Machine Learning with Enormous "Synthetic" Data Sets: Predicting Glass Transition Temperature of Polyimides Using Graph Convolutional Neural Networks.

In the present work, we address the problem of utilizing machine learning (ML) methods to predict the thermal properties of polymers by establishing "structure-property" relationships. Having focused on a particular class of heterocyclic polymers, namely polyimides (PIs), we developed a graph convolutional neural network (GCNN), being one of the most promising tools for working with big data, to predict the PI glass transition temperature T g as an example of the fundamental property of polymers. To train the GCNN, we propose an original methodology based on using a "transfer learning" approach with an enormous "synthetic" data set for pretraining and a small experimental data set for its fine-tuning. The "synthetic" data set contains more than 6 million combinatorically generated repeating units of PIs and theoretical values of their T g values calculated using the well-established Askadskii's quantitative structure-property relationship (QSPR) computational scheme. Additionally, an experimental data set for 214 PIs was also collected from the literature for training, fine-tuning, and validation of the GCNN. Both "synthetic" and experimental data sets are included into a PolyAskInG database (Polymer Askadskii's Intelligent Gateway). By using the PolyAskInG database, we developed GCNN which allows estimation of T g of PI with a mean absolute error (MAE) of about 20 K, which is 1.5 times lower than in the case of Askadskii QSPR analysis (33 K). To prove the efficiency and usability of the proposed GCNN architecture and training methodology for predicting polymer properties, we also employed "transfer learning" to develop alternative GCNN pretrained on proxy-characteristics taken from the popular quantum-chemical QM9 database for small compounds and fine-tuned on an experimental T g values data set from PolyAskInG database. The obtained results indicate that pretraining of GCNN on the "synthetic" polymer data set provides MAE which is almost twice as low as that in the case of using the QM9 data set in the pretraining stage (∼41 K). Furthermore, we address the questions associated with the influence of the differences in the size of the experimental and "synthetic" data sets (so-called "reality gap" problem), as well as their chemical composition on the training quality. Our results state the overall priority of using polymer data sets for developing deep neural networks, and GCNN in particular, for efficient prediction of polymer properties. Moreover, our work opens up a challenge for the theoretically supported generation of large "synthetic" data sets of polymer properties for the training of the complex ML models. The proposed methodology is rather versatile and may be generalized for predicting other properties of different polymers and copolymers synthesized through the polycondensation reaction.

Computer Simulations of Deep Eutectic Solvents: Challenges, Solutions, and Perspectives.

Deep eutectic solvents (DESs) are one of the most rapidly evolving types of solvents, appearing in a broad range of applications, such as nanotechnology, electrochemistry, biomass transformation, pharmaceuticals, membrane technology, biocomposite development, modern 3D-printing, and many others. The range of their applicability continues to expand, which demands the development of new DESs with improved properties. To do so requires an understanding of the fundamental relationship between the structure and properties of DESs. Computer simulation and machine learning techniques provide a fruitful approach as they can predict and reveal physical mechanisms and readily be linked to experiments. This review is devoted to the computational research of DESs and describes technical features of DES simulations and the corresponding perspectives on various DES applications. The aim is to demonstrate the current frontiers of computational research of DESs and discuss future perspectives.

Validation of the Brief International Cognitive Assessment for Multiple Sclerosis (BICAMS) in the Russian Population.

OBJECTIVE: Cognitive dysfunction is common in multiple sclerosis (MS). The Brief International Cognitive Assessment for MS (BICAMS) battery of tests has been suggested as a measure for the evaluation of the cognitive status of MS patients. This study aims to validate the BICAMS battery in the Russian population of MS patients. METHODS: Age- and sex-matched MS patients (n = 98) and healthy individuals (n = 86) were included in the study. Symbol Digit Modalities Test (SDMT), California Verbal Learning Test, 2nd edition (CVLT-II) and the Brief Visuospatial Memory Test - Revised (BVMT-R) were administered to all participants. The battery was readministered 1 month later to 44 MS patients to investigate the test-retest reliability. RESULTS: MS patients exhibited a significantly lower performance in testing with BICAMS than the control group in all three neuropsychological tests. Test-retest reliability was good for SDMT and CVLT-II (r = .82 and r = .85, respectively) and adequate for BVMT-R (r = .70). Based on the proposed criterion for impairment as z score below 1.5 SD the mean of the control group, we found that 34/98 (35%) of MS patients were found impaired at least in one cognitive domain. Patients with Expanded Disability Status Scale score ≥3.5 performed significantly worse than controls (SDMT, p < .0001; CVLT-II, p = .03; BVMT-R, p = .0004), while those with ≤3.0 scores did not. CONCLUSION: This study demonstrates that the BICAMS battery is a valid instrument to identify cognitive impairment in MS patients and it can be recommended for routine use in the Russian Federation.

Transport Properties of Thermoplastic R-BAPB Polyimide: Molecular Dynamics Simulations and Experiment.

The present work evaluates the transport properties of thermoplastic R-BAPB polyimide based on 1,3-bis(3,3',4,4'-dicarboxyphenoxy)benzene (dianhydride R) and 4,4'-bis(4-aminophenoxy)biphenyl (diamine BAPB). Both experimental studies and molecular dynamics simulations were applied to estimate the diffusion coefficients and solubilities of various gases, such as helium (He), oxygen (O2), nitrogen (N2), and methane (CH4). The validity of the results obtained was confirmed by studying the correlation of the experimental solubilities and diffusion coefficients of He, O2, and N2 in R-BAPB, with their critical temperatures and the effective sizes of the gas molecules, respectively. The solubilities obtained in the molecular dynamics simulations are in good quantitative agreement with the experimental data. A good qualitative relationship between the simulation results and the experimental data is also observed when comparing the diffusion coefficients of the gases. Analysis of the Robeson plots shows that R-BAPB has high selectivity for He, N2, and CO2 separation from CH4, which makes it a promising polymer for developing gas-separation membranes. From this point of view, the simulation models developed and validated in the present work may be put to effective use for further investigations into the transport properties of R-BAPB polyimide and nanocomposites based on it.

Cushing's Syndrome: A Historic Review of the Treatment Strategies and Corresponding Outcomes in a Single Tertiary Center over the Past Half-Century.

Cushing's syndrome (CS) is associated with serious comorbidities and an increased mortality rate that could be reduced only if strict biochemical control is achieved. The aim of this study was to show the 50-year experience of a single tertiary center in the management of CS patients - the different treatment modalities used over the years and the corresponding outcomes. It was a retrospective study of a large cohort of patients from the Bulgarian CS database: 613 patients (374 with ACTH-dependent and 239 with ACTH-independent CS). Pituitary surgery was applied to 242 patients with Cushing's disease (CD) with initial remission rate of 74% of which 10% relapsed. Approximately 36% manifested with active disease during the long-term follow-up (26% with persistent disease, 10% relapses) most of which were subjected to a secondary treatment (13.6% to pituitary resurgery, 14% to pituitary radiotherapy, and 5.4% to bilateral adrenalectomy). A total of 294 CD patients received medical therapy with overall remission rates for the most commonly used drugs: dopamine agonists 20%, pasireotide 30%, and ketoconazole 63%. Significant improvement of results was achieved by combining drugs with different mechanisms of action. Regardless of the progress in the neurosurgery and radiotherapy techniques and new drugs discovery, the management of patients with CS remains a real challenge for physicians. Not only patients with adrenal carcinoma but also significant percentage of subjects with persistent and recurrent Cushing's disease often require a polymodal approach and the efforts of a multidisciplinary highly qualified, experienced, and motivated team in order to achieve a long-term remission.

Method of estimation of turbulence characteristic scales.

We propose an optical method that uses phase data of a laser beam obtained from a Shack-Hartmann sensor to estimate both the inner and outer scales of turbulence. The method is based on the sequential analysis of normalized correlation functions of Zernike coefficients. It allows the exclusion C(n)(2) from the analysis and reduces the solution of a two-parameter problem to a sequential solution of two single-parameter problems. The method has been applied to estimate the outer and inner scales of turbulence induced in the water cell.