Grammar Variational Autoencoder

Kusner, Matt J.; Paige, Brooks; Hernández-Lobato, José Miguel

doi:10.48550/arxiv.1703.01925

Cited by 51 publications

(92 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We cannot provide a similar discussion for the GG-GAN comparison as the paper does not release the values for the individual metrics. 97.7 2.5 GG-GAN 7 16.6 3G-GAN (ours) 78.5 53.9 62.0 26.2…”

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Permutation Equivariant Generative Adversarial Networks for Graphs

Yoann¹,

Gregorova²,

Kalousis³

2021

Preprint

View full text Add to dashboard Cite

One of the most discussed issues in graph generative modeling is the ordering of the representation. One solution consists of using equivariant generative functions, which ensure the ordering invariance. After having discussed some properties of such functions, we propose 3G-GAN, a 3-stages model relying on GANs and equivariant functions. The model is still under development. However, we present some encouraging exploratory experiments and discuss the issues still to be addressed.

show abstract

Section: Methodsmentioning

confidence: 99%

“…By definition, auto-regressive models are not permutation invariant. Some use a canonical sequential representation such as the SMILES representation (for instance: [20,17,5,7,2]). Others are generating directly some graph objects such nodes or edges [25,18,9,21,10,8,6].…”

Section: Related Workmentioning

confidence: 99%

Permutation Equivariant Generative Adversarial Networks for Graphs

Yoann¹,

Gregorova²,

Kalousis³

2021

Preprint

View full text Add to dashboard Cite

show abstract

“…Jin et al, [18] used a model called Junction Tree VAE (JT-VAE), which first generated a junction tree that serves as a scaffold for the molecule and then specifies the subgraphs as nodes of the junction tree. Their method of coarse-to-fine generation of molecules significantly outperformed previous SMILES-based approaches such as Gram-marVAE [16], Character VAE [22], or graph generating GraphVAE in validity of generated molecules. Another model , DeLinker, which is a gated graph neural network (GGNN) that links two molecular fragments together to form more complex compounds [23] also incorporates 3D structural information to constrain the linking process.…”

Section: Small Molecule Designmentioning

confidence: 99%

“…One possible approach for correction is sequence-to-sequence learning with attention, where the input is a wrong sequence and the target is the corrected one [15]. Grammar VAE (GVAE) [16] was also used to generate SMILES that follow syntactic constraints given by a context-free grammar. Finally, syntaxdirected VAE (SDVAE) [17]) can make use of attribute grammar to enforce syntactic and semantic constraints on generated SMILES.…”

Section: Small Molecule Designmentioning

confidence: 99%

Deep Denerative Models for Drug Design and Response

Zadorozhny,

Nuzhna

2021

Preprint

View full text Add to dashboard Cite

Designing new chemical compounds with desired pharmaceutical properties is a challenging task and takes years of development and testing. Still, a majority of new drugs fail to prove efficient in vivo. Recent success of deep generative modeling holds promises of generation and optimization of new molecules. In this review paper, we provide an overview of the current generative models, and describe necessary biological and chemical terminology, including molecular representations needed to understand the field of drug design and drug response. We present commonly used chemical and biological databases, and tools for generative modeling. Finally, we summarize the current state of generative modeling for drug design and drug response prediction, highlighting the state-of-art approaches and limitations the field is currently facing.

show abstract

“…Modern deep neural network (DNN) models have been used in various molecular applications, such as high-throughput screening for drug discovery, [1][2][3][4] de novo molecular design [5][6][7][8][9][10][11][12] and planning chemical reaction. [13][14][15] DNNs show comparable or sometimes better performance than traditional approaches grounded on quantum chemical theories in predicting some molecular properties, [16][17][18][19][20] if a vast amount of well-qualified data is secured.…”

Section: Introductionmentioning

confidence: 99%

Uncertainty quantification of molecular property prediction with Bayesian neural networks

Ryu,

Kwon,

Kim

2019

Preprint

View full text Add to dashboard Cite

Deep neural networks have outperformed existing machine learning models in various molecular applications. In practical applications, it is still difficult to make confident decisions because of the uncertainty in predictions arisen from insufficient quality and quantity of training data. Here, we show that Bayesian neural networks are useful to quantify the uncertainty of molecular property prediction with three numerical experiments. In particular, it enables us to decompose the predictive variance into the model-and data-driven uncertainties, which helps to elucidate the source of errors. In the logP predictions, we show that data noise affected the data-driven uncertainties more significantly than the model-driven ones. Based on this analysis, we were able to find unexpected errors in the Harvard Clean Energy Project dataset. Lastly, we show that the confidence of prediction is closely related to the predictive uncertainty by performing on bio-activity and toxicity classification problems.

show abstract

Grammar Variational Autoencoder

Cited by 51 publications

References 0 publications

Permutation Equivariant Generative Adversarial Networks for Graphs

Permutation Equivariant Generative Adversarial Networks for Graphs

Deep Denerative Models for Drug Design and Response

Uncertainty quantification of molecular property prediction with Bayesian neural networks

Contact Info

Product

Resources

About