Double-Shift: A Low-Power DNN Weights Storage and Access Framework based on Approximate Decomposition and Quantization

One major challenge in deploying Deep Neural Network (DNN) in resource-constrained applications, such as edge nodes, mobile embedded systems, and IoT devices, is its high energy cost. The emerging approximate computing methodology can effectively reduce the energy consumption during the computing process in DNN. However, a recent study shows that the weight storage and access operations can dominate DNN's energy consumption due to the fact that the huge size of DNN weights must be stored in the high-energy-cost DRAM. In this paper, we propose Double-Shift, a low-power DNN weight storage and access framework, to solve this problem. Enabled by approximate decomposition and quantization, Double-Shift can reduce the data size of the weights effectively. By designing a novel weight storage allocation strategy, Double-Shift can boost the energy efficiency by trading the energy consuming weight storage and access operations for low-energy-cost computations. Our experimental results show that Double-Shift can reduce DNN weights to 3.96%–6.38% of the original size and achieve an energy saving of 86.47%–93.62%, while introducing a DNN classification error within 2%.

Thermoelectric generation with reduced pollutants made possible by bio-inspired computing

The debate to establish a balance between the generation of electricity and the preservation of the environment is, extraordinarily, important. This article proposes, as a short-term solution, the replacement of diesel oil by natural gas in thermoelectric generation. Natural gas emits 75% less pollutants to the environment than diesel and has a similar energetic efficiency. As a strategy for this replacement to occur safely, the computational modeling was developed in a Bioinspired Computing methodology, called Genetic Algorithm (GA). The GA incorporated all the variables of the electricity and natural gas networks, presented in the mathematical modeling. The result was a significant reduction in the level of pollutants emitted, with high stability in the electrical power system.

The involvement of records managers in cloud computing decisions: A cross-sectional study of New Zealand records managers

<p>Research problem: Cloud computing has become an important topic in many organisations, due to the benefits it can provide to businesses and their operations. This increased interest in cloud computing is also reflected in the records management profession. However, records managers using cloud computing need to be aware of many factors that could negatively affect control of their records, and be able to manage these potential implications. This study aims to discover the level of involvement that records managers have in decision-making relating to cloud computing, and also to determine how informed records managers are about the implications of cloud computing. Methodology: The research design used was a cross-sectional study, with an online web survey being distributed to members of the NZRecords mailing list (an e-mail list for the New Zealand recordkeeping community). Results: The results of this study highlight that records managers have low levels of involvement in cloud computing decision-making, and mostly do not believe that their opinions will influence decisions about cloud computing in their organisations. The findings of the survey reveal awareness of the potential implications of cloud computing is high, although more resources and training should be made available to these records managers, especially in the area of portability and interoperability of records in the cloud. Implications: Requests are made for additional training resources to be made available. Suggestions are made for further research into the factors affecting records managers’ involvement in cloud computing decisions.</p>

The involvement of records managers in cloud computing decisions: A cross-sectional study of New Zealand records managers

<p>Research problem: Cloud computing has become an important topic in many organisations, due to the benefits it can provide to businesses and their operations. This increased interest in cloud computing is also reflected in the records management profession. However, records managers using cloud computing need to be aware of many factors that could negatively affect control of their records, and be able to manage these potential implications. This study aims to discover the level of involvement that records managers have in decision-making relating to cloud computing, and also to determine how informed records managers are about the implications of cloud computing. Methodology: The research design used was a cross-sectional study, with an online web survey being distributed to members of the NZRecords mailing list (an e-mail list for the New Zealand recordkeeping community). Results: The results of this study highlight that records managers have low levels of involvement in cloud computing decision-making, and mostly do not believe that their opinions will influence decisions about cloud computing in their organisations. The findings of the survey reveal awareness of the potential implications of cloud computing is high, although more resources and training should be made available to these records managers, especially in the area of portability and interoperability of records in the cloud. Implications: Requests are made for additional training resources to be made available. Suggestions are made for further research into the factors affecting records managers’ involvement in cloud computing decisions.</p>

Reduced Dilation-Erosion Perceptron for Binary Classification

Dilation and erosion are two elementary operations from mathematical morphology, a non-linear lattice computing methodology widely used for image processing and analysis. The dilation-erosion perceptron (DEP) is a morphological neural network obtained by a convex combination of a dilation and an erosion followed by the application of a hard-limiter function for binary classification tasks. A DEP classifier can be trained using a convex-concave procedure along with the minimization of the hinge loss function. As a lattice computing model, the DEP classifier assumes the feature and class spaces are partially ordered sets. In many practical situations, however, there is no natural ordering for the feature patterns. Using concepts from multi-valued mathematical morphology, this paper introduces the reduced dilation-erosion (r-DEP) classifier. An r-DEP classifier is obtained by endowing the feature space with an appropriate reduced ordering. Such reduced ordering can be determined using two approaches: one based on an ensemble of support vector classifiers (SVCs) with different kernels and the other based on a bagging of similar SVCs trained using different samples of the training set. Using several binary classification datasets from the OpenML repository, the ensemble and bagging r-DEP classifiers yielded mean higher balanced accuracy scores than the linear, polynomial, and radial basis function (RBF) SVCs as well as their ensemble and a bagging of RBF SVCs.

Performance Enhancement of Cloud Computing: Methodology & Tool

This paper describes the testing process employed for testing the in-house developed cloud by using the Google open source tool PerfKit and employing techniques for increasing the performance. Though new tools for testing cloud are emerging into the market, there are aspects which are suited for manual testing and some which can be speeded up using automatic testing tools for various cloud environments for Infrastructure, Platform and Software services. This paper brings out the techniques best suited to test different features of Cloud computing environment and to figure out the lacuna in performance of cloud services. The authors also try to bring out solutions to improve the performance of cloud (recommend) by using various tools to figure out the debugging and analysis process guidelines to follow while fine tuning the performance of private clouds