Abstract

Solubility of drug molecules is related to pharmacokinetic properties such as absorption and distribution, which affects the amount of drug that is available in the body for its action. Computational or experimental evaluation of solvation free energies of drug-like molecules/solute that quantify solubilities is an arduous task and hence development of reliable computationally tractable models is sought after in drug discovery tasks in pharmaceutical industry. Here, we report a novel method based on graph neural network to predict solvation free energies. Previous studies considered only the solute for solvation free energy prediction and ignored the nature of the solvent, limiting their practical applicability. The proposed model is an end-to-end framework comprising three phases namely, message passing, interaction and prediction phases. In the first phase, message passing neural network was used to compute inter-atomic interaction within both solute and solvent molecules represented as molecular graphs. In the interaction phase, features from the preceding step is used to calculate a solute-solvent interaction map, since the solvation free energy depends on how (un)favorable the solute and solvent molecules interact with each other. The calculated interaction map that captures the solute-solvent interactions along with the features from the message passing phase is used to predict the solvation free energies in the final phase. The model predicts solvation free energies involving a large number of solvents with high accuracy. We also show that the interaction map captures the electronic and steric factors that govern the solubility of drug-like molecules and hence is chemically interpretable.

Abstract

Recent years have witnessed utilization of modern machine learning approaches for predicting properties of material using available datasets. However, to identify potential candidates for material discovery, one has to systematically scan through a large chemical space and subsequently calculate the properties of all such samples. On the other hand, generative methods are capable of efficiently sampling the chemical space and can generate molecules/materials with desired properties. In this study, we report a deep learning based inorganic material generator (DING) framework consisting of a generator module and a predictor module. The generator module is developed based upon conditional variational autoencoders (CVAE) and the predictor module consists of three deep neural networks trained for predicting enthalpy of formation, volume per atom and energy per atom chosen to demonstrate the proposed method. The predictor and generator modules have been developed using a one hot key representation of the material composition. A series of tests were done to examine the robustness of the predictor models, to demonstrate the continuity of the latent material space, and its ability to generate materials exhibiting target property values. The DING architecture proposed in this paper can be extended to other properties based on which the chemical space can be efficiently explored for interesting materials/molecules.

Abstract

We examine the mutual exclusion problem of concurrency through the systematic application of modern feedback control theory, by revisiting the classical problem involving mutual exclusion: the Generalised Dining Philosophers problem. The result is a modular development of the solution using the notions of system and system composition in a formal setting that employs simple equational reasoning. The modular approach separates the solution architecture from the algorithmic minutiae and has the benefit of simplifying the design and correctness proofs. Two variants of the problem are considered: centralised and distributed topology with N philosophers. In each case, solving the Generalised Dining Philosophers reduces to designing an appropriate feedback controller.

Abstract

Unlike CPUs and GPUs, it is possible to use custom fixed-point data types, specified as a tuple (α, β), on FPGAs. The parameters α and β denote the number of integral and fractional bitwidths respectively. The power and area savings while performing arithmetic operations on fixed-point data types are well known to be significant over using floating-point data types. In this paper, we propose a hybrid approach involving interval analysis and SMT solvers, for estimating integral bitwidths at different compute stages, in an image processing pipeline, specified using a domain-specific language (DSL) such as PolyMage. The DSL specification facilitates the compiler analysis to infer the underlying computational structure with ease. We also propose a simple and practical profile-driven greedy heuristic search technique for fractional bitwidth analysis. Using the Horn-Schunck Optical Flow benchmark program, we demonstrate where the conventional range analysis approaches fail, and how we overcome them using the hybrid technique proposed in this paper. The integral bitwidth estimates provided by the hybrid technique on the optical flow benchmark are up to 3x times better when compared with interval analysis.

Abstract

Finding the centrality measures of nodes in a graph is a problem of fundamental importance due to various applications from social networks, biological networks, and transportation networks. Given the large size of such graphs, it is natural to use parallelism as a recourse. There have been several studies that show how to compute the various centrality measures of nodes in a graph on parallel architectures, including multi-core systems and GPUs. However, as these graphs evolve and change, it is pertinent to study how to update the centrality measures on changes to the underlying graph. In this paper, we show novel parallel algorithms for updating the betweenness- and closeness-centrality values of nodes in a dynamic graph. Our algorithms process a batch of updates in parallel by extending the approach of handling a single update for etweennessand closeness-centrality by Jamour et al. [16] and Sarıyüce et al. [27], respectively. Besides, our algorithms incorporate mechanisms to exploit the structural properties of graphs for enhanced performance.

Abstract

Proteresis, a complementary effect to well-known hysteresis, is an interesting phenomenon to be explored. As we know, the hysteric device is a bistable switch whose states depend on the system's history whereas the proteretic device states' depend on the future. Though an all-optical implementation of the hysteric device has been demonstrated but its complementary behavior has yet to be contrived. Therefore, this work presents the design and development of an all-optical proteretic system using semiconductor ring lasers. The design was achieved by providing a constraint feedforward path to a hysteric feedback system. Additionally, the role of various parameters was studied in order to achieve proteresis. It is observed that the proteretic effect can be achieved only for specific conditions otherwise the system behaves as a traditional hysteric device. The experimental results match well with the theoretical Simulink results.

Abstract

Detecting presence of proteins in a given sample is an essential part of diagnostics. The use of the avidin–biotin interaction is becoming an increasingly common method for the detection of proteins. Among avidin group, Streptavidin forms strongest non-covalent biological interaction with biotin. Therefore, it is stable in wide range of temperatures and exhibits similar binding as antigen–antibody binding. The strong interaction has led to a large number of diagnostic applications using Streptavidin–Biotin technology. Permittivity of Biotin–Streptavidin combination shows a frequency dependency in several dispersion regions like regions. Traditional and standard biological testing use immunoassay technology (ELISA) method use chemicals, and molecules which modifies or destroy properties of sensing bio molecule are cumbersome (several reaction steps) and time consuming. We present design and development …

Abstract

Pre-detection of hypertension mostly considers the measurement of Brachial Artery Blood Pressure (BABP). Although being a standard vital, it is still considered a poor alternative for Central Blood Pressure (CBP). However, CBP is measured invasively during the process of cardiac catheterization (Cath). Though cuff-less techniques to estimate BABP are widely employed, CBP estimation has not been explored yet. Moreover, to best of our knowledge intermittent CBP estimation has not been proposed earlier. Therefore, we present a cuff-less and beat-by-beat CBP estimation technique using linear regression analysis on features extracted from continuous Electrocardiogram (ECG) and Photoplethysmograph (PPG) signals. Unlike for BABP estimation, 30 supplementary features to conventional pulse transit time such as ST-interval, P-systolic peak interval, etc., were extracted to enhance CBP accuracy. This extraction was done using Haar wavelet along with modulus maxima. Feature selection has been done using the wrapper technique and reduced using principal component analysis. Segregation of each beat was achieved with the help of constraints developed based on iteration and backtracing. This model estimates Systolic CBP with a validation error of 0.109±2.37 mmHg and Diastolic CBP with an error of 0.031±2.102 mmHg for 33 Cath lab patients.

Abstract

Since the eminent work of J. S. Bell, spatial correlations in quantum mechanics have gained much progress. Nevertheless, up to now, there is no agreement on the nature of temporal correlations. In this paper, based on the entangled-history theory, we prove that temporal correlations are just quantum channels. Moreover, by using the state tomography technology, the quantum channel can be uniquely determined. What is more, through the entanglement of formation, temporal correlations can be quantified. Cases when the temporal correlation is strongest and weakest are investigated, too.

Abstract

In this work, we extensively study the problem of broadcasting of entanglement as state dependent versus state independent cloners. We start by re-conceptualizing the idea of state dependent quantum cloning machine (SD-QCM), and in that process, we introduce different types of SD-QCMs, namely, orthogonal and non-orthogonal cloners. We derive the conditions for which the fidelity of these cloners will become independent of the input state. We note that such a construction allows us to maximize the cloning fidelity at the cost of having partial information of the input state. In the discussion on broadcasting of entanglement, we start with a general two qubit state as our resource and later we consider a specific example of Bell diagonal state. We apply both state dependent and state independent cloners (orthogonal and non-orthogonal), locally and non locally, on input resource state and obtain a range for broadcasting of entanglement in terms of the input state parameters. Our results highlight several instances where the state dependent cloners outperform their state independent counterparts in broadcasting entanglement. Our study provides a comparative perspective on the broadcasting of entanglement via cloning in two qubit scenario, when we have some knowledge of the resource ensemble versus a situation when we have no such information.