Hasil untuk "Property"

Chaoji Chen, Yudi Kuang, Shuze Zhu et al.

Kevin Yang, Kyle Swanson, Wengong Jin et al.

Advancements in neural machinery have led to a wide range of algorithmic solutions for molecular property prediction. Two classes of models in particular have yielded promising results: neural networks applied to computed molecular fingerprints or expert-crafted descriptors and graph convolutional neural networks that construct a learned molecular representation by operating on the graph structure of the molecule. However, recent literature has yet to clearly determine which of these two methods is superior when generalizing to new chemical space. Furthermore, prior research has rarely examined these new models in industry research settings in comparison to existing employed models. In this paper, we benchmark models extensively on 19 public and 16 proprietary industrial data sets spanning a wide variety of chemical end points. In addition, we introduce a graph convolutional model that consistently matches or outperforms models using fixed molecular descriptors as well as previous graph neural architectures on both public and proprietary data sets. Our empirical findings indicate that while approaches based on these representations have yet to reach the level of experimental reproducibility, our proposed model nevertheless offers significant improvements over models currently used in industrial workflows.

I. Bell, J. Wronski, S. Quoilin et al.

Over the last few decades, researchers have developed a number of empirical and theoretical models for the correlation and prediction of the thermophysical properties of pure fluids and mixtures treated as pseudo-pure fluids. In this paper, a survey of all the state-of-the-art formulations of thermophysical properties is presented. The most-accurate thermodynamic properties are obtained from multiparameter Helmholtz-energy-explicit-type formulations. For the transport properties, a wider range of methods has been employed, including the extended corresponding states method. All of the thermophysical property correlations described here have been implemented into CoolProp, an open-source thermophysical property library. This library is written in C++, with wrappers available for the majority of programming languages and platforms of technical interest. As of publication, 110 pure and pseudo-pure fluids are included in the library, as well as properties of 40 incompressible fluids and humid air. The source code for the CoolProp library is included as an electronic annex.

M. Schmidt, M. Torn, S. Abiven et al.

Seyone Chithrananda, Gabriel Grand, Bharath Ramsundar

GNNs and chemical fingerprints are the predominant approaches to representing molecules for property prediction. However, in NLP, transformers have become the de-facto standard for representation learning thanks to their strong downstream task transfer. In parallel, the software ecosystem around transformers is maturing rapidly, with libraries like HuggingFace and BertViz enabling streamlined training and introspection. In this work, we make one of the first attempts to systematically evaluate transformers on molecular property prediction tasks via our ChemBERTa model. ChemBERTa scales well with pretraining dataset size, offering competitive downstream performance on MoleculeNet and useful attention-based visualization modalities. Our results suggest that transformers offer a promising avenue of future work for molecular representation learning and property prediction. To facilitate these efforts, we release a curated dataset of 77M SMILES from PubChem suitable for large-scale self-supervised pretraining.

Nadime Francis, Alastair Green, P. Guagliardo et al.

Richard Baraniuk, M. Davenport, R. DeVore et al.

Shorey Peterson, A. A. Berle, G. Means

C. Harris

Zhuanxin Ding, C. Granger, R. Engle

H. D. Soto

O. Hart, John. Moore

S. Gordon

The chief aim of this paper is to examine the economic theory of natural resource utilization as it pertains to the fishing industry. It will appear, I hope, that most of the problems associated with the words “conservation” or “depletion” or “overexploitation” in the fishery are, in reality, manifestations of the fact that the natural resources of the sea yield no economic rent. Fishery resources are unusual in the fact of their common-property nature; but they are not unique, and similar problems are encountered in other cases of common-property resource industries, such as petroleum production, hunting and trapping, etc. Although the theory presented in the following pages is worked out in terms of the fishing industry, it is, I believe, applicable generally to all cases where natural resources are owned in common and exploited under conditions of individualistic competition.

H. Demsetz

When a transaction is concluded in the marketplace, two bundles of property rights are exchanged. A bundle of rights often attaches to a physical commodity or service, but it is the value of the rights that determines the value of what is exchanged. Questions addressed to the emergence and mix of the components of the bundle of rights are prior to those commonly asked by economists. Economists usually take the bundle of property rights as a datum and ask for an explanation of the forces determining the price and the number of units of a good to which these rights attach.

Edella Schlager, E. Ostrom

The term "common-property resource" is an example of a term repeatedly used to refer to property owned by a government or by no one. It is also used for property owned by a community of resource users. Such usage leads to confusion in scientific study and policy analysis. In this paper we develop a conceptual schema for arraying property-rights regimes that distinguishes among diverse bundles of rights ranging from authorized user, to claimant, to proprietor, and to owner. We apply this conceptual schema to analyze findings from a variety of empirical settings including the Maine lobster industry.

A. Agrawal

T. Besley

Matthew B. Dwyer, G. Avrunin, J. Corbett

Model checkers and other finite-state verification tools allow developers to detect certain kinds of errors automatically. Nevertheless, the transition of this technology from research to practice has been slow. While there are a number of potential causes for reluctance to adopt such formal methods, we believe that a primary cause is that practitioners are unfamiliar with specification processes, notations, and strategies. In a recent paper, we proposed a pattern-based approach to the presentation, codification and reuse of property specifications for finite-state verification. Since then, we have carried out a survey of available specifications, collecting over 500 examples of property specifications. We found that most are instances of our proposed patterns. Furthermore, we have updated our pattern system to accommodate new patterns and variations of existing patterns encountered in this survey. This paper reports the results of the survey and the current status of our pattern system.

Esther Heid, Kevin P. Greenman, Yunsie Chung et al.

Deep learning has become a powerful and frequently employed tool for the prediction of molecular properties, thus creating a need for open-source and versatile software solutions that can be operated by nonexperts. Among the current approaches, directed message-passing neural networks (D-MPNNs) have proven to perform well on a variety of property prediction tasks. The software package Chemprop implements the D-MPNN architecture and offers simple, easy, and fast access to machine-learned molecular properties. Compared to its initial version, we present a multitude of new Chemprop functionalities such as the support of multimolecule properties, reactions, atom/bond-level properties, and spectra. Further, we incorporate various uncertainty quantification and calibration methods along with related metrics as well as pretraining and transfer learning workflows, improved hyperparameter optimization, and other customization options concerning loss functions or atom/bond features. We benchmark D-MPNN models trained using Chemprop with the new reaction, atom-level, and spectra functionality on a variety of property prediction data sets, including MoleculeNet and SAMPL, and observe state-of-the-art performance on the prediction of water-octanol partition coefficients, reaction barrier heights, atomic partial charges, and absorption spectra. Chemprop enables out-of-the-box training of D-MPNN models for a variety of problem settings in fast, user-friendly, and open-source software.

Sheng Wang, Yuzhi Guo, Yuhong Wang et al.

With the rapid progress of AI in both academia and industry, Deep Learning has been widely introduced into various areas in drug discovery to accelerate its pace and cut R&D costs. Among all the problems in drug discovery, molecular property prediction has been one of the most important problems. Unlike general Deep Learning applications, the scale of labeled data is limited in molecular property prediction. To better solve this problem, Deep Learning methods have started focusing on how to utilize tremendous unlabeled data to improve the prediction performance on small-scale labeled data. In this paper, we propose a semi-supervised model named SMILES-BERT, which consists of attention mechanism based Transformer Layer. A large-scale unlabeled data has been used to pre-train the model through a Masked SMILES Recovery task. Then the pre-trained model could easily be generalized into different molecular property prediction tasks via fine-tuning. In the experiments, the proposed SMILES-BERT outperforms the state-of-the-art methods on all three datasets, showing the effectiveness of our unsupervised pre-training and great generalization capability of the pre-trained model.