data-driven finding, not an identity derived from the model definitions. The main caveat—that the reported predictive standard deviation reflects only the residual DGP variance and does not propagate uncertainty from the deterministic encoder-decoder prior—is a real limitation of the uncertainty quantification, but it is not a circular step: the prior is not defined in terms of the test outputs, and no fitted parameter is renamed as a prediction. The self-citations to the authors' prior work ([11], [14], [26]) are contextual or correspond to implemented methods and datasets; the paper's own prior-versus-no-prior comparisons (HDGP P vs HDGP NP) provide the actual evidence for the prior's benefit, so no load-bearing claim reduces to a self-citation. Overall, the derivation chain is self-contained and empirically falsifiable, and no circularity is present.

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

The central claim rests on the BIRDSHOT dataset and on the DGP with a learned encoder-decoder prior. The free parameters are the DGP architecture choice (reduction parameter), the variational and kernel parameters fit during training, and the encoder-decoder weights that define the prior. These are all fit to the same training data, which is standard supervised learning but means the 'prior' is not an independent physical constraint. The axioms are the reliability of the experimental data, the usefulness of auxiliary computational tasks, the adequacy of variational inference, and the treatment of the prior as deterministic with no propagated uncertainty. No new physical entities are introduced.

free parameters (4)
  • DGP reduction parameter = Not specified numerically; selected via cross-validation (Section 4.3)
    Controls dimensionality reduction between DGP layers; the authors tuned it per task and report only that the best value was chosen by highest Spearman and lowest RMSE.
  • DGP kernel hyperparameters and variational parameters = Learned via ELBO optimization (Section 4.2)
    Lengthscales, noise variances, and variational parameters are fit to the training data; the central performance claim depends on these fitted values.
  • Encoder-decoder prior network weights = Trained on the fixed training split (Section 2.3)
    The 'prior' is a neural network fit to the same training data; its predictions are subtracted before DGP training.
  • cGP kernel hyperparameters = Learned by maximum likelihood (Section 4.1)
    Baseline GP hyperparameters are fit to each task; included for completeness since the benchmark compares against them.
assumptions (4)
  • domain assumption The BIRDSHOT dataset's experimental properties are measured reliably and consistently across two campaigns; all properties from a given alloy come from the same physical ingot.
    The comparison treats these measurements as ground truth for fitting and evaluation; introduced in Section 2.1.
  • domain assumption Auxiliary computational descriptors (VarvYS, SFE, VEC, depth of penetration) carry useful information for predicting the main experimental properties, despite having unquantified model-form uncertainty.
    The HDGP P-All configuration's advantage is attributed to auxiliary tasks; Section 2.3 and Tables 2 and 3.
  • standard math Variational inference for DGPs (ELBO optimization, 10 latent GPs, two layers) gives a faithful approximation of the true posterior for this dataset.
    The DGP implementation relies on BoTorch's variational approximations; Section 4.3. This is standard but not exact.
  • ad hoc to paper The encoder-decoder prior is deterministic and its prediction is re-added at test time; no uncertainty from the prior is propagated into the final predictive variance.
    The paper treats the prior as a fixed function; the predictive variance reported by the DGP is therefore only the residual variance, not the total uncertainty (Section 2.3, Figure 5).

how reviews work

0 comments
Cite this review

Pith. "Pith review of Accurate and Uncertainty-Aware Multi-Task Prediction of HEA Properties Using Prior-Guided Deep Gaussian Processes." pith.science (2026). https://pith.science/paper/DF4K2OGZ

@misc{pith2026250614828,
  author       = {Pith},
  title        = {Pith review of: Accurate and Uncertainty-Aware Multi-Task Prediction of HEA Properties Using Prior-Guided Deep Gaussian Processes},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DF4K2OGZ}},
  note         = {Machine review of arXiv:2506.14828}
}
read the original abstract

Surrogate modeling techniques have become indispensable in accelerating the discovery and optimization of high-entropy alloys(HEAs), especially when integrating computational predictions with sparse experimental observations. This study systematically evaluates the fitting performance of four prominent surrogate models conventional Gaussian Processes(cGP), Deep Gaussian Processes(DGP), encoder-decoder neural networks for multi-output regression and XGBoost applied to a hybrid dataset of experimental and computational properties in the AlCoCrCuFeMnNiV HEA system. We specifically assess their capabilities in predicting correlated material properties, including yield strength, hardness, modulus, ultimate tensile strength, elongation, and average hardness under dynamic and quasi-static conditions, alongside auxiliary computational properties. The comparison highlights the strengths of hierarchical and deep modeling approaches in handling heteroscedastic, heterotopic, and incomplete data commonly encountered in materials informatics. Our findings illustrate that DGP infused with machine learning-based prior outperform other surrogates by effectively capturing inter-property correlations and input-dependent uncertainty. This enhanced predictive accuracy positions advanced surrogate models as powerful tools for robust and data-efficient materials design.

Figures

Figures reproduced from arXiv: 2506.14828 by the authors.

Figure 1
Figure 1. Schematic overview of the BIRDSHOT data acquisition workflow. The dataset results from a multi-institutional e [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Barycentric Plot of available experimental data points in Al–V–Cr–Mn–Fe–Co–Ni–Cu composition space. [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Pairwise plot illustrating correlations between di [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Schematic representation of the two-layered variational Deep Gaussian Process (DGP) architecture with Encoder-Decoder based [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 5
Figure 5. Figure 5: HDGP-P-All model results: (a) Predicted VarvYS at 298 K; (b) SFE calculation; (c) Average VEC; (d) Yield Strength (MPa); [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: (a) Without prior injection the points within the red circle are further from actual value. (b) With prior injection the points within [PITH_FULL_IMAGE:figures/full_fig_p010_6.png]
Figure 7
Figure 7. Figure 7: Schematic representation of the two-layered variational Deep Gaussian Process (DGP) architecture. Each variational Multi-task [PITH_FULL_IMAGE:figures/full_fig_p013_7.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

28 extracted references · 27 canonical work pages

  1. [1]

    D. B. Miracle, O. N. Senkov, A critical review of high entropy alloys and related concepts, Acta Materialia 122 (2017) 448–511

    work page 2017

  2. [2]

    E. P. George, D. Raabe, R. O. Ritchie, High-entropy alloys, Nature reviews materials 4 (8) (2019) 515–534

    work page 2019

  3. [3]

    Elkatatny, W

    S. Elkatatny, W. Abd-Elaziem, T. A. Sebaey, M. A. Darwish, A. Hamada, Machine-learning synergy in high-entropy alloys: A review, Journal of Materials Research and Technology (2024)

    work page 2024

  4. [4]

    C. Wen, Y . Zhang, C. Wang, D. Xue, Y . Bai, S. Antonov, L. Dai, T. Lookman, Y . Su, Machine learning assisted design of high entropy alloys with desired property, Acta Materialia 170 (2019) 109–117

    work page 2019

  5. [5]

    J. Wang, H. Kwon, H. S. Kim, B.-J. Lee, A neural network model for high entropy alloy design, npj Computational Materials 9 (1) (2023) 60

    work page 2023

  6. [6]

    T. Gao, J. Gao, S. Yang, L. Zhang, Data-driven design of novel lightweight refractory high-entropy alloys with superb hardness and corrosion resistance, npj Computational Materials 10 (1) (2024) 256

    work page 2024

  7. [7]

    C. E. Rasmussen, C. K. Williams, Gaussian processes for machine learning, MIT Press, 2006

    work page 2006

  8. [8]

    E. V . Bonilla, K. M. Chai, C. Williams, Multi-task gaussian process prediction, in: Advances in Neural Information Processing Systems, V ol. 20, 2008

    work page 2008

Show all 28 references
  1. [9]

    Damianou, N

    A. Damianou, N. D. Lawrence, Deep gaussian processes, in: Artificial Intelligence and Statistics, 2013, pp. 207–215

    2013

  2. [10]

    T. Chen, C. Guestrin, Xgboost: A scalable tree boosting system, in: Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 2016, pp. 785–794

    2016

  3. [11]

    Attari, R

    V . Attari, R. Arroyave, Decoding non-linearity and complexity: Deep tabular learning approaches for materials science, arXiv preprint arXiv:2411.18717 (2024). 14

    2024

  4. [12]

    Rao, P.-Y

    Z. Rao, P.-Y . Tung, R. Xie, W. Ye, H. Zhang, A. Ferrari, T. Klaver, F. Körmann, P. T. Sukumar, A. K. da Silva, Y . Chen, Z. Li, D. Ponge, J. Neugebauer, O. Gutfleisch, S. Bauer, D. Raabe, Machine learning– enabled high-entropy alloy discovery, Science 378 (6615) (2022) 78–85....

    2022 doi

  5. [13]

    Wang, et al., Multi-objective bayesian optimization for materials design and discovery, Journal of Materials Research 35 (7) (2020) 900–920

    H. Wang, et al., Multi-objective bayesian optimization for materials design and discovery, Journal of Materials Research 35 (7) (2020) 900–920

    2020

  6. [14]

    S. M. A. A. Alvi, J. Janssen, D. Khatamsaz, D. Perez, D. Allaire, R. Arroyave, Hierarchical gaussian process-based bayesian optimization for materials discovery in high entropy alloy spaces, Acta Materialia (2025) 120908

    2025

  7. [15]

    Khatamsaz, B

    D. Khatamsaz, B. Vela, R. Arroyave, Multi-objective bayesian alloy design using multi-task gaussian processes, Materials Letters 351 (2023) 135067

    2023

  8. [16]

    Y . Ban, J. Hou, X. Wang, G. Zhao, An effective multitask neural network for predicting mechanical prop- erties of steel, SSRN Preprint 4539167 (2023)

    2023

  9. [17]

    Khatamsaz, A

    D. Khatamsaz, A. Molkeri, R. Couperthwaite, J. James, R. Arroyave, D. Allaire, A. Srivastava, Efficiently exploiting process-structure-property relationships in material design by multi-information source fusion, Acta Materialia 206 (2021) 116619

    2021

  10. [18]

    S. F. Ghoreishi, A. Molkeri, R. Arroyave, D. Allaire, A. Srivastava, Efficient use of multiple information sources in material design, Acta Materialia 180 (2019) 260–271

    2019

  11. [19]

    T. Z. Khan, T. Kirk, G. Vazquez, P. Singh, A. V . Smirnov, D. D. Johnson, K. Youssef, R. Arróyave, Towards stacking fault energy engineering in FCC high entropy alloys, Acta Materialia 224 (2022) 117472

    2022

  12. [20]

    Varvenne, A

    P. Varvenne, A. Luque, W. A. Curtin, Theory of strengthening in fcc high entropy alloys, Acta Materialia 118 (2016) 164–176.doi:10.1016/j.actamat.2016.07.040

    2016 doi

  13. [21]

    M. A. Álvarez, L. Rosasco, N. D. Lawrence, Kernels for vector-valued functions: A review, Foundations and Trends®in Machine Learning 4 (3) (2012) 195–266

    2012

  14. [22]

    Salimbeni, M

    H. Salimbeni, M. Deisenroth, Doubly stochastic variational inference for deep gaussian processes, in: Advances in Neural Information Processing Systems, 2017, pp. 4588–4599

    2017

  15. [23]

    Hastings, M

    T. Hastings, M. Mulukutla, D. Khatamsaz, D. Salas, W. Xu, D. Lewis, N. Person, M. Skokan, B. Miller, J. Paramore, et al., Accelerated multi-objective alloy discovery through efficient bayesian methods: Ap- plication to the fcc high entropy alloy space

  16. [24]

    Mulukutla, A

    M. Mulukutla, A. N. Person, S. V oigt, L. Kuettner, B. Kappes, D. Khatamsaz, R. Robinson, D. S. Mula, W. Xu, D. Lewis, et al., Illustrating an effective workflow for accelerated materials discovery, Integrating Materials and Manufacturing Innovation (2024) 1–21

    2024

  17. [25]

    Arroyave, D

    R. Arroyave, D. Khatamsaz, B. Vela, R. Couperthwaite, A. Molkeri, P. Singh, D. D. Johnson, X. Qian, A. Srivastava, D. Allaire, A perspective on bayesian methods applied to materials discovery and design, MRS Communications 12 (6) (2022) 1037–1049

    2022

  18. [26]

    B. Vela, D. Khatamsaz, C. Acemi, I. Karaman, R. Arróyave, Data-augmented modeling for yield strength of refractory high entropy alloys: A bayesian approach, Acta Materialia 261 (2023) 119351

    2023

  19. [27]

    Hensman, N

    J. Hensman, N. Fusi, N. D. Lawrence, Gaussian processes for big data, in: Uncertainty in Artificial Intel- ligence, 2013

    2013

  20. [28]

    Balandat, B

    M. Balandat, B. Karrer, D. R. Jiang, S. Daulton, B. Letham, A. G. Wilson, E. Bakshy, Botorch: A frame- work for efficient monte-carlo bayesian optimization, in: Advances in Neural Information Processing Systems, V ol. 33, 2020, pp. 21524–21538. 15

    2020

Pith tools

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