Search

With increasingly complex neural network architectures and heterogeneous device characteristics, finding a reasonable graph partitioning and device placement strategy is challenging. There have been prior attempts at learned approaches for solving device placement, these approaches are computationally expensive, unable to handle large graphs consisting over 50000 nodes, and do not generalize well to unseen graphs. To address all these limitations, we propose an efficient single-shot, generalized deep RL method (SGDP) based on a scalable sequential attention mechanism over a graph neural network that is transferable to new graphs. On a diverse set of representative deep learning models, our method on average achieves 20% improvement over human placement and 18% improvement over the prior art with 15× faster convergence. We are the first to demonstrate super human performance on 8-layer recurrent neural network language model and 8-layer GNMT consisting of over 50000 nodes, on 8-GPUs. We provide rationales and sensitivity study on model architecture selections. [ABSTRACT FROM AUTHOR]

Copyright of Fibres & Textiles in Eastern Europe is the property of Sciendo and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

This article investigates an optimal design for a paper-feeding system that minimizes both jamming and simultaneous feeding of multiple papers, hereafter referred to as the multi-feeding rate. A total of 11 design parameters for the paper transfer device, the paper separation device, and the paper guide path are selected and analysed in this study. A test jig for feeding and transferring papers is manufactured to obtain experimental data for use in parameter analysis and design optimization. Back-propagation neural network-based causality analysis is employed to extract five dominant variables among 11 design parameters, and the results of causality analysis are compared with sensitivity results obtained from the analysis of means in the context of experimental design. Five-variable, second-order polynomial based approximate meta-models for jam rate and multi-feeding rate are then constructed, and numerical optimization is performed using NSGA-II, a non-dominated sorting genetic algorithm. Finally, two numerical Pareto optimal solutions are verified via experimental testing. [ABSTRACT FROM AUTHOR]

Papermaking is an industrial process that is becoming more competitive nowadays. In this process there are numerous techniques and measurements to indicate paper quality. To increase competitiveness a good control of paper quality is needed through paper properties predictions from different process measurements. However, complex physico-chemical processes take place during papermaking, and thus, paper property predictions are not easy to obtain, especially in the wet-end area. In the wet end flocculation takes place, which will determine the floc properties during the formation of the sheet, and therefore, it will influence retention, drainage and formation. These strongly affect the runnability of the machine and the properties of the final product, and thus, using wet-end measurements for the predictions implies advanced data treatment. Artificial neural networks have been used in this article to predict newsprint paper properties from wet-end parameters. Results show that formation and strength properties can be robustly predicted from pulp properties at the headbox, flocculation parameters and machine speed. [ABSTRACT FROM AUTHOR]

In this paper, we develop a method for recognizing face images by combining wavelet decomposition, Fisherface method, and fuzzy integral. The proposed approach is comprised of four main stages. The first stage uses the wavelet decomposition that helps extract Intrinsic features of face images. As a result of this decomposition, we obtain four subimages (namely approximation, horizontal, vertical, and diagonal detailed images). The second stage of the approach concerns the application of the Fisherface method to these four decompositions. The choice of the Fisherface method in this setting is motivated by its insensitivity to large variation in light direction, face pose, and facial expression. The two last phases are concerned with the aggregation of the individual classifiers by means of the fuzzy integral. Both Sugeno and Choquet type of fuzzy integral are considered as the aggregalion method. In the experiments we use n-fold cross-validation to assure high consistency of the produced classification outcomes. The experimental results obtained for the Chungbuk National University (CNU) and Yale University face databases reveal that the approach presented in this paper yields better classification performance in comparison to the results obtained by other classifiers. [ABSTRACT FROM AUTHOR]