Pith. sign in

REVIEW 5 major objections 6 minor 38 references

Prob-cGAN: A Probabilistic Conditional Generative Adversarial Network for LSD1 Inhibitor Activity Prediction

T0 review · 5 major / 6 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read Prob-cGAN, a probabilistic conditional GAN with noise-injected generation, predicts LSD1 inhibitor activity with reported R2 0.8131 and RMSE 0.4838, beating Random Forest and Smiles-Transformer on ChEMBL data.

desk verdict The two headline results disagree with each other, so the claimed outperformance is currently unsupported; worth engaging only if the authors release code and a consistent set of numbers. read the letter →

arxiv 2411.15483 v1 pith:R625V3Y2 submitted 2024-11-23 cs.CE

Hanyang Wang This is my paper · ORCID
classification cs.CE
keywords LSD1conditionalgenerativeadversarialnetworkactivitypredictionprobabilisticregressionSMILESMorganfingerprintChEMBLQSAR
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 proposes Prob-cGAN, a probabilistic conditional generative adversarial network that predicts the pChEMBL activity of LSD1 inhibitors from a combination of Morgan fingerprints and SMILES word embeddings. On 931 ChEMBL compounds split 80/20, the model reports an $R^2$ of 0.8131 and an RMSE of 0.4838, outperforming Random Forest, SVR, XGBoost, and a Smiles-Transformer baseline. The authors argue that injecting noise into the generator and using an f-GAN discriminator lets the model absorb distributional shift and output an uncertainty estimate rather than a single point. If this holds, it gives drug-discovery teams a way to rank candidate LSD1 inhibitors with a measure of confidence on small, noisy datasets.

What carries the argument

The load-bearing object is the Prob-cGAN architecture itself. It starts from a fused descriptor: 512-bit Morgan fingerprints concatenated with 300-dimensional Word2Vec SMILES embeddings (812 dimensions total), which an autoencoder compresses to 203 dimensions. The generator concatenates a random noise vector $z$ with the hidden representation at each layer, so the network learns where noise best represents uncertainty; the discriminator uses separate pathways for true and generated samples and is trained with an f-GAN loss, a variational divergence-minimization objective. Together these pieces let the model produce a conditional distribution of activity rather than a single point estimate, which is the mechanism the paper credits for its resilience to distribution shift.

What would settle it

Run the same pipeline with scaffold-based splitting or repeated stratified cross-validation on the same 931 compounds; if the $R^2$ drops near the Random Forest or cGAN baseline under splits that separate chemical series, the claimed generalization advantage is not supported. Also, reproducing the exact configuration should yield one consistent $R^2$; the paper currently reports 0.8131 in Table 1 and 0.739 in the abstract, so a reader can check which number a faithful re-run produces.

Watch

Extended reading notes

Core claim

The paper's central discovery is that recasting inhibitor-activity regression as probabilistic conditional generation—where the generator receives noise at every layer and a dual-path f-GAN discriminator scores real versus generated activity—yields markedly better predictions of LSD1 inhibitor activity than standard regressors. On a random 80/20 split of 931 ChEMBL compounds, the authors report $R^2 = 0.8131$ and RMSE = 0.4838, versus 0.6686 and 0.6441 for Random Forest and 0.6097 and 0.6990 for Smiles-Transformer; the abstract reports a separate best result of $R^2 = 0.739$ and RMSE = 0.562 against the same baselines. The authors attribute the improvement to the model's ability to approximate unknown conditional densities and to its probabilistic output, which gives an uncertainty estimate for each predicted activity.

Load-bearing premise

The evaluation assumes that one random 80/20 split of 931 ChEMBL compounds yields a test set from the same distribution as training, so the reported $R^2$ reflects generalization rather than shared scaffolds or assay artifacts.

Editorial extensions

If this is right

  • If the claimed accuracy transfers, Prob-cGAN can rank LSD1 inhibitor candidates by predicted pChEMBL before synthesis, reducing the number of compounds that need experimental testing.
  • The probabilistic output gives each prediction an uncertainty, which is useful for prioritizing compounds whose predictions are most reliable.
  • The architecture's tolerance for unknown density functions means it can be applied to other small epigenetic datasets where activity distributions are irregular.
  • Combining structural fingerprints with SMILES text embeddings appears to matter: ablations removing word embeddings and the autoencoder drop $R^2$ to 0.258, so the fusion pipeline is a key component.

Reading between the lines

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

  • The reported superiority rests on a single random split; a scaffold split or repeated cross-validation would test whether the advantage survives on chemical series absent from training.
  • Because the abstract and Section 5 give different headline numbers (0.739/0.562 versus 0.8131/0.4838), a faithful re-run needs to state which configuration produced which; the mechanism claim is independent of that discrepancy but the precise magnitude is not.
  • If noise injection is the active ingredient, the same architecture should generalize to other epigenetic targets with small datasets; comparing Prob-cGAN against Random Forest on, say, EZH2 or HDAC inhibitors would be a direct transfer test.
  • The f-GAN divergence choice is not isolated in the ablation; swapping in a different divergence term would show whether the probabilistic GAN framing or the specific divergence drives the gain.
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

5 major / 6 minor

Summary. This paper proposes Prob-cGAN, a conditional generative adversarial network with a noise-injection generator and f-GAN discriminator, for predicting LSD1 inhibitor activity (pChEMBL) from a dataset of 931 ChEMBL compounds. Input features combine Morgan fingerprints and Word2Vec embeddings, reduced by an autoencoder to 203 dimensions. The authors report that Prob-cGAN substantially outperforms several machine learning baselines and a Smiles-Transformer. The paper does not provide code, data, hyperparameters, or a detailed training protocol.

Significance. The proposed application of a cGAN with probabilistic outputs to QSAR-style activity prediction is topical, and the ablation study (Table 2) attempts to quantify contributions of autoencoder and word embedding components. If the reported results were reproducible and robust to scaffold splits, the method could be a useful addition to the LSD1 modeling toolbox. However, the significance as presented is undercut by an unresolved internal inconsistency in the headline numbers and by an evaluation protocol that does not support the claim of superior generalization. No code or data are released, so the empirical contribution cannot be verified or built upon by other researchers.

major comments (5)
  1. [Abstract vs. Section 4.1/Table 1] The paper reports two incompatible sets of headline results for the same comparison. The abstract states that Prob-cGAN achieves a top-1 R2 of 0.739 and RMSE of 0.562, against Smiles-Transformer (0.591/0.708) and cGAN (0.488/0.791). Table 1 and Section 5.1 report Prob-cGAN at 0.8131 (±0.0008) and 0.4838 (±0.0010), and Smiles-Transformer at 0.6097 (±0.0078) and 0.6990 (±0.0070). The cGAN numbers in Table 2 (0.488/0.791) agree with the abstract but the Prob-cGAN and Smiles-Transformer numbers differ by up to 0.074 in R2 and 0.078 in RMSE. The concept of 'top-1' is never defined, and no explanation is offered for which table the abstract is quoting. Because the paper's central claim is that Prob-cGAN is superior, and the two sets of numbers support different effect sizes, the claim as presented is not supported by a stable evidence base. The authors must reconcile these numbers and specify the exact experimental condition (e.g., best seed vs. mean over runs) for each reported metric.
  2. [Section 4] The evaluation uses a single random 80/20 split of 931 ChEMBL compounds, with no scaffold split, temporal split, or repeated cross-validation. Random splitting of a set of inhibitors commonly overestimates performance on new chemical series because structurally similar compounds appear in both training and test sets. The claim in Section 6 that Prob-cGAN is a 'powerful instrument for molecular property predictions' requires demonstration on a scaffold-based or external test set; otherwise the reported R2 cannot be taken as evidence of generalization to novel LSD1 inhibitors.
  3. [Sections 2-4 / reproducibility] The manuscript does not provide code, data, hyperparameter values, random seeds, or training details for any of the models, including the deep baselines. The feature pipeline (Morgan fingerprint radius 3, length 512, Word2Vec dimension 300, autoencoder dimension 203) is described, but the network architectures, learning rates, batch sizes, number of epochs, and regularization are omitted. The reported standard deviations (e.g., 0.0008 for Prob-cGAN in Table 1) cannot be interpreted without knowing how many runs and which seeds they are over. This is below the reproducibility standard expected for a machine-learning benchmark paper.
  4. [Section 3] The paper advertises 'probabilistic predictions' and uncertainty estimation as the key novelty, but the evaluation never assesses these outputs. No calibration plot, coverage probability, interval width, or any uncertainty metric is reported; only point-estimate R2 and RMSE are given. To substantiate the probabilistic claim, the authors should specify how the predictive distribution is derived from the generator and evaluate it with a proper scoring rule or calibration test.
  5. [Table 1] Most baselines in Table 1 (SVR, Ridge, Random Forest, Decision Tree, XGBoost, MLP, K-Neighbours) are reported as single values with no standard deviation or confidence interval, while Prob-cGAN is reported as a mean over some runs. This makes the 'significantly outperforms' statement in Section 5.1 unsupported; without error bars on the baselines, even if the mean differences are large, one cannot assess statistical significance. In addition, the manuscript does not compare against previously published LSD1-specific deep learning models (e.g., the GCN and transfer-learning models cited in the introduction), which limits the benchmark's completeness.
minor comments (6)
  1. [Section 5.1] Section 5.1 repeatedly refers to 'Table 3', but the actual results appear in Table 1; Table 3 does not exist in the manuscript.
  2. [Abstract] The term 'top-1 R2' is undefined; is it the best performance over multiple runs or something else?
  3. [Throughout] The spelling of the model name is inconsistent ('Prob-cGAN', 'prob-cGAN', 'proposed prob-cGAN'); please unify.
  4. [Section 5.1 heading] The section heading contains a typo: 'Comparsion' should be 'Comparison'.
  5. [Section 2] Section 2 states that a pre-trained Word2Vec model is used, but gives no details about the training corpus or model version; this feature cannot be reproduced by readers.
  6. [Section 4] Section 4 contains a grammatical fragment: 'the corresponding biological activity, To ensure that the dataset is suitable...' — please rephrase the sentence.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: this is an empirical benchmark whose claims rest on external data and standard baselines, not on arguments equivalent to their own inputs.

full rationale

Prob-cGAN (arXiv:2411.15483) is an empirical machine-learning benchmark, not a derivation. I inspected the manuscript for all seven circularity patterns. There is no equation in which an output variable is defined in terms of the predicted quantity, no parameter fitted to a subset and then renamed as a prediction, no load-bearing self-citation, no imported uniqueness theorem, no ansatz smuggled in via citation, and no re-labeling of a known result as a new organization. The model pipeline is described concretely: Morgan fingerprints are combined with Word2Vec embeddings, an autoencoder reduces the 812-dimensional features to 203 dimensions, and a generator with noise injection plus an f-GAN discriminator is trained to predict pChEMBL values. The evaluation compares the trained model against SVR, Ridge, Random Forest, Decision Tree, XGBoost, MLP, K-Neighbours, Smiles-Transformer, and a cGAN ablation using R2 and RMSE on a single random 80/20 split of ChEMBL data. These are external, falsifiable comparisons, so the central claim cannot reduce by construction to the paper's own assumptions. The most serious issues are the unexplained inconsistency between the abstract (R2=0.739, RMSE=0.562) and Table 1 (R2=0.8131, RMSE=0.4838) for the same comparison, the single random split without scaffold splitting or repeated cross-validation, absent code/data, and the asserted but unevaluated probabilistic output. Those are correctness and reproducibility concerns, not circularity. No circular step can be quoted because none exists; the manuscript is self-contained against external benchmarks, so the honest finding is a score of 0.

Assumptions & free parameters 4 free parameters · 3 assumptions · 0 invented entities

The central claim rests on data quality, the representativeness of a single random split, and implicit convergence of the GAN. All architecture dimensions are hand-set without sensitivity analysis, and no training hyperparameters are reported.

free parameters (4)
  • Morgan fingerprint dimensions = L=512, radius=3
    Chosen in Section 2 without sensitivity analysis; the central comparison may depend on these hand-set values.
  • Word2Vec embedding dimension = 300
    Uses a pretrained model; the source and training corpus are not specified, so the embedding may not be reproducible.
  • Autoencoder latent dimension = 203
    Reduces 812 to 203 dimensions, chosen by hand; no ablation shows this is optimal, and the ablation drops it entirely rather than varying it.
  • Neural network hyperparameters
    No learning rate, batch size, epochs, optimizer, or random seed is reported; every training choice is an unreported free parameter.
assumptions (3)
  • domain assumption ChEMBL v28 labels for the 931 LSD1 compounds are accurate.
    Section 4 describes filtering but not the exact criteria; if labels are noisy the R2 comparison is meaningless.
  • domain assumption The 80/20 random split is representative.
    Section 4: 'randomly divided into two subsets'; without scaffold splitting, test molecules may closely resemble training molecules.
  • domain assumption GAN training converges to a useful conditional predictor.
    The paper asserts f-GAN training works but gives no convergence checks, mode collapse analysis, or sensitivity to initialization.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Prob-cGAN: A Probabilistic Conditional Generative Adversarial Network for LSD1 Inhibitor Activity Prediction." pith.science (2026). https://pith.science/paper/R625V3Y2

@misc{pith2026241115483,
  author       = {Pith},
  title        = {Pith review of: Prob-cGAN: A Probabilistic Conditional Generative Adversarial Network for LSD1 Inhibitor Activity Prediction},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/R625V3Y2}},
  note         = {Machine review of arXiv:2411.15483}
}
abstract

The inhibition of Lysine-Specific Histone Demethylase 1 (LSD1) is a promising strategy for cancer treatment and targeting epigenetic mechanisms. This paper introduces a Probabilistic Conditional Generative Adversarial Network (Prob-cGAN), designed to predict the activity of LSD1 inhibitors. The Prob-cGAN was evaluated against state-of-the-art models using the ChEMBL database, demonstrating superior performance. Specifically, it achieved a top-1 $R^2$ of 0.739, significantly outperforming the Smiles-Transformer model at 0.591 and the baseline cGAN at 0.488. Furthermore, it recorded a lower $RMSE$ of 0.562, compared to 0.708 and 0.791 for the Smiles-Transformer and cGAN models respectively. These results highlight the potential of Prob-cGAN to enhance drug design and advance our understanding of complex biological systems through machine learning and bioinformatics.

Figures

Figures reproduced from arXiv: 2411.15483 by the authors.

Figure 1
Figure 1. Overview of the proposed Probabilistic Conditional Generative Adversarial Networks (CGAN) with molecule [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Model results of activity prediction. 5 Result 5.1 Result: Comparsion against Machine Learning Algorithms The proposed Prob-cGAN is compared with state-of-the-art models for the prediction of the activity of the LSD1 inhibitor. The results are summarized in [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Model results of activity prediction distribution [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

38 extracted references · 29 canonical work pages

  1. [1]

    Snapshot: histone modifications

    He Huang, Benjamin R Sabari, Benjamin A Garcia, C David Allis, and Yingming Zhao. Snapshot: histone modifications. Cell, 159(2):458–458, 2014

    work page 2014

  2. [2]

    Histone demethylation mediated by the nuclear amine oxidase homolog lsd1

    Yujiang Shi, Fei Lan, Caitlin Matson, Peter Mulligan, Johnathan R Whetstine, Philip A Cole, Robert A Casero, and Yang Shi. Histone demethylation mediated by the nuclear amine oxidase homolog lsd1. Cell, 119(7):941–953, 2004

    work page 2004

  3. [3]

    Opposing lsd1 complexes function in developmental gene activation and repression programmes

    Jianxun Wang, Kathleen Scully, Xiaoyan Zhu, Ling Cai, Jie Zhang, Gratien G Prefontaine, Anna Krones, Kenneth A Ohgi, Ping Zhu, Ivan Garcia-Bassets, et al. Opposing lsd1 complexes function in developmental gene activation and repression programmes. Nature, 446(7138):882–887, 2007

    work page 2007

  4. [4]

    Emerging roles of epigenetic regulator sin3 in cancer

    N Bansal, G David, E Farias, and S Waxman. Emerging roles of epigenetic regulator sin3 in cancer. Advances in cancer research, 130:113–135, 2016

    work page 2016

  5. [5]

    Lsd1 inhibition suppresses the growth of clear cell renal cell carcinoma via upregulating p21 signaling

    Liangsong Zhu, Jianfeng Wang, Wen Kong, Jiwei Huang, Baijun Dong, Yiran Huang, Wei Xue, and Jin Zhang. Lsd1 inhibition suppresses the growth of clear cell renal cell carcinoma via upregulating p21 signaling. Acta Pharmaceutica Sinica B, 9(2):324–334, 2019

    work page 2019

  6. [6]

    Functional interplay between histone demethylase and deacetylase enzymes

    Min Gyu Lee, Christopher Wynder, Daniel A Bochar, Mohamed-Ali Hakimi, Neil Cooch, and Ramin Shiekhattar. Functional interplay between histone demethylase and deacetylase enzymes. Molecular and cellular biology, 26(17):6395–6402, 2006

    work page 2006

  7. [7]

    The re-expression of the epigenetically silenced e-cadherin gene by a polyamine analogue lysine-specific demethylase-1 (lsd1) inhibitor in human acute myeloid leukemia cell lines

    Tracy Murray-Stewart, Patrick M Woster, and Robert A Casero. The re-expression of the epigenetically silenced e-cadherin gene by a polyamine analogue lysine-specific demethylase-1 (lsd1) inhibitor in human acute myeloid leukemia cell lines. Amino acids, 46:585–594, 2014

    work page 2014

  8. [8]

    Inhibition of lysine-specific demethylase 1 by polyamine analogues results in reexpression of aberrantly silenced genes

    Yi Huang, Eriko Greene, Tracy Murray Stewart, Andrew C Goodwin, Stephen B Baylin, Patrick M Woster, and Robert A Casero Jr. Inhibition of lysine-specific demethylase 1 by polyamine analogues results in reexpression of aberrantly silenced genes. Proceedings of the National Academy of Sciences, 104(19):8023–8028, 2007

    work page 2007

Show all 38 references
  1. [9]

    Quantitative structure-activity relationship methods: Perspectives on drug discovery and toxicology

    Roger Perkins, Hong Fang, Weida Tong, and William J Welsh. Quantitative structure-activity relationship methods: Perspectives on drug discovery and toxicology. Environmental Toxicology and Chemistry: An International Journal, 22(8):1666–1679, 2003

    2003

  2. [10]

    Discovery of new inhibitors of aldose reductase from molecular docking and database screening

    Giulio Rastelli, Anna Maria Ferrari, Luca Costantino, and Maria Cristina Gamberini. Discovery of new inhibitors of aldose reductase from molecular docking and database screening. Bioorganic & medicinal chemistry, 10(5):1437– 1450, 2002

    2002

  3. [11]

    Investigating the binding mode of reversible lsd1 inhibitors derived from stilbene derivatives by 3d-qsar, molecular docking, and molecular dynamics simulation

    Yongtao Xu, Zihao He, Min Yang, Yunlong Gao, Linfeng Jin, Meiting Wang, Yichao Zheng, Xiaoyuan Lu, Songjie Zhang, Chang Wang, et al. Investigating the binding mode of reversible lsd1 inhibitors derived from stilbene derivatives by 3d-qsar, molecular docking, and molecular dyna...

    2019

  4. [12]

    Investi- gating the binding mechanism of (4-cyanophenyl) glycine derivatives as reversible lsd1 by 3d-qsar, molecular docking and molecular dynamics simulations

    Zhi-Zheng Wang, Chao-Ya Ma, Jing Yang, Qi-Bing Gao, Xu-Dong Sun, Lina Ding, and Hong-Min Liu. Investi- gating the binding mechanism of (4-cyanophenyl) glycine derivatives as reversible lsd1 by 3d-qsar, molecular docking and molecular dynamics simulations. Journal of Molecular ...

    2019

  5. [13]

    Inductive transfer learning for molecular activity prediction: Next-gen qsar models with molpmofit

    Xinhao Li and Denis Fourches. Inductive transfer learning for molecular activity prediction: Next-gen qsar models with molpmofit. Journal of Cheminformatics, 12(1):1–15, 2020

    2020

  6. [14]

    Prediction of pharmacological activities from chemical structures with graph convolutional neural networks

    Miyuki Sakai, Kazuki Nagayasu, Norihiro Shibui, Chihiro Andoh, Kaito Takayama, Hisashi Shirakawa, and Shuji Kaneko. Prediction of pharmacological activities from chemical structures with graph convolutional neural networks. Scientific reports, 11(1):525, 2021

    2021

  7. [15]

    A structure- based platform for predicting chemical reactivity

    Frederik Sandfort, Felix Strieth-Kalthoff, Marius Kühnemund, Christian Beecks, and Frank Glorius. A structure- based platform for predicting chemical reactivity. Chem, 6(6):1379–1390, 2020. 7

    2020

  8. [16]

    Neural word embedding as implicit matrix factorization

    Omer Levy and Yoav Goldberg. Neural word embedding as implicit matrix factorization. Advances in neural information processing systems, 27, 2014

    2014

  9. [17]

    Towards a universal smiles representation-a standard method to generate canonical smiles based on the inchi

    Noel M O’Boyle. Towards a universal smiles representation-a standard method to generate canonical smiles based on the inchi. Journal of cheminformatics, 4:1–14, 2012

    2012

  10. [18]

    Decision tree methods: applications for classification and prediction

    Yan-Yan Song and LU Ying. Decision tree methods: applications for classification and prediction. Shanghai archives of psychiatry, 27(2):130, 2015

    2015

  11. [19]

    What is a support vector machine? Nature biotechnology, 24(12):1565–1567, 2006

    William S Noble. What is a support vector machine? Nature biotechnology, 24(12):1565–1567, 2006

    2006

  12. [20]

    Xgboost: extreme gradient boosting

    Tianqi Chen, Tong He, Michael Benesty, Vadim Khotilovich, Yuan Tang, Hyunsu Cho, Kailong Chen, Rory Mitchell, Ignacio Cano, Tianyi Zhou, et al. Xgboost: extreme gradient boosting. R package version 0.4-2 , 1(4):1–4, 2015

    2015

  13. [21]

    Probabilistic regression using conditional generative adversarial networks, 2020

    Joel Oskarsson. Probabilistic regression using conditional generative adversarial networks, 2020

    2020

  14. [22]

    Benchmarking regression methods: A comparison with cgan

    Karan Aggarwal, Matthieu Kirchmeyer, Pranjul Yadav, S Sathiya Keerthi, and Patrick Gallinari. Benchmarking regression methods: A comparison with cgan. arXiv preprint arXiv:1905.12868, 2019

    1905 arXiv

  15. [23]

    Word2vec

    Kenneth Ward Church. Word2vec. Natural Language Engineering, 23(1):155–162, 2017

    2017

  16. [24]

    Greedy layer-wise training of deep networks

    Yoshua Bengio, Pascal Lamblin, Dan Popovici, and Hugo Larochelle. Greedy layer-wise training of deep networks. Advances in neural information processing systems, 19, 2006

    2006

  17. [25]

    Review on studies of machine learning algorithms

    Peiyuan Xu. Review on studies of machine learning algorithms. In Journal of Physics: Conference Series, volume 1187, page 052103. IOP Publishing, 2019

    2019

  18. [26]

    The chembl database in 2017

    Anna Gaulton, Anne Hersey, Michał Nowotka, A Patricia Bento, Jon Chambers, David Mendez, Prudence Mutowo, Francis Atkinson, Louisa J Bellis, Elena Cibrián-Uhalte, et al. The chembl database in 2017. Nucleic acids research, 45(D1):D945–D954, 2017

    2017

  19. [27]

    Advantages of the mean absolute error (mae) over the root mean square error (rmse) in assessing average model performance

    Cort J Willmott and Kenji Matsuura. Advantages of the mean absolute error (mae) over the root mean square error (rmse) in assessing average model performance. Climate research, 30(1):79–82, 2005

    2005

  20. [28]

    Applied regression analysis, volume 326

    Norman R Draper and Harry Smith. Applied regression analysis, volume 326. John Wiley & Sons, 1998

    1998

  21. [29]

    Assessing model fit by cross-validation

    Douglas M Hawkins, Subhash C Basak, and Denise Mills. Assessing model fit by cross-validation. Journal of chemical information and computer sciences, 43(2):579–586, 2003

    2003

  22. [30]

    Classification and regression by randomforest

    Andy Liaw, Matthew Wiener, et al. Classification and regression by randomforest. R news, 2(3):18–22, 2002

    2002

  23. [31]

    Generative adversarial nets

    Ian Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron Courville, and Yoshua Bengio. Generative adversarial nets. Advances in neural information processing systems, 27, 2014

    2014

  24. [32]

    Auto-encoding variational bayes

    Diederik P Kingma and Max Welling. Auto-encoding variational bayes. arXiv preprint arXiv:1312.6114, 2013

    2013 arXiv

  25. [33]

    Smiles transformer: Pre-trained molecular fingerprint for low data drug discovery

    Shion Honda, Shoi Shi, and Hiroki R Ueda. Smiles transformer: Pre-trained molecular fingerprint for low data drug discovery. arXiv preprint arXiv:1911.04738, 2019

    1911 arXiv

  26. [34]

    The rise of deep learning in drug discovery

    Hongming Chen, Ola Engkvist, Yinhai Wang, Marcus Olivecrona, and Thomas Blaschke. The rise of deep learning in drug discovery. Drug discovery today, 23(6):1241–1250, 2018

    2018

  27. [35]

    Deep learning in drug discovery

    Erik Gawehn, Jan A Hiss, and Gisbert Schneider. Deep learning in drug discovery. Molecular informatics, 35(1):3–14, 2016

    2016

  28. [36]

    f-gan: Training generative neural samplers using variational divergence minimization

    Sebastian Nowozin, Botond Cseke, and Ryota Tomioka. f-gan: Training generative neural samplers using variational divergence minimization. Advances in neural information processing systems, 29, 2016

    2016

  29. [37]

    Reducing the dimensionality of data with neural networks

    Geoffrey E Hinton and Ruslan R Salakhutdinov. Reducing the dimensionality of data with neural networks. science, 313(5786):504–507, 2006

    2006

  30. [38]

    Convolutional networks on graphs for learning molecular fingerprints

    David K Duvenaud, Dougal Maclaurin, Jorge Iparraguirre, Rafael Bombarell, Timothy Hirzel, Alán Aspuru-Guzik, and Ryan P Adams. Convolutional networks on graphs for learning molecular fingerprints. Advances in neural information processing systems, 28, 2015. 8

    2015

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.