HybridDTA: Hybrid Data Fusion through Pairwise Training for Drug-Target Affinity Prediction

Estimating drug-target binding affinity (DTA) is crucial for various tasks, including drug design, drug repurposing, and lead optimization. Advanced works adopt machine learning techniques, especially deep learning, to DTA estimation by utilizing the existing assay data. These powerful techniques make it possible to screen a massive amount of potential drugs with limited computation cost. However, a typical DNN-based training paradigm directly minimizes the distances between the estimated scores and the ground truths, suffering from the issue of data inconsistency. The data inconsistency caused by various measurements, e.g., Kd, Ki, and IC50, as well as experimental conditions, e.g., reactant concentration and temperature, severely hinders the effective utilization of existing data, thus deteriorating the performance of DTA prediction. We propose a novel paradigm for effective training on hybrid DTA data to alleviate the data inconsistency issue. Since the ranking orders of the affinity scores with respect to measurements and experimental batches are more consistent, we adopt a pairwise paradigm to enable the DNNs to learn from ranking orders instead. We expect this paradigm can effectively blend datasets with various measurements and experimental batches to achieve better performances. For the sake of verifying the proposed paradigm, we compare it with the previous paradigm for various model backbones on multiple DTA datasets. The experimental results demonstrate the superior performance of our proposed paradigm. The ablation studies also show the effectiveness of the design of the proposed training paradigm.

Using the REA Data Model in Order to Reduce Design Complexity in an Inventory Management Relational Database, and the Need for a Relational Database Complexity Metric.

With the extensive use of relational databases in the business environment there is a need to reduce database complexity in order to avoid data inconsistency and redundancy, which can provide a company with unreliable and/or meaningless data and information. The use of the REA Data Model in database design can significantly help with this problem. The model can eliminate the need for unnecessary data artifacts which should only be generated by the system when needed. This paper also addresses the need for a Relational Database Complexity Metric. A simple and easy to understand metric is presented.

Reducing dairy data inconsistency through Regional Modeling Approach (RMA): a case from North-Western part of Bangladesh

In Bangladesh, the transformation of dairy farming from livelihood-oriented to enterprise-driven farming system might require deeper understanding on the regional differences in terms of regional potential for further dairy development. This, however, entails detailed data on dairy farm at regional level. Since the data are relatively very scarce in one hand and on the other hand, even available, are contradicting among various sources in terms of data accuracy and precision, the application of the regional modeling on the data and extrapolates to the national data and vice-versa is one of the ways to identify the possible options to improve the data availability and quality. Considering this, the current study was undertaken to assess the data inconsistency by comparing the dairy herd structure and its milk production at regional level and propose a validation tool to arrive at the national data by using the regional findings. The International Farm Comparison Network (IFCN) Regional Modeling Approach (RMA) along with the locally developed Integrated Dairy Research Network (IDRN) farm model was used. The primary data was collected from three divisions (9 districts) from the North-Western part of the country. The results revealed that proportion of household farm dominates over family and business farm while considering the total dairy cow as unit for defining the farm type. The share of the cross bred cows to the local cows is 74.6% and 24.4%, respectively. However, the proportion of lactating cows over dry cows and heifer seems to be higher in local cows (48.8%) than cross breed cows (34.2%). The average milk production for all regions is 4.49 lit/day/cow while that for cross breed is 6.23 lit and local 1.71 lit/day/cow. Using regional model and its coefficient on average milk production, herd composition, proportion of lactating cows on total milk production of DLS and IDRN revealed that IDRN new model estimates 36.5% lower milk than the DLS in 2019 and 33.5% lower in 2018. The IDRN version 1.0 and 2.0 model difference was found to 15.4% and 18.3% lower for 2018 and 2019, respectively. The model setup, calibration and validation are time-demanding and challenging tasks for these large set of data, given the scale intensive data requirements, and the need to ensure the reliability data from multiple regions. This study concludes that regional modeling is quite useful for validating the regional share of the milk production and national milk production. However, this study would recommend for using standardized for data collection, validation and thus conducting further study on the other regions and finally including all regions of the country. Bang. J. Anim. Sci. 2020. 49 (2): 128-141

Multiauthority Traceable Ring Signature Scheme for Smart Grid Based on Blockchain

As the next-generation power grid system, the smart grid can realize the balance of supply and demand and help in communication security and privacy protection. However, real-time power consumption data collection might expose the users’ privacy information, such as their living habits and economic conditions. In addition, during the process of data transmission, it may lead to data inconsistency between the user side and the storage side. Blockchain provides tamper-resistant and traceable characteristics for solving these problems, and ring signature schemes provide an anonymous authentication mechanism. Therefore, in this work, we consider the applications of ring signature scheme in smart grid based on blockchain. We introduce the notion of multi-authority traceable ring signature (MA-TRS) scheme for distributed setting. In our scheme, there is an auditing node that can distinguish the identity of the real signer from the ring without any secret information. Last but not least, we prove that the proposed scheme is unforgeable, anonymous, and traceable.

The Improved Least Square Support Vector Machine Based on Wolf Pack Algorithm and Data Inconsistency Rate for Cost Prediction of Substation Projects

Accurate and stable cost forecasting of substation projects is of great significance to ensure the economic construction and sustainable operation of power engineering projects. In this paper, a forecasting model based on the improved least squares support vector machine (ILSSVM) optimized by wolf pack algorithm (WPA) is proposed to improve the accuracy and stability of the cost forecasting of substation projects. Firstly, the optimal features are selected through the data inconsistency rate (DIR), which helps reduce redundant input vectors. Secondly, the wolf pack algorithm is used to optimize the parameters of the improved least square support vector machine. Lastly, the cost forecasting method of WPA-DIR-ILSSVM is established. In this paper, 88 substation projects in different regions from 2015 to 2017 are chosen to conduct the training tests to verify the validity of the model. The results indicate that the new hybrid WPA-DIR-ILSSVM model presents better accuracy, robustness, and generality in cost forecasting of substation projects.