Efficient Embedding Research Articles

Efficient embedding of service function chains in space-division multiplexing elastic optical networks

Virtual Network Functions (VNFs) are implemented as software instances on generic servers in Data Centers (DCs). A network service is implemented by an ordered list of VNFs, called Service Function Chain (SFC). SFC embedding decides which DC to place/instantiate a VNF, and how to route traffic through the instantiated VNFs/DCs. Benefits from its high capacity and flexibility, the Space-Division Multiplexing Elastic Optical Networks (SDM-EONs) have been proposed as a promising solution to provide connection service between DCs. In this paper, the problem of SFC embedding in SDM-EONs is studied. We first formulate an Integer Linear Programming (ILP) to minimize the Maximum Index of the Utilized Frequency Slot (MIUFS) for a given set of SFC requests. To achieve scalability, a greedy algorithm is proposed to embed SFCs separately. For each SFC, the lowest cost embedding solution can be found by enumeration would be selected for service embedding. A novel stretch algorithm is then designed to reduce the computing complexity further. The key idea is to place VNFs close to the least-cost path between the source and destination of a request. Three strategies to measure the path cost are evaluated. Simulation results show that our stretch algorithm gives the best overall performance in terms of running time, bandwidth consumption, MIUFS and request blocking probability.

Information Embedding for Secure Manufacturing: Challenges and Research Opportunities

Abstract The digitization of manufacturing has transformed the product realization process across many industries, from aerospace and automotive to medicine and healthcare. While this progress has accelerated product development cycles and enabled designers to create products with previously unachievable complexity and precision, it has also opened the door to a broad array of unique security concerns, from theft of intellectual property to supply chain attacks and counterfeiting. To address these concerns, information embedding (e.g., watermarks and fingerprints) has emerged as a promising solution that enhances product security and traceability. Information embedding techniques involve storing unique and secure information within parts, making these parts easier to track and to verify for authenticity. However, a successful information embedding scheme requires information to be transmitted in physical parts both securely and in a way that is accessible to end sers. Ensuring these qualities introduces unique computational and engineering challenges. For instance, these qualities require the designer of the embedding scheme to have an accurate model of the cyber-physical processes needed to embed information during manufacturing and read it later in the product life cycle, as well as models of the phenomena that may degrade that information through natural wear-and-tear, or through adversarial attacks. This article discusses challenges and research opportunities for the engineering design and manufacturing community in developing methods for efficient information embedding in manufactured products.

Efficient reconfigurable architecture to enhance medical image security

Medical images are one of the most critical and sensitive types of data in information systems. For the secure storage and transfer of medical images, confidentiality is the most important aspect. This paper presents efficient embedding technique to enhance medical image security. The Gaussian filters are used as preprocessing to remove high frequency components and then applied to cumulative distribution function (CDF) 5/3 wavelet to obtain LL band features. Similarly, the LL band features of cover image are obtained. The alpha bending technique combines both the LL band features of cover and secret image to form first level of encryption and now with other high frequency bands of LH, HL and HH applied to Inverse CDF 5/3 obtains an encrypted image which is then transferred along with key obtained through other bands of LH, HL and HH. The key generated acts as additional level of security and similarly, at the receiver the opposite operations are performed to obtain the original secrete image. The performance is measured in terms of Peak signal-to-noise ratio (PSNR) and is compared with existing techniques to validate the results. Further, the entire architecture is synthesized on spartan 6 field-programmable gate array (FPGA) board to compare the hardware utilizations.

Graph Neural Network-Based Efficient Subgraph Embedding Method for Link Prediction in Mobile Edge Computing

Link prediction is critical to completing the missing links in a network or to predicting the generation of new links according to current network structure information, which is vital for analyzing the evolution of a network, such as the logical architecture construction of MEC (mobile edge computing) routing links of a 5G/6G access network. Link prediction can provide throughput guidance for MEC and select appropriate c nodes through the MEC routing links of 5G/6G access networks. Traditional link prediction algorithms are always based on node similarity, which needs predefined similarity functions, is highly hypothetical and can only be applied to specific network structures without generality. To solve this problem, this paper proposes a new efficient link prediction algorithm PLAS (predicting links by analysis subgraph) and its GNN (graph neural network) version PLGAT (predicting links by graph attention networks) based on the target node pair subgraph. In order to automatically learn the graph structure characteristics, the algorithm first extracts the h-hop subgraph of the target node pair, and then predicts whether the target node pair will be linked according to the subgraph. Experiments on eleven real datasets show that our proposed link prediction algorithm is suitable for various network structures and is superior to other link prediction algorithms, especially in some 5G MEC Access networks datasets with higher AUC (area under curve) values.

Efficient Medical Knowledge Graph Embedding: Leveraging Adaptive Hierarchical Transformers and Model Compression

Medical knowledge graphs have emerged as essential tools for representing complex relationships among medical entities. However, existing methods for learning embeddings from medical knowledge graphs, such as DistMult, RotatE, ConvE, InteractE, JointE, and ConvKB, may not adequately capture the unique challenges posed by the domain, including the heterogeneity of medical entities, rich hierarchical structures, large-scale, high-dimensionality, and noisy and incomplete data. In this study, we propose an Adaptive Hierarchical Transformer with Memory (AHTM) model, coupled with a teacher–student model compression approach, to effectively address these challenges and learn embeddings from a rich medical knowledge dataset containing diverse entities and relationship sets. We evaluate the AHTM model on this newly constructed “Med-Dis” dataset and demonstrate its superiority over baseline methods. The AHTM model achieves substantial improvements in Mean Rank (MR) and Hits@10 values, with the highest MR value increasing by nearly 56% and Hits@10 increasing by 39%. Furthermore, we observe similar performance enhancements on the “FB15K-237” and “WN18RR” datasets. Our model compression approach, incorporating knowledge distillation and weight quantization, effectively reduces the model’s storage and computational requirements, making it suitable for resource-constrained environments. Overall, the proposed AHTM model and compression techniques offer a novel and effective solution for learning embeddings from medical knowledge graphs and enhancing our understanding of complex relationships among medical entities, while addressing the inadequacies of existing approaches.

Representing Hierarchical Structured Data Using Cone Embedding

Extracting hierarchical structure in graph data is becoming an important problem in fields such as natural language processing and developmental biology. Hierarchical structures can be extracted by embedding methods in non-Euclidean spaces, such as Poincaré embedding and Lorentz embedding, and it is now possible to learn efficient embedding by taking advantage of the structure of these spaces. In this study, we propose embedding into another type of metric space called a metric cone by learning an only one-dimensional coordinate variable added to the original vector space or a pre-trained embedding space. This allows for the extraction of hierarchical information while maintaining the properties of the pre-trained embedding. The metric cone is a one-dimensional extension of the original metric space and has the advantage that the curvature of the space can be easily adjusted by a parameter even when the coordinates of the original space are fixed. Through an extensive empirical evaluation we have corroborated the effectiveness of the proposed cone embedding model. In the case of randomly generated trees, cone embedding demonstrated superior performance in extracting hierarchical structures compared to existing techniques, particularly in high-dimensional settings. For WordNet embeddings, cone embedding exhibited a noteworthy correlation between the extracted hierarchical structures and human evaluation outcomes.

Vacuum-assisted tissue embedding for whole-heart imaging.

The use of combined optical imaging and tissue sectioning has potential for use in visualizing heart-wide fine structures at single-cell resolution. However, existing tissue preparation methods fail to generate ultrathin cavity-containing cardiac tissue slices with minimal deformation. This study developed an efficient vacuum-assisted tissue embedding method to prepare high-filled, agarose-embedded whole-heart tissue. Utilizing optimized vacuum parameters, we achieved 94% filled whole-heart tissue with the thinnest cut slice of 5 µm. We subsequently imaged a whole mouse heart sample using vibratome-integrated fluorescence micro-optical sectioning tomography (fMOST) with a voxel size of 0.32 µm × 0.32 µm × 1 µm. The imaging results indicated that the vacuum-assisted embedding method enabled whole-heart tissue to withstand long-term thin cutting while ensuring that slices were consistent and of high quality.

Improving BERT model for requirements classification by bidirectional LSTM-CNN deep model

In the last decade, requirements classification has emerged as hot research topic in Requirements Engineering (RE). Early identification of software requirements helps the development team in the design of software systems. Manual identification and classification of these requirements is time-consuming and labor-intensive task. To address this issue, machine and deep learning techniques have been studied for automatic classification of requirements. Furthermore, an efficient word embedding representation of the input data is a major concern in automatic approaches. This research work presents a novel requirements classification model that integrates Bidirectional Encoder Representations from Transformers (BERT) with Bidirectional Long-short Term Memory (BiLSTM) with Convolutional Neural Network (CNN) layer called BERT-BiCNN. In BERT-BiCNN, BERT is used as word embedding layer that extract the full combination of contextual semantics. BiLSTM obtains contextual information. The CNN mechanism has been employed to reduce the dimensionality of feature space by selecting the important features. The effectiveness of BERT-BiCNN is evaluated on the PROMISE dataset. On comparative analysis, it can be concluded that proposed approach outperforms six recent deep learning based architectures in binary and multi-class classification.

Real-Time Stereo 3D Car Detection With Shape-Aware Non-Uniform Sampling

Pseudo-LiDAR based stereo 3D detectors have gained popularity due to their high accuracy. However, these methods need dense depth supervision and suffer from inferior speed. To solve these two issues, a recently introduced RTS3D builds an efficient 4D feature-consistency embedding (FCE) space for the object intermediate representation without depth supervision, where FCE space performs uniform sampling to generate feature sampling points, which ignores the importance of different object regions. In this paper, we observe that, compared with the inner region, the outer region of the object plays a more important role for accurate 3D detection. To fully exploit the useful information from the outer region, we propose a novel shape-aware non-uniform sampling strategy. Instead of uniform sampling, our proposed non-uniform sampling strategy performs dense sampling in outer region and sparse sampling in inner region. Therefore, more points are sampled from the outer region and more useful features are extracted for 3D detection. In addition, we design a high-level semantic enhanced FCE module to exploit more contextual information and suppress noise better. As a result, it further improves feature discrimination of each sampling point. Experimental results on the KITTI dataset show the effectiveness of the proposed method. Compared with the baseline RTS3D, our proposed method has 2.57% improvement on car AP <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$_{3d}$ </tex-math></inline-formula> almost without extra network parameters. Moreover, our proposed method outperforms the state-of-the-art methods without extra supervision at a real-time speed.

Reference-free phylogeny from sequencing data

MotivationClustering of genetic sequences is one of the key parts of bioinformatics analyses. Resulting phylogenetic trees are beneficial for solving many research questions, including tracing the history of species, studying migration in the past, or tracing a source of a virus outbreak. At the same time, biologists provide more data in the raw form of reads or only on contig-level assembly. Therefore, tools that are able to process those data without supervision need to be developed.ResultsIn this paper, we present a tool for reference-free phylogeny capable of handling data where no mature-level assembly is available. The tool allows distance calculation for raw reads, contigs, and the combination of the latter. The tool provides an estimation of the Levenshtein distance between the sequences, which in turn estimates the number of mutations between the organisms. Compared to the previous research, the novelty of the method lies in a newly proposed combination of the read and contig measures, a new method for read-contig mapping, and an efficient embedding of contigs.

Improved Object Re-Identification via More Efficient Embeddings

Object reidentification (ReID) in cluttered rigid scenes is a challenging problem especially when same-looking objects coexist in the scene. ReID is accepted to be one of the most powerful tools for matching the correct identities to each individual object when issues such as occlusion, missed detections, multiple same-looking objects coexisting in the same scene, and disappearance of objects from the view and/or revisiting the same region arise. We propose a novel framework towards more efficient object ReID, improved object reidentification (IO-ReID), to perform object ReID in challenging scenes with real-time processing in mind. The proposed approach achieves distinctive and efficient object embedding via training with the triplet loss, with input from both the foreground/background split by bounding box, and the full input image. With extensive experiments on two datasets serving for Object ReID, we demonstrate that the proposed method, IO-ReID, obtains a higher ReID accuracy and runs faster compared to the state-of-the-art methods on object ReID.

Distributed Graph Embedding with Information-Oriented Random Walks

Graph embedding maps graph nodes to low-dimensional vectors, and is widely adopted in machine learning tasks. The increasing availability of billion-edge graphs underscores the importance of learning efficient and effective embeddings on large graphs, such as link prediction on Twitter with over one billion edges. Most existing graph embedding methods fall short of reaching high data scalability. In this paper, we present a general-purpose, distributed, information-centric random walk-based graph embedding framework, DistGER, which can scale to embed billion-edge graphs. DistGER incrementally computes information-centric random walks. It further leverages a multi-proximity-aware, streaming, parallel graph partitioning strategy, simultaneously achieving high local partition quality and excellent workload balancing across machines. DistGER also improves the distributed Skip-Gram learning model to generate node embeddings by optimizing the access locality, CPU throughput, and synchronization efficiency. Experiments on real-world graphs demonstrate that compared to state-of-the-art distributed graph embedding frameworks, including KnightKing, DistDGL, and Pytorch-BigGraph, DistGER exhibits 2.33×--129× acceleration, 45% reduction in cross-machines communication, and &gt;10% effectiveness improvement in downstream tasks.

NORD: NOde Ranking-based efficient virtual network embedding over single Domain substrate networks

Network virtualization (NV) allows the service providers (SPs) to partition the substrate resources in the form of isolated virtual networks (VNs) comprising multiple correlated virtual machines (VMs) and virtual links (VLs), capturing the dependencies. Though NV brought about multiple benefits, such as service isolation, improved quality-of-service (QoS), secure communication, and better utilization of substrate resources, it also introduced numerous research challenges. In this regard, one of the predominant challenges is assigning resources to the virtual components, i.e., VMs and VLs, also termed virtual network embedding (VNE). VNE comprises two closely related sub-problems, (i.) VM embedding and (ii.) VL embedding, and both the problems have been demonstrated to be NP-Hard. In the context of VNE, maximizing the revenue to cost ratio remains the focal point for the SPs as it not only boosts acceptance of VNRs but also effectively utilizes the substrate resources. However, the existing literature on VNE suffers from the following pitfalls: (i.) They only consider system resources or (ii.) limited topological attributes. However, both attributes are quintessential in accurately capturing the VNRs and the substrate network dependencies, thereby augmenting the revenue to cost ratio. This paper proposes an efficient VNE strategy called, NOde Ranking-based efficient virtual network embedding over single Domain substrate networks (NORD), to maximize the revenue to cost ratio. To address the problem of VM embedding, NORD utilizes a hybrid entropy and the technique for order of preference by similarity to ideal solution (TOPSIS) based ranking strategy for VMs and servers considering both system and topological attributes that effectively capture the dependencies. Once the ranking is generated, A greedy VM embedding followed by shortest path VL embedding completes the assignment. Simulation results confirm that NORD attains a 40% and 61% increment in average acceptance and revenue-to-cost ratios compared to the baselines.

Reads2Vec: Efficient Embedding of Raw High-Throughput Sequencing Reads Data.

The massive amount of genomic data appearing for SARS-CoV-2 since the beginning of the COVID-19 pandemic has challenged traditional methods for studying its dynamics. As a result, new methods such as Pangolin, which can scale to the millions of samples of SARS-CoV-2 currently available, have appeared. Such a tool is tailored to take as input assembled, aligned, and curated full-length sequences, such as those found in the GISAID database. As high-throughput sequencing technologies continue to advance, such assembly, alignment, and curation may become a bottleneck, creating a need for methods that can process raw sequencing reads directly. In this article, we propose Reads2Vec, an alignment-free embedding approach that can generate a fixed-length feature vector representation directly from the raw sequencing reads without requiring assembly. Furthermore, since such an embedding is a numerical representation, it may be applied to highly optimized classification and clustering algorithms. Experiments on simulated data show that our proposed embedding obtains better classification results and better clustering properties contrary to existing alignment-free baselines. In a study on real data, we show that alignment-free embeddings have better clustering properties than the Pangolin tool and that the spike region of the SARS-CoV-2 genome heavily informs the alignment-free clusterings, which is consistent with current biological knowledge of SARS-CoV-2.

Efficient Knowledge Graph Embedding Training Framework with Multiple GPUs

When training a large-scale knowledge graph embedding (KGE) model with multiple graphics processing units (GPUs), the partition-based method is necessary for parallel training. However, existing partition-based training methods suffer from low GPU utilization and high input/output (10) overhead between the memory and disk. For a high 10 overhead between the disk and memory problem, we optimized the twice partitioning with fine-grained GPU scheduling to reduce the 10 overhead between the CPU memory and disk. For low GPU utilization caused by the GPU load imbalance problem, we proposed balanced partitioning and dynamic scheduling methods to accelerate the training speed in different cases. With the above methods, we proposed fine-grained partitioning KGE, an efficient KGE training framework with multiple GPUs. We conducted experiments on some benchmarks of the knowledge graph, and the results show that our method achieves speedup compared to existing framework on the training of KGE.

How to Realize Efficient and Scalable Graph Embeddings via an Entropy-Driven Mechanism

Graph embedding is becoming widely adopted as an efficient way to learn graph representations required to solve graph analytics problems. However, most existing graph embedding methods, owing to computation-efficiency challenges for large-scale graphs, generally employ a one-size-fits-all strategy to extract information, resulting in a large amount of redundant or inaccurate representations. In this work, we propose HuGE+, an efficient and scalable graph embedding method enabled by an entropy-driven mechanism. Specifically, HuGE+ leverages hybrid-property heuristic random walk to capture node features, which considers both information content of nodes and the number of common neighbors in each walking step. More importantly, to guarantee information effectiveness of sampling, HuGE+ adopts two heuristic methods to decide the random walk length and the number of walks per node, respectively. Extensive experiments on real-world graphs demonstrate that HuGE+ achieves both efficiency and performance advantages over recent popular graph embedding approaches. For three downstream graph tasks, our approach not only offers >10% average gains, but also exhibits 23×–127× speedup over existing sampling-based methods. In addition, HuGE+ significantly reduces memory footprint by an average of 68.9%, facilitating training for billion-node-scale graph embeddings.

An efficient electrostatic embedding QM/MM method using periodic boundary conditions based on particle-mesh Ewald sums and electrostatic potential fitted charge operators.

Hybrid quantum mechanics/molecular mechanics (QM/MM) models are successful at describing the properties and reactivity of biological macromolecules. Combining ab initio QM/MM methods and periodic boundary conditions (PBC) is currently the optimal approach for modeling chemical processes in an infinite environment, but frequently, these models are too time-consuming for general applicability to biological systems in a solution. Here, we define a simple and efficient electrostatic embedding QM/MM model in PBC, combining the benefits of electrostatic potential fitted atomic charges and particle-mesh Ewald sums, which can efficiently treat systems of an arbitrary size at a reasonable computational cost. To illustrate this, we apply our scheme to extract the lowest singlet excitation energies from a model for Arabidopsis thaliana cryptochrome 1 containing circa 93 000 atoms, accurately reproducing the experimental absorption maximum.

An Efficient Sequential Embedding ConvNet for Rotating Machinery Intelligent Fault Diagnosis

Effective monitoring of critical components of rotating machinery can bring considerable economic benefits and ensure personal safety. However, effective information in mechanical vibration signals is usually submerged by noise interference and non-stationary operating conditions. This causes constructed time-series vibration samples to produce temporal covariate shift (TCS) phenomenon, resulting in differences in the distribution of the training and testing sets. Focusing on this issue, this study proposes an efficient sequential embedding ConvNet (SECN), which can be mainly divided into three layers. Sequential embedding (SE) is firstly proposed to reduce the distribution differences between samples by trainable linear dimensional boosting, where SE aims to realize the secondary representation of samples information distribution via the expansion of a low-dimensional space to a high-dimensional space. Then a single layer of multiscale separable convolution is introduced to extract information over different perceptual fields. Finally the fully connected layer is synthesized by global average pooling to complete the classification task. Experiments and comparisons on three datasets verified that SECN has obvious advantages in terms of accuracy, noise immunity and sample number dependence. The key principles of the proposed SE are explored by ablation and quantitative experiments. The SE applied in the proposed SECN model shows great potential as a basic framework for intelligent fault diagnosis models.

LSTM-DGWO-Based Sentiment Analysis Framework for Analyzing Online Customer Reviews.

Sentiment analysis furnishes consumer concerns regarding products, enabling product enhancement development. Existing sentiment analysis using machine learning techniques is computationally intensive and less reliable. Deep learning in sentiment analysis approaches such as long short term memory has adequately evolved, and the selection of optimal hyperparameters is a significant issue. This study combines the LSTM with differential grey wolf optimization (LSTM-DGWO) deep learning model. The app review dataset is processed using the bidirectional encoder representations from transformers (BERT) framework for efficient word embeddings. Then, review features are extracted by the genetic algorithm (GA), and the optimal review feature set is extracted using the firefly algorithm (FA). Finally, the LSTM-DGWO model categorizes app reviews, and the DGWO algorithm optimizes the hyperparameters of the LSTM model. The proposed model outperformed conventional methods with a greater accuracy of 98.89%. The findings demonstrate that sentiment analysis can be practically applied to understand the customer's perception of enhancing products from a business perspective.

Retracted: BRFP: An Efficient and Universal Sentence Embedding Learning Model Method Supporting Fused Syntax Combined with Graph Embedding Representation Algorithm

Retracted: BRFP: An Efficient and Universal Sentence Embedding Learning Model Method Supporting Fused Syntax Combined with Graph Embedding Representation Algorithm