Rethinking Embedded Blocks for Machine Learning Applications

The underlying goal of FPGA architecture research is to devise flexible substrates that implement a wide variety of circuits efficiently. Contemporary FPGA architectures have been optimized to support networking, signal processing, and image processing applications through high-precision digital signal processing (DSP) blocks. The recent emergence of machine learning has created a new set of demands characterized by: (1) higher computational density and (2) low precision arithmetic requirements. With the goal of exploring this new design space in a methodical manner, we first propose a problem formulation involving computing nested loops over multiply-accumulate (MAC) operations, which covers many basic linear algebra primitives and standard deep neural network (DNN) kernels. A quantitative methodology for deriving efficient coarse-grained compute block architectures from benchmarks is then proposed together with a family of new embedded blocks, called MLBlocks. An MLBlock instance includes several multiply-accumulate units connected via a flexible routing, where each configuration performs a few parallel dot-products in a systolic array fashion. This architecture is parameterized with support for different data movements, reuse, and precisions, utilizing a columnar arrangement that is compatible with existing FPGA architectures. On synthetic benchmarks, we demonstrate that for 8-bit arithmetic, MLBlocks offer 6× improved performance over the commercial Xilinx DSP48E2 architecture with smaller area and delay; and for time-multiplexed 16-bit arithmetic, achieves 2× higher performance per area with the same area and frequency. All source codes and data, along with documents to reproduce all the results in this article, are available at http://github.com/raminrasoulinezhad/MLBlocks .

Combining Graph Convolutional Neural Networks and Label Propagation

Label Propagation Algorithm (LPA) and Graph Convolutional Neural Networks (GCN) are both message passing algorithms on graphs. Both solve the task of node classification, but LPA propagates node label information across the edges of the graph, while GCN propagates and transforms node feature information. However, while conceptually similar, theoretical relationship between LPA and GCN has not yet been systematically investigated. Moreover, it is unclear how LPA and GCN can be combined under a unified framework to improve the performance. Here we study the relationship between LPA and GCN in terms of feature/label influence , in which we characterize how much the initial feature/label of one node influences the final feature/label of another node in GCN/LPA. Based on our theoretical analysis, we propose an end-to-end model that combines GCN and LPA. In our unified model, edge weights are learnable, and the LPA serves as regularization to assist the GCN in learning proper edge weights that lead to improved performance. Our model can also be seen as learning the weights of edges based on node labels, which is more direct and efficient than existing feature-based attention models or topology-based diffusion models. In a number of experiments for semi-supervised node classification and knowledge-graph-aware recommendation, our model shows superiority over state-of-the-art baselines.

An improved performance routing protocol based on delay for MANETs in smart cities

Mobile ad-hoc networks (MANETs) is a set of mobile devices that can self-configuration, self-established parameters to transmission in-network. Although limited inability, MANETs have been applied in many domains to serve humanity in recent years, such as disaster recovery, forest fire, military, intelligent traffic, or IoT ecosystems. Because of the movement of network devices, the system performance is low. In order to MANETs could more contribution in the future of the Internet, the routing is a significant problem to enhance the performance of MANETs. In this work, we proposed a new delay-based protocol aim enhance the system performance, called performance routing protocol based on delay (PRPD). In order to analyze the efficiency of the proposed solution, we compared the proposed protocol with traditional protocols. Experiment results showed that the PRPD protocol improved packet delivery ratio, throughput, and delay compared to the traditional protocols.

Optimization of Ideological and Political Education under the Epidemic via Mobile Learning Auxiliary Platform in the Era of Digitization

In the digitized era, life has become simpler with the increased information technology. The Education Department in the whole world is facing a tremendous revolution with the development. The traditional classroom study is converted to a modernized and digitized classroom with visualization. This modernization has increased the learning capability of the students with an increase in student and teacher interaction. From this teaching and learning process, most colleges and universities have improved performance in preparing course materials, effective teaching, and independent learning among the students in the theoretical courses. Ideological and political education (IPE) is a theoretical subject that is taught and understood at higher education institutions, such as colleges and universities. A hybrid hierarchical K -means clustering for optimizing clustering with unsupervised machine learning is proposed to analyze the student performance and concluded that the proposed algorithm shows improved performance than the K -means algorithm.

Deep Learning-Based Estimation of Reverberant Environment for Audio Data Augmentation

This paper proposes an audio data augmentation method based on deep learning in order to improve the performance of dereverberation. Conventionally, audio data are augmented using a room impulse response, which is artificially generated by some methods, such as the image method. The proposed method estimates a reverberation environment model based on a deep neural network that is trained by using clean and recorded audio data as inputs and outputs, respectively. Then, a large amount of a real augmented database is constructed by using the trained reverberation model, and the dereverberation model is trained with the augmented database. The performance of the augmentation model was verified by a log spectral distance and mean square error between the real augmented data and the recorded data. In addition, according to dereverberation experiments, the proposed method showed improved performance compared with the conventional method.

Plasmonic–perovskite solar cells, light emitters, and sensors

The field of plasmonics explores the interaction between light and metallic micro/nanostructures and films. The collective oscillation of free electrons on metallic surfaces enables subwavelength optical confinement and enhanced light–matter interactions. In optoelectronics, perovskite materials are particularly attractive due to their excellent absorption, emission, and carrier transport properties, which lead to the improved performance of solar cells, light-emitting diodes (LEDs), lasers, photodetectors, and sensors. When perovskite materials are coupled with plasmonic structures, the device performance significantly improves owing to strong near-field and far-field optical enhancements, as well as the plasmoelectric effect. Here, we review recent theoretical and experimental works on plasmonic perovskite solar cells, light emitters, and sensors. The underlying physical mechanisms, design routes, device performances, and optimization strategies are summarized. This review also lays out challenges and future directions for the plasmonic perovskite research field toward next-generation optoelectronic technologies.