Pith. sign in

REVIEW 5 major objections 4 minor 43 references

MetaMolGen: A Neural Graph Motif Generation Model for De Novo Molecular Design

T0 review · 5 major / 4 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read This paper claims that a meta-learned SMILES generator combining Reptile and Conditional Neural Processes can generate valid, diverse, property-aligned molecules from as few as 1,000 training samples.

desk verdict Few-shot validity gain is real, but the property-control claim is contradicted by the paper's own Table II. read the letter →

arxiv 2504.15587 v2 pith:W6ITXSVO submitted 2025-04-22 cs.LG cs.AI

Zimo Yan , Jie Zhang , Zheng Xie , Chang Liu , Yizhen Liu , Yiping Song This is my paper · ORCID
classification cs.LGcs.AI
keywords moleculargenerationfew-shotlearningmeta-learningReptilealgorithmConditionalNeuralProcessesSMILESproperty-conditionedfeaturestandardization
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

MetaMolGen claims that a molecular generator can be made data-efficient by treating each small molecular dataset as a meta-learning task: a Reptile-style first-order update learns a shared initialization, a Conditional Neural Process context encoder turns a handful of molecules into a task representation, and a learnable standardization layer keeps feature scales stable. The paper's experiments, on subsets of 1,000 to 60,000 molecules from ChEMBL, QM9, ZINC, and MOSES, report that this design yields higher validity, uniqueness, drug-likeness, and an overall score above RNN and MolGPT baselines, while generating in 0.05 hours. It also claims property-conditioned generation: a lightweight MLP projects target properties such as LogP, TPSA, QED, and SAS into the decoder's initial state, producing molecules with reported property similarities of 0.98 to 1.00 for four drug targets. A sympathetic reader would care because data-scarce settings such as orphan drug discovery are exactly where existing generative models demand millions of training examples.

What carries the argument

The mechanism is a pipeline of four coupled pieces: a learnable standardization layer, $X'=(X-\mu_\theta)/(\sigma_\theta+\epsilon)$, that whitens molecular descriptors; a CNP-style context encoder that mean-pools encoded support molecules into a task embedding; the Reptile update $\theta \leftarrow \theta + \epsilon \cdot \frac{1}{M}\sum_m(\theta'_m-\theta)$, which moves the shared initialization toward parameters that adapt in a few inner steps; and an MLP property projector that injects the target property vector into the initial hidden state of a two-layer LSTM SMILES decoder. The standardization layer is the component whose removal the ablation identifies as degrading validity, novelty, diversity, and conditional success, and Reptile is what nominally supplies the few-shot adaptation.

What would settle it

Run the conditional-generation protocol on property targets outside the training distribution, for example an extreme TPSA or a multimodal LogP mixture, and check whether validity stays near 75% and property mean absolute deviation remains comparable to MolGPT; the paper's own limitation section says new property domains may require retraining, so a failed transfer would directly refute the claimed generalization of the conditioning mechanism.

Watch

Extended reading notes

Core claim

MetaMolGen's central claim is that data scarcity in molecular design can be attacked by meta-learning rather than by larger corpora. The model turns each small molecular dataset into a task, adapts a shared initialization with the first-order Reptile update, aggregates standardized molecular descriptors through a Conditional Neural Process context encoder, and decodes SMILES with a two-layer LSTM. The paper reports that this recipe raises validity in the 1,000-sample regime to 43.6%, against 34.8% for RNN and 1.6% for MolGPT, reaches 83.6% validity at 10,000 samples, keeps uniqueness above 99.9%, and posts an Overall Score of 0.7143, above MolGAN's 0.5419 and MolGPT's 0.4836, while generating 40 to 200 times faster. For property conditioning, a lightweight MLP projector injects the target property vector into the decoder's initial hidden state, and the paper reports mean absolute errors of 0.15 in LogP and 0.32 in molecular weight, with average property similarities of 0.98 to 1.00 across four drug targets; it also reports that TPSA control is looser, with a mean absolute deviation of 123.4 versus MolGPT's 3.52. The paper presents this as evidence that fast adaptation plus feature normalization, not training-set scale, is the route to few-shot, property-aware molecular generation.

Load-bearing premise

The load-bearing premise is that an autoregressive SMILES generator nudged by a latent property vector will keep its outputs both syntactically valid and aligned with the target property, a premise the paper's own tables only partially confirm (validity 75.12%, TPSA mean absolute deviation 123.4 versus 3.52 for MolGPT).

Editorial extensions

If this is right

  • At 1,000 to 10,000 training molecules, MetaMolGen keeps validity between 43.6% and 83.6%, where RNN stays below 61% and MolGPT mostly below 26%, so few-shot molecular design becomes a viable target.
  • The validity/uniqueness trade-off is shifted: near-perfect uniqueness (99.92%) and high drug-likeness (0.82) are achieved at the price of a validity drop to 75.12% versus MolGAN's 97.4%, an exchange many drug-discovery pipelines would accept.
  • Property-conditioned generation works without post-hoc filtering: target-property similarities of 0.98 to 1.00 on Aspirin, Tamiflu, Amoxicillin, and Chloroquine, with reported conditional generation success rates above 95% under hydrogen-bond donor/acceptor constraints.
  • The standardization layer is load-bearing: ablation shows removing it lowers validity, novelty, diversity, and conditional success, with the largest gaps in the 1,000 to 6,000-sample regime.
  • Generation time drops to 0.05 hours versus 10.2 hours for ORGAN and 2.12 hours for MolGAN, making the model practical for iterative design loops.

Reading between the lines

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

  • Because the paper's own Table II shows TPSA alignment is much looser than LogP or QED alignment, a testable extension is to replace the single latent projection with per-property scaling or a small property-specific adapter, and check whether TPSA mean absolute deviation moves from 123.4 toward MolGPT's 3.52.
  • The meta-training uses random subsets as tasks; a harder and more realistic test would build tasks from disjoint scaffolds, since the paper concedes that task-specific adaptation is coarse-grained and that entirely new property domains may require retraining.
  • The observed unimodal property distributions suggest much of the few-shot gain may come from standardization plus a well-behaved loss surface, so an ablation that swaps the learnable standardization for ordinary z-scoring would isolate how much of the improvement is meta-learning versus preprocessing.
  • A natural next step the paper leaves implicit is using the same CNP-plus-Reptile backbone for non-SMILES targets such as reaction pathways or polymer strings, where the property projector would condition on yield or bandgap instead of drug-likeness.
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 / 4 minor

Summary. MetaMolGen is presented as a first-order meta-learning (Reptile) framework for few-shot, property-conditioned SMILES generation. Molecular descriptors are standardized via a learnable normalization layer, encoded by a CNP-style context encoder, and decoded by an LSTM; a property projector is claimed to condition generation on target properties such as LogP, TPSA, SAS, and QED. The paper provides seven theorems on convergence and generalization, an Overall Score aggregating seven metrics, and experiments on ChEMBL, QM9, ZINC, and MOSES comparing MetaMolGen with ORGAN, MolGAN, RNN, and MolGPT. The central claims are that MetaMolGen balances validity and uniqueness, adapts efficiently under low data, and achieves accurate property matching.

Significance. The strongest credible result in the manuscript is the low-data validity trend in Table IV: MetaMolGen's validity rises from 43.6% at 1,000 training molecules to 83.6% at 10,000, consistently above RNN and MolGPT, and the standardization ablation in Figure 7 suggests a benefit in low-data settings. If the property-control results were as reported, the model would be a useful contribution to few-shot molecular design. However, the property-control claim is internally contradicted by Table II, several headline superiority claims are not supported by the reported diversity and uniqueness numbers, the Overall Score that carries the main comparison is defined relative to the models being compared, and the theoretical results that are used to explain the method's benefits are not proven. These are not presentation issues; they affect the paper's central claims.

major comments (5)
  1. [§VII-C, Table II, Figure 5] The conditional property-control claim in the abstract and Contribution 2 is contradicted by the paper's own Table II. For TPSA, MetaMolGen has MAD 123.40 versus MolGPT's 3.52; for LogP, MAD 1.54 versus 0.23; for QED, MAD 1.00 versus 0.056. Yet Section VII-C and Figure 5 report property similarities of 0.98-1.00, including 'Property Similarity: 1.0' for displayed molecules. A TPSA MAD of 123.4 means the generated molecules are far outside the target property range, so these two sets of numbers cannot both describe the same conditional generation setup. Section VIII-A's description of the gap as 'slightly higher' is also inconsistent with the reported magnitudes. Because property matching is listed as a core advantage, this contradiction directly undermines the central contribution.
  2. [Table IV, Figure 4, Table I] The abstract's claim that MetaMolGen 'consistently generates valid and diverse SMILES sequences ... outperforming conventional baselines' is not supported by the diversity and uniqueness columns. In Table IV, MetaMolGen diversity ranges from 0.8285 to 0.8415 across training sizes 1,000-10,000, while RNN ranges from 0.8536 to 0.8852 and MolGPT from 0.8481 to 0.8914; MetaMolGen is the least diverse model at every size, including the 5,000-molecule row in Table I (0.8337 versus 0.8742 for RNN and 0.8637 for MolGPT). Similarly, Table I reports MetaMolGen uniqueness 99.92%, below MolGPT's 100.00%. These numbers contradict Contribution 2's claim of significant improvement in uniqueness and diversity.
  3. [§IV-B, Eq. (3), Eq. (6)] The conditional generation mechanism is not defined or trained to enforce property alignment. The only training loss stated is token-level reconstruction (Eq. (3), L_recon); there is no term that penalizes mismatch between the generated molecule's properties and the conditioning vector z. The property projector is described only qualitatively in Section VII-C as an 'MLP-based property conditioning module,' with no equation, input/output specification, or loss contribution, and Eq. (6) merely factorizes the decoder distribution. In addition, the text says this formulation 'ensures syntactic validity,' but Table I reports validity of only 75.12% and Table IV reports 43.6% at 1,000 training molecules, so the claimed guarantee does not hold. The conditional property control results therefore lack a supporting mechanism in the method.
  4. [§VI-D, Eq. (9)] The Overall Score defined in Eq. (9) is computed by min-max normalizing each metric over the observed values of the baselines and MetaMolGen itself, with zero-range metrics arbitrarily set to 0.5. This makes the score relative to the particular set of models in the table and lets one extreme value dominate: MetaMolGen's generation time of 0.05 h versus ORGAN's 10.2 h (Table I) yields a near-max normalized time score by construction, mechanically inflating the Overall Score. Since Contribution 2 and multiple conclusions in Section VII-A rely on 'superior Overall Score,' this self-referential metric cannot support the claimed superiority. A fixed, pre-specified aggregation or a benchmark-provided normalization is needed.
  5. [§V-B, Appendix B, Theorems 2-4, 7] The theoretical analysis that is invoked to explain the benefits of standardization is not established. In the proof of Theorem 3, the key step (Appendix B, Eq. (24)) asserts that after normalization E[J^T J] approximately cI; this is exactly the claim needing proof and does not follow from unit-variance inputs. The conclusion that the condition number is drastically reduced and that the iteration complexity improves to O(log 1/epsilon) is therefore unsupported. Theorem 2's proof rests on an unproven approximation (Eq. (14)) relating gradient variance to Hessian eigenvalues and input variances, and Theorem 4 assumes without justification that empirical loss variance is the relevant quantity in the PAC-Bayes bound. Theorem 7 combines O(L/k), a PAC-Bayes term, a gradient-noise term, and an approximation term without stating the assumptions under which these four errors can be added. Standard results (Theorems 1 and 6) are correct but do not establish anything specific to MetaMolGen.
minor comments (4)
  1. [§VI-E] Section VI-E ends with 'please refer to https:.'; this is an incomplete URL and there is no code or data link, which hampers reproducibility.
  2. [§VII-C, Figure 5] The 'property similarity' metric is never defined; the paper should specify how it is computed and report it alongside Table II so that the reader can reconcile the two sets of numbers.
  3. [Eq. (1)] Equation (1) states that X' is distributed as N(0,I), but empirical standardization only yields zero mean and unit variance; it does not make the distribution Gaussian. The text should say approximately standardized or state a normality assumption.
  4. [Appendix B, Lemma 2] Lemma 2 (Stability of Task Encoder) is not used in any subsequent proof; either use it in the analysis or remove it to avoid an unused lemma.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the generative model is trained with a standard reconstruction objective and the central claims are tested against external baselines.

full rationale

The derivation chain is self-contained. MetaMolGen is trained by token-level cross-entropy (Eq. 3) over SMILES with Reptile-style updates (Eqs. 4-5), and inference is a standard autoregressive factorization (Eq. 6). No parameter is fitted to the evaluation metrics and then reported as a prediction; validity, uniqueness, and diversity are computed directly on generated SMILES and compared with independent baselines (Table I, Table IV). The theoretical section relies on standard smoothness, convexity, and PAC-Bayes arguments citing external sources (e.g., Descent Lemma [33], PAC-Bayes [35]); although some proofs are heuristic sketches, none assume the result being proved. The conditional-generation section contains internal inconsistencies (Table II reports TPSA MAD of 123.40 while Section VII-C claims average property similarities near 1.0), but that is a correctness or reproducibility problem, not circularity. The custom Overall Score (Eq. 9) min-max normalizes each metric over all models including MetaMolGen, which is a self-referential aggregation rather than an independent test; however, the central generation results and the abstract's validity/diversity claims are not derived from that score, so this does not constitute a circular derivation.

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

The model relies on standard ML assumptions plus several unverified premises; the core empirical claim rests on hyperparameter choices and the transferability of normalization statistics. No new physical or chemical entity is introduced; the property projector and task encoder are neural network modules, not hypothesized objects.

free parameters (4)
  • Inner-loop learning rate alpha = 0.01
    Set by hand in Section VI-C.3; convergence claims depend on alpha <= 2/L without estimating L.
  • Outer-loop Adam learning rate and weight decay = 0.001 and 0.01
    Standard choices in Section VI-C.3; no sensitivity analysis is reported.
  • Sampling temperature tau = not reported
    Decoding diversity in Eq. 7 depends on tau, but the paper never states the value used in experiments.
  • Learnable standardization statistics mu_theta and sigma_theta = learned during training
    Eq. 8; they are fitted to the training data and assumed to transfer across datasets.
assumptions (5)
  • standard math Assumption 1: L-smooth losses
    Used for Theorems 1, 3, and 6; unverified for the actual LSTM and MLP objective.
  • domain assumption Assumption 2: strong convexity of the expected meta-loss
    Unrealistic for deep networks; required for the convergence rate claims in Section V-B.
  • ad hoc to paper Normalized inputs give E[J^T J] approximately c I
    Postulated inside the proof of Theorem 3 without derivation; the condition number reduction depends on it.
  • domain assumption Token-level SMILES training implies syntactic validity
    Eq. 6 claims grammar-compliant token transitions; Table I validity of 75.12% shows the assumption is only partially true.
  • domain assumption Descriptors normalize comparably across datasets
    Training on ChEMBL and testing on QM9, ZINC, and MOSES relies on standardized features transferring across chemical spaces.

how reviews work

0 comments
Cite this review

Pith. "Pith review of MetaMolGen: A Neural Graph Motif Generation Model for De Novo Molecular Design." pith.science (2026). https://pith.science/paper/W6ITXSVO

@misc{pith2026250415587,
  author       = {Pith},
  title        = {Pith review of: MetaMolGen: A Neural Graph Motif Generation Model for De Novo Molecular Design},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/W6ITXSVO}},
  note         = {Machine review of arXiv:2504.15587}
}
read the original abstract

Molecular generation plays an important role in drug discovery and materials science, especially in data-scarce scenarios where traditional generative models often struggle to achieve satisfactory conditional generalization. To address this challenge, we propose MetaMolGen, a first-order meta-learning-based molecular generator designed for few-shot and property-conditioned molecular generation. MetaMolGen standardizes the distribution of graph motifs by mapping them to a normalized latent space, and employs a lightweight autoregressive sequence model to generate SMILES sequences that faithfully reflect the underlying molecular structure. In addition, it supports conditional generation of molecules with target properties through a learnable property projector integrated into the generative process.Experimental results demonstrate that MetaMolGen consistently generates valid and diverse SMILES sequences under low-data regimes, outperforming conventional baselines. This highlights its advantage in fast adaptation and efficient conditional generation for practical molecular design.

Figures

Figures reproduced from arXiv: 2504.15587 by the authors.

Figure 1
Figure 1. Distribution of Key Molecular Properties [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Overview of the MetaMolGen training process. [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Comparison of key generation metrics across differ [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: Few-shot performance comparison of MetaMolGen, RNN, and MolGPT across training set sizes (1k–10k) on six key [PITH_FULL_IMAGE:figures/full_fig_p011_4.png]
Figure 5
Figure 5. Figure 5: Representative molecules generated by the MetaMolGen model using the condition vectors of Aspirin, Tamiflu, [PITH_FULL_IMAGE:figures/full_fig_p012_5.png]
Figure 6
Figure 6. Figure 6: Distributions of generated molecular properties under [PITH_FULL_IMAGE:figures/full_fig_p012_6.png]
Figure 7
Figure 7. Figure 7: Ablation study on the effect of the standardization layer [PITH_FULL_IMAGE:figures/full_fig_p013_7.png]
Figure 8
Figure 8. Figure 8: Training loss trajectories with and without feature standardization. Feature standardization leads to smoother convergence [PITH_FULL_IMAGE:figures/full_fig_p021_8.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

43 extracted references · 35 canonical work pages

  1. [1]

    B. K. Shoichet, ”Virtual screening of chemical libraries,” Nature, vol. 432, no. 7019, pp. 862–865, 2004

    work page 2004

  2. [2]

    Scior, A

    T. Scior, A. Bender, G. Tresadern, et al., ”Recognizing pitfalls in virtual screening: a critical review,” J. Chem. Inf. Model. , vol. 52, no. 4, pp. 867–881, 2012

    work page 2012

  3. [3]

    P. G. Polishchuk, T. I. Madzhidov, and A. Varnek, ”Estimation of the size of drug-like chemical space based on GDB-17 data,” J. Comput.- Aided Mol. Des. , vol. 27, pp. 675–679, 2013

    work page 2013

  4. [4]

    J. L. Reymond, R. Van Deursen, L. C. Blum, et al., ”Chemical space as a source for new drugs,” MedChemComm, vol. 1, no. 1, pp. 30–38, 2010

    work page 2010

  5. [5]

    Zhang, M

    J. Zhang, M. Li, Y . Xu, H. He, Q. Li, and T. Wang, ”StrucGCN: Struc- tural enhanced graph convolutional networks for graph embedding,” Inf. Fusion, vol. 117, p. 102893, 2025

    work page 2025

  6. [6]

    M. H. S. Segler, T. Kogej, C. Tyrchan, et al., ”Generating focused molecule libraries for drug discovery with recurrent neural networks,” ACS Cent. Sci. , vol. 4, no. 1, pp. 120–131, 2018

    work page 2018

  7. [7]

    Bagal, R

    V . Bagal, R. Aggarwal, P. K. Vinod, et al., ”MolGPT: Molecular generation using a transformer-decoder model,” J. Chem. Inf. Model. , vol. 62, no. 9, pp. 2064–2076, 2022

    work page 2022

  8. [8]

    W. Jin, R. Barzilay, and T. Jaakkola, ”Junction tree variational autoen- coder for molecular graph generation,” in Proc. Int. Conf. Mach. Learn. (ICML), PMLR, pp. 2323–2332, 2018

    work page 2018

Show all 43 references
  1. [9]

    S. Choi, S. Seo, B. J. Kim, et al., ”PIDiff: Physics informed diffusion model for protein pocket-specific 3D molecular generation,” Comput. Biol. Med., vol. 180, p. 108865, 2024

    2024

  2. [10]

    H. H. Loeffler, J. He, A. Tibo, et al., ”Reinvent 4: Modern AI–driven generative molecule design,” J. Cheminform., vol. 16, no. 1, p. 20, 2024

    2024

  3. [11]

    Lee and K

    M. Lee and K. Min, ”MGCV AE: multi-objective inverse design via molecular graph conditional variational autoencoder,” J. Chem. Inf. Model., vol. 62, no. 12, pp. 2943–2950, 2022

    2022

  4. [12]

    G. L. Guimaraes, B. Sanchez-Lengeling, P. L. C. Farias, and A. Aspuru-Guzik, ”Objective-Reinforced Generative Adversarial Net- works (ORGAN) for Sequence Generation Models,” arXiv preprint arXiv:1705.10843, 2017

    2017 arXiv

  5. [13]

    De Cao and T

    N. De Cao and T. Kipf, ”MolGAN: An implicit generative model for small molecular graphs,” arXiv preprint arXiv:1805.11973, 2018

    2018 arXiv

  6. [14]

    M. J. Kusner, D. He, H. Larochelle, ”Grammar Variational Autoen- coder,” Proceedings of the 34th International Conference on Machine Learning, vol. 70, pp. 1945-1954, 2017

    1945

  7. [15]

    Nichol, J

    A. Nichol, J. Achiam, and J. Schulman, ”On first-order meta-learning algorithms,” arXiv preprint arXiv:1803.02999 , 2018

    2018 arXiv

  8. [16]

    Garnelo, D

    M. Garnelo, D. Rosenbaum, C. Maddison, et al., ”Conditional neural processes,” in Proc. Int. Conf. Mach. Learn. (ICML) , PMLR, pp. 1704– 1713, 2018

    2018

  9. [17]

    J. You, B. Liu, Z. Ying, et al., ”Graph convolutional policy network for goal-directed molecular graph generation,” in Adv. Neural Inf. Process. Syst. (NeurIPS), vol. 31, 2018

    2018

  10. [18]

    Xie and J

    T. Xie and J. C. Grossman, ”Crystal graph convolutional neural networks for an accurate and interpretable prediction of material properties,” Phys. Rev. Lett., vol. 120, no. 14, p. 145301, 2018

    2018

  11. [19]

    Weininger, ”SMILES, a chemical language and information system

    D. Weininger, ”SMILES, a chemical language and information system

  12. [20]

    Introduction to methodology and encoding rules,” J. Chem. Inf. Comput. Sci., vol. 28, no. 1, pp. 31–36, 1988

    1988

  13. [21]

    J. Lim, S. Ryu, J. W. Kim, et al., ”Molecular generative model based on conditional variational autoencoder for de novo molecular design,” J. Cheminform., vol. 10, pp. 1–9, 2018

    2018

  14. [22]

    W. Jin, R. Barzilay, and T. Jaakkola, ”Hierarchical generation of molec- ular graphs using structural motifs,” in Proc. Int. Conf. Mach. Learn. (ICML), PMLR, pp. 4839–4848, 2020

    2020

  15. [23]

    Zhavoronkov, Y

    A. Zhavoronkov, Y . A. Ivanenkov, A. Aliper, et al., ”Deep learning enables rapid identification of potent DDR1 kinase inhibitors,” Nat. Biotechnol., vol. 37, no. 9, pp. 1038–1040, 2019

    2019

  16. [24]

    C. Finn, P. Abbeel, and S. Levine, ”Model-agnostic meta-learning for fast adaptation of deep networks,” in Proc. 34th Int. Conf. Mach. Learn. (ICML), 2017, pp. 1126–1135

    2017

  17. [25]

    Bottou, F

    L. Bottou, F. E. Curtis, and J. Nocedal, ”Optimization methods for large- scale machine learning,” SIAM Rev., vol. 60, no. 2, pp. 223–311, 2018

    2018

  18. [26]

    T. M. Hospedales, A. Antoniou, P. Micaelli, and A. J. Storkey, ”Meta-learning in neural networks: A survey,” IEEE Trans. Pattern Anal. Mach. Intell. , vol. 44, no. 9, pp. 5149–5169, Sep. 2022, doi: 10.1109/TPAMI.2021.3079209. Available: https://arxiv.org/abs/2004.05439

    2022

  19. [27]

    H. Kim, A. Mnih, J. Schwarz, et al., ”Attentive neural processes,” arXiv preprint arXiv:1901.05761, 2019

    1901 arXiv

  20. [28]

    Gordon, W

    J. Gordon, W. P. Bruinsma, A. Y . K. Foong, et al., ”Convolutional conditional neural processes,” arXiv preprint arXiv:1910.13556, 2019

    1910 arXiv

  21. [29]

    Foong, W

    A. Foong, W. Bruinsma, J. Gordon, et al., ”Meta-learning stationary stochastic process prediction with convolutional neural processes,” in Adv. Neural Inf. Process. Syst. (NeurIPS), vol. 33, pp. 8284–8295, 2020

    2020

  22. [30]

    S. T. Radev, M. Schmitt, L. Schumacher, et al., ”BayesFlow: Amor- tized Bayesian workflows with neural networks,” arXiv preprint arXiv:2306.16015, 2023

    2023 arXiv

  23. [31]

    E. J. Bjerrum, ”SMILES enumeration as data augmentation for neural network modeling of molecules,” arXiv preprint arXiv:1703.07076, 2017

    2017 arXiv

  24. [32]

    Ozturk, E

    H. Ozturk, E. Ozkirimli, and E. Ozkirimli, ”Exploring chemical space using natural language processing methodologies for drug discovery,” Drug Discov. Today, vol. 25, no. 4, pp. 689–705, 2020

    2020

  25. [33]

    Nesterov, Introductory Lectures on Convex Optimization: A Basic Course, Springer, 2013

    Y . Nesterov, Introductory Lectures on Convex Optimization: A Basic Course, Springer, 2013

    2013

  26. [34]

    Beck, First-Order Methods in Optimization , Society for Industrial and Applied Mathematics, 2017

    A. Beck, First-Order Methods in Optimization , Society for Industrial and Applied Mathematics, 2017

    2017

  27. [35]

    Bousquet and A

    O. Bousquet and A. Elisseeff, ”Stability and generalization,” J. Mach. Learn. Res., vol. 2, no. 3, pp. 499–526, 2002

    2002

  28. [36]

    D. A. McAllester, ”Some PAC-Bayesian theorems,” in Proc. 11th Annu. Conf. Comput. Learn. Theory , 1998, pp. 230–234

    1998

  29. [37]

    Gaulton, L

    A. Gaulton, L. J. Bellis, A. P. Bento, et al., ”ChEMBL: a large-scale bioactivity database for drug discovery,” Nucleic Acids Res., vol. 40, no. D1, pp. D1100–D1107, 2012

    2012

  30. [38]

    Ramakrishnan, P

    R. Ramakrishnan, P. O. Dral, M. Rupp, et al., ”Quantum chemistry structures and properties of 134 kilo molecules,” Sci. Data, vol. 1, no. 1, p. 1, 2014

    2014

  31. [39]

    J. J. Irwin and B. K. Shoichet, ”ZINC a free database of commercially available compounds for virtual screening,” J. Chem. Inf. Model. , vol. 45, no. 1, pp. 177–182, 2005

    2005

  32. [40]

    clean leads

    D. Polykovskiy, A. Zhebrak, B. Sanchez-Lengeling, et al., ”Molecular sets (MOSES): a benchmarking platform for molecular generation models,” Front. Pharmacol., vol. 11, p. 565644, 2020. APPENDIX APPENDIX A: B ENCHMARK AND BASELINE DETAILS A. Benchmark Datasets MetaMolGen is ev...

    2020

  33. [41]

    Visualization of the effect of standardization on training dynamics through loss trajectories (Figure 8)

  34. [42]

    Full metric breakdowns across training sizes for few-shot molecular generation (Table IV)

  35. [43]

    Effect of Feature Standardization on Meta-Training Loss

    Detailed results from the ablation study on the effect of the standardization layer (Table V). Effect of Feature Standardization on Meta-Training Loss. Figure 8 shows the training loss trajectories of MetaMolGen with and without feature standardization. Feature standardization...

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.