Linear Embedding Research Articles

Trends and driving forces of agricultural carbon emissions: A case study of Anhui, China.

To facilitate accurate prediction and empirical research on regional agricultural carbon emissions, this paper uses the LLE-PSO-XGBoost carbon emission model, which combines the Local Linear Embedding (LLE), Particle Swarm Algorithm (PSO) and Extreme Gradient Boosting Algorithm (XGBoost), to forecast regional agricultural carbon emissions in Anhui Province under different scenarios. The results show that the regional agricultural carbon emissions in Anhui Province generally show an upward and then downward trend during 2000-2021, and the regional agricultural carbon emissions in Anhui Province in 2030 are expected to fluctuate between 11,342,100 tones and 14,445,700 tones under five different set scenarios. The projections of regional agricultural carbon emissions can play an important role in supporting the development of local regional agriculture, helping to guide the input and policy guidance of local rural low-carbon agriculture and promoting the development of rural areas towards a resource-saving and environment-friendly society.

Quantitative blade damage detection based on multisource domain and multistage joint transfer

For blade damage detection under variable rotational speeds of centrifugal fans, transfer learning under working conditions appears to be more effective than deep learning. However, damage detection models with small samples and single-source domain transfer are still characterized by insufficiencies, such as negative transfer and poor accuracy. Therefore, in this paper, a quantitative blade damage detection method based on multisource domain and multistage joint transfer is proposed. First, an anchor adapter is constructed using linear discriminant analysis projection combined with a feature similarity index metric to acquire the weight matrix of multisource domain-target domain data. Second, a multisource domain feature extractor based on the fusion of vibroacoustic information is established, obtaining the feature set of the target domain data. Then, the feature set is filtered through information gain and max-relevance and min-redundancy to remove the negative transfer features and is combined with an improved supervised locally linear embedding to build a subspace structure preservation model for the alignment between the source and target domain feature distributions. Finally, the classifier is fine-tuned with small sample data for quantitative damage detection of blades with variable rotational speeds. The proposed method is verified using experimental data from two centrifugal fans. The results show that the detection accuracy of the proposed model is significantly higher than that of any comparison model with a single source domain or single stage. Compared with other transfer models, the proposed method is characterized by higher detection accuracy and generalization performance.

Fluid classification with dynamic graph convolution network by local linear embedding well logging data

Fluid prediction is pivotal in exploration, aiding in the identification of targets and estimating reserve potential. To enhance well logging data processing, we employ local linear embedding (LLE) for dimensionality reduction. LLE effectively reduces data dimensionality by identifying local linear relationships and preserving essential local structure in a low-dimensional space, which is particularly advantageous for log data that often contains formation-specific information, including fluid content. The process of dimensionality reduction through LLE retains vital stratigraphic information, which is key for insightful subsequent analyses. Next, we utilize a dynamic graph convolutional network (DGCN) integrated with a multi-scale temporal self-attention (TSA) module for fluid classification on the reduced data. This multi-scale temporal self-attention module is specifically designed to capture time series information inherent in well logging data, allowing the model to autonomously learn and interpret temporal dependencies and evolutionary patterns in the data. This enhances the accuracy of fluid prediction, particularly in the context of varying rock layer characteristics over time. Our methodology, combining LLE with DGCN-TSA, has demonstrated high accuracy in applications such as Tarim Oilfield logging data analysis. It amalgamates advanced technologies with a robust generalization ability. In practical applications, this approach provides steadfast support for oil and gas exploration, significantly contributing to the refinement of fluid prediction accuracy.

P153 Artificial Intelligence to predict interleukin-23 signalling activity from H&E stained whole slide images of Inflammatory Bowel Disease

Abstract Background Interleukin23 (IL-23) is a cytokine that plays a crucial role in the pathogenesis of inflammatory bowel disease (IBD), making it a highly validated therapeutic target. Understanding the role of IL-23 in IBD at the histopathological level is crucial for determining effective treatment strategies, providing insights into IBD patients who fail to respond to targeted therapies, or predicting those who are likely to lose response. In this context, there is a surge in utilizing artificial intelligence (AI) for histopathological data in IBD and other disease indications. Here, we present an automated computer vision approach to predict IL-23 signalling activity directly from routinely stained Hematoxylin and Eosin (H&E) images Methods A total of 1502 samples with matched clinical data and H&E biopsy images were included from 991 Crohn’s disease (CD) and 511 ulcerative colitis (UC) samples. IL-23 signalling activity was calculated using gene set variation analysis on RNA-seq data collected from the same tissue biopsies. The data were obtained from the IBD Plexus program of the Crohn’s & Colitis Foundation. The proposed approach is based on vision transformers (ViTs) which is a type of deep learning model. ViTs divide the input image into fixed-size patches, transform it into linear embedding, and analyze it with the self-attention mechanism. This enables ViTs to incorporate relationships between different patches of the input image to identify regions predictive of IL-23. Our approach was trained in a weakly supervised manner to automatically identify tissue regions that correlate with IL-23 signaling activity. The model produces interpretable heatmaps to interrogate model predictions and allow clinicians to visualize and interpret the significance of different tissue regions predictive of IL-23 Results We performed 5-fold cross-validation on the splits obtained at the patient level, retaining the data distribution of IL-23 signaling activity, biopsy location, and diagnosis. We separately validated the performance of the proposed model on both disease categories, including CD and UC. The proposed approach achieved an area under the curve (AUC) of 0.82 ± 0.04 on unseen data from CD and an AUC of 0.80 ± 0.02 for UC. The 5-fold results for both disease categories are shown below Conclusion The presented results highlight the significance of computational pathology algorithms to identify IL-23 signalling activity from H&E images. Pathological interpretation from the heatmaps may help understand disease pathomechanism and optimize the treatment options for IBD patients by timely identification of IL-23 status. We are further validating the clinical utility of such heatmaps and expanding the use of H&E to predict other patient-centric endpoints

A Telemedicine Analytic Framework for Fully and Semi-automatic Alzheimer's Disease Screening using Clock Drawing Test.

More than 6 million Americans are at risk for Alzheimer's Disease Related Dementias (ADRD), most of whom are 65 or older. The clock drawing test (CDT) is a quick, simple, and effective technique that has the potential advantage of self-management and screening for ADRD patients. Current CDT-based ADRD screening studies focus more on efficacy, involving many handcrafted features, ignoring data modalities, and lacking validation. This paper aims to propose a unified telemedicine framework for fully and semi-automatic effective early ADRD screening based on multimodal and agile data fusion, focusing on the interpretability and validation of the model by using gradient-weighted class activation mapping (Grad-CAM) and locally linear embedding (LLE). The datasets for this work include 1,662 samples of CDT images and related demographic and cognitive information. The fully automatic case involving only CDT images can achieve the highest AUC of 81% with a 75% recall rate in binary screening. The multimodal data fusion in the semi-automatic case can achieve up to 90% AUC with an 83% recall rate. The visualization of the Convolutional Neural Networks (CNN) shows that it can automatically obtain critical information about the outline, scale, and clock hands from CDT images, and the analysis of structured features shows that the memory test is key to effective ADRD screening.

Cooperative spectrum sensing based on locally linear embedding and Adaboost in dynamic fading channel

Cooperative spectrum sensing based on locally linear embedding and Adaboost in dynamic fading channel

Operating Performance Assessment of Complex Nonlinear Industrial Process Based on Kernel Locally Linear Embedding PLS

Operating Performance Assessment of Complex Nonlinear Industrial Process Based on Kernel Locally Linear Embedding PLS

Cooperative spectrum sensing based on locally linear embedding and Adaboost in dynamic fading channel

Cooperative spectrum sensing based on locally linear embedding and Adaboost in dynamic fading channel

Buildings, valuated matroids, and tropical linear spaces

AbstractAffine Bruhat–Tits buildings are geometric spaces extracting the combinatorics of algebraic groups. The building of parameterizes flags of subspaces/lattices in or, equivalently, norms on a fixed finite‐dimensional vector space, up to homothety. It has first been studied by Goldman and Iwahori as a piecewise‐linear analogue of symmetric spaces. The space of seminorms compactifies the space of norms and admits a natural surjective restriction map from the Berkovich analytification of projective space that factors the natural tropicalization map. Inspired by Payne's result that the analytification is the limit of all tropicalizations, we show that the space of seminorms is the limit of all tropicalized linear embeddings and prove a faithful tropicalization result for compactified linear spaces. The space of seminorms is in fact the tropical linear space associated to the universal realizable valuated matroid.

Generative adversarial networks and hessian locally linear embedding for geometric variations management in manufacturing

Generative adversarial networks and hessian locally linear embedding for geometric variations management in manufacturing

Novel parameter-free and parametric same degree distribution-based dimensionality reduction algorithms for trustworthy data structure preserving

As an effective dimensionality reduction method, Same Degree Distribution (SDD) has been demonstrated to be able to maintain better data structure than other dimensionality reduction methods, including Principal Component Analysis (PCA), Multidimensional Scaling (MDS), Isomap, Locally Linear Embedding (LLE), Laplacian Eigen-maps (LE), Uniform Manifold Approximation and Projection (UMAP) and t-Stochastic Neighbour Embedding (t-SNE). In addition, SDD does not require tuning the number of neighbours or perplexity to scale the structure capturing performance. Instead, it requires tuning the degree of degree-distribution ranging in e certain interval. Hence, tuning the degree of degree-distribution makes SDD a less costly method than other methods that require tuning the number of neighbours or perplexity. Although these advantages, SDD is still an expensive method compared with parameter-free methods such as PCA and MDS. A parameter-free SDD is proposed based on standard SDD, with two main differences: 1) it does not require tuning the degree of degree-distribution in the entire range from 1 to 15, but only uses degree 1; and 2) it re-scales the pairwise distances in the range [0, 2] instead of range [0, 1]. A theoretical analysis is presented to prove the better performance of parameter-free SDD. In addition, the performances of the proposed parameter-free SDD and the standard SDD have been experimentally compared in terms of structure capturing and computational time. This paper also proposes a parametric version of SDD using a deep neural network approach to learn the mapping based on the samples of the original data and their corresponding embedded representations in a low dimensional space. Comparative experiments have been undertaken with SDD and other methods such as Isomap, t-SNE and UMAP to demonstrate the effectiveness of the proposed parametric SDD with several popular synthetic and real datasets such as Churn, SEER Breast Cancer, AVletters (LIPS Reading) and MNIST.

Fault detection using generalized autoencoder with neighborhood restriction for electrical drive systems of high-speed trains

Over the past two decades, fault detection of high-speed trains has become an active issue in the transportation area. Recent work has demonstrated the benefits of autoencoder for fault detection in high-speed trains. The latent variable of autoencoder is generally viewed as features generated by the learning process, without explainable interpretation and assessment of the latent variable with respect to its information quality. In the context of representation learning, the latent variable of autoencoder plays an important role towards an fault detection. There is a lack of a methodical framework for explainable applications of AE-based technique to approach fault detection issues. This paper proposes a generalized autoencoder based on neighborhood restriction (NRGAE) to minimize the information loss caused by data compression and reconstruction. The main objective of this paper is to establish a framework for learning autoencoder-based fault detection in the electrical drive systems of high-speed trains. Meanwhile, this work introduces the concepts of mutual information and local linear embedding for electrical drive systems and fault detection specifications. A case study on an electrical drive system is given to illustrate the capability of proposed fault detection.

Comparison of dimensionality reduction techniques for multi-variable spatiotemporal flow fields

In the field of fluid mechanics, it is a potential consensus that nonlinear dimensionality reduction (DR) techniques outperform linear methods. However, this conclusion has been obtained based on simple fluid phenomena and an incomplete evaluation system of dimensionality reduction algorithms. In this study, we use an improved evaluation system of DR methods to compare and evaluate the performance of four DR methods, including two linear techniques: Principal Component Analysis (PCA) and Independent Component Analysis (ICA), and two non-linear techniques: Isometric Mapping (ISOMAP) and Locally Linear Embedding (LLE). The four methods are applied to analyze a complex hydrodynamic flow field with cavitation by considering the joint features of multiple variables. Results show that LLE can capture redundant features that do not contribute to understanding the characteristics of flow fields, while ISOMAP is more suitable for handling datasets with multiple scales than LLE. PCA and ISOMAP can successfully capture the characteristics and evolution of flow fields. In addition, 3D supplementary information can assist ICA in improving the problem of identifying unstable flow field states.

Online evaluation method of resistance spot welding quality based on locally linear embedding algorithm

To address issues of low efficiency, poor feedback timeliness, and unsuitability for fast-paced, high-volume manufacturing of the traditional quality inspection methods of resistance spot welding, an online evaluation method of resistance spot welding quality based on a locally linear embedding algorithm is studied for mild steel resistance spot welding to achieve cost reduction and efficiency improvement. During welding tests, voltage and current were simultaneously collected to calculate the welding power signal. We study the variation pattern of the dynamic power curve. The dynamic power signal was subjected to locally linear embedding and manual feature extraction. The collected features were then used as input to build random forest models and CatBoost models for the online weld quality evaluation, respectively. The results show that the classification models with the feature volumes constructed by locally linear embedding as input have higher assessment accuracy than manually extracted features. The locally linear embedding method can effectively eliminate the subjective influence brought by manual extraction and has better reliability. The CatBoost model based on the locally linear embedding method using the welding power signal can quickly and effectively achieve online quality assessment of mild steel spot welding, providing a further breakthrough in spot welding quality evaluation technology.

Learning Canonical Embeddings for Unsupervised Shape Correspondence With Locally Linear Transformations.

We present a new approach to unsupervised shape correspondence learning between pairs of point clouds. We make the first attempt to adapt the classical locally linear embedding algorithm (LLE)-originally designed for nonlinear dimensionality reduction-for shape correspondence. The key idea is to find dense correspondences between shapes by first obtaining high-dimensional neighborhood-preserving embeddings of low-dimensional point clouds and subsequently aligning the source and target embeddings using locally linear transformations. We demonstrate that learning the embedding using a new LLE-inspired point cloud reconstruction objective results in accurate shape correspondences. More specifically, the approach comprises an end-to-end learnable framework of extracting high-dimensional neighborhood-preserving embeddings, estimating locally linear transformations in the embedding space, and reconstructing shapes via divergence measure-based alignment of probability density functions built over reconstructed and target shapes. Our approach enforces embeddings of shapes in correspondence to lie in the same universal/canonical embedding space, which eventually helps regularize the learning process and leads to a simple nearest neighbors approach between shape embeddings for finding reliable correspondences. Comprehensive experiments show that the new method makes noticeable improvements over state-of-the-art approaches on standard shape correspondence benchmark datasets covering both human and nonhuman shapes.

Hybrid quantum algorithms for flow problems.

For quantum computing (QC) to emerge as a practically indispensable computational tool, there is a need for quantum protocols with end-to-end practical applications-in this instance, fluid dynamics. We debut here a high-performance quantum simulator which we term QFlowS (Quantum Flow Simulator), designed for fluid flow simulations using QC. Solving nonlinear flows by QC generally proceeds by solving an equivalent infinite dimensional linear system as a result of linear embedding. Thus, we first choose to simulate two well-known flows using QFlowS and demonstrate a previously unseen, full gate-level implementation of a hybrid and high precision Quantum Linear Systems Algorithms (QLSA) for simulating such flows at low Reynolds numbers. The utility of this simulator is demonstrated by extracting error estimates and power law scaling that relates [Formula: see text] (a parameter crucial to Hamiltonian simulations) to the condition number [Formula: see text] of the simulation matrix and allows the prediction of an optimal scaling parameter for accurate eigenvalue estimation. Further, we include two speedup preserving algorithms for a) the functional form or sparse quantum state preparation and b) in situ quantum postprocessing tool for computing nonlinear functions of the velocity field. We choose the viscous dissipation rate as an example, for which the end-to-end complexity is shown to be [Formula: see text], where [Formula: see text] is the size of the linear system of equations, [Formula: see text] is the solution error, and [Formula: see text] is the error in postprocessing. This work suggests a path toward quantum simulation of fluid flows and highlights the special considerations needed at the gate-level implementation of QC.

Fault Detection-Based Multiple Local Manifold Learning and Its Application to Blast Furnace Ironmaking Process

Process safety plays a vital role in the modern process industry. To prevent undesired accidents caused by malfunctions or other disturbances in complex industrial processes, considerable attention has been paid to data-driven fault detection techniques. To explore the underlying manifold structure, manifold learning methods including Laplacian eigenmaps, locally linear embedding, and Hessian eigenmaps have been utilized in data-driven fault detection. However, only the partial local structure information is extracted from the aforementioned methods. This paper proposes fused local manifold learning (FLML), which synthesizes the typical manifold learning methods to find the underlying manifold structure from different angles. A more comprehensive local structure is discovered under a unified framework by constructing an objection optimization function for process data dimension reduction. The proposed method takes advantage of different manifold learning methods. Based on the proposed dimension reduction method, a new data-driven fault detection method is developed. Hotelling’s T2 and Q statistics are established for the purpose of fault detection. Experiments on an industrial benchmark Tennessee Eastman process whose average MDR and average FAR of FLML T2 are 7.58% and 0.21% and a real blast furnace ironmaking process whose MDR and FAR of FLML T2 are 2.80% and 0.00% are carried out to demonstrate the superiority and effectiveness of the proposed method.

Application of dimension reduction methods on propeller performance prediction model

Efficient and effective prediction of propeller performance, especially for hydrodynamic performance and fluctuating pressure, have always been areas of interest for ship engineers and researchers. With the rapid blooming of machine learning based surrogate models, it is important to identify proper features that can represent crucial information from numerous features of propellers' geometry design. This paper used five different dimension reduction methods to conduct the feature selection process from propellers geometry. The reduced features are then used as inputs for a random forest based surrogate model to predict both the hydrodynamic performance and fluctuation pressure under cavitation. The experiment results on three test propellers showed that dimension reduction methods such as factor analysis, principal component analysis and random forest feature selection could yield a higher precision for the surrogate model compare to using all features or the main features selected by experience, while manifold learning methods such as isometric feature mapping and locally linear embedding were not good at extracting propellers' geometry. This paper innovatively proposes dimension reduction methods that can automatically extract main features from propellers' complex geometry. With these main features selected, Artificial Intelligence models can yield a higher precision on propellers’ performance predication. Furthermore, these dimension reduction methods have high generalization ability on different types of propellers, which enables a smarter and more cost-efficient way for the preliminary design process of ship propellers.

Neighborhood preserving embedding with autoencoder

Neighborhood preserving embedding (NPE) is a classical method for dimensionality reduction (DR), and it is a linear version of the locally linear embedding method. However, NPE and all its variants only consider the one-way mapping from high-dimensional space to low-dimensional space. The projected low-dimensional data may not accurately and effectively “represent” the original samples. To address this problem, we improve NPE based on linear autoencoder. The conventional projection of NPE is considered as the encoding stage, and the decoder stage is a reconstruction from the low-dimensional space to the original high-dimensional space, which is the key to maintaining more significant information. Based on this, we propose a new NPE method called NPEAE (neighborhood preserving embedding with autoencoder) in this paper. NPEAE performs excellently in face recognition, handwritten character categorization, object classification, etc. The experiments on MNIST, COIL-20, the Extended Yale B, Olivetti Research Laboratory (ORL), and FERET show that NPEAE has a higher recognition accuracy than other comparative methods.

Human Activity Recognition based on Local Linear Embedding and Geodesic Flow Kernel on Grassmann manifolds

Human Activity Recognition (HAR) plays a crucial role in various applications(e.g., medical treatment, video surveillance and sports monitoring). Transfer learning is a promising solution to cross-domain identification problems in HAR. However, existing methods usually ignore the negative transfer caused by using the features of each source domain in equal proportions, as well as the distribution difference between the source and target domains. In this paper, an HAR method based on manifold learning is proposed. Firstly, the similarity between the domain and multiple source domains is calculated using the Multi-Kernel-Maximum Mean Difference (MK-MMD), and the source domain most similar to the target domain is selected as the optimal source domain in the transfer task. Secondly, Locally Linear Embedding (LLE) is leveraged to reduce the dimensionality of both optimal source domain and target domain data to remove redundant information, and the Geodesic Flow Kernel (GFK) is utilized to project low-dimensional data into the Grassmann manifold space and reduce the distribution difference between the two domains. Finally, the source domain action training model is applied to the target domain. Three public datasets (i.e., PAMAP2, OPPORTUNITY and UCI DSADS) are utilized to validate the effectiveness of the proposed approach. Experimental results are presented to demonstrate that the proposed HAR method can predict a large number of unlabeled samples in the target domain while preserving the original data structure.