Manifold Projection Research Articles

Stability and fold bifurcation in a three-species model with mutual interference on parasitoids and hyperparasitoids

This study establishes a novel three-species host–parasitoid–hyperparasitoid system with mutual interference effects on parasitoids and hyperparasitoids. We derive the asymptotic stability conditions for the equilibrium points and classify the impacts of hyperparasitoids on the dynamical behavior by comparing these stability conditions. Additionally, we demonstrate the presence of a fold bifurcation in the proposed three-dimensional model using a novel center manifold projection method. Finally, we verify our findings through numerical simulations.

IoT-driven load forecasting with machine learning for logistics planning

IoT-driven load forecasting with machine learning for logistics planning

(Invited) Theory-Informed AI/ML for Battery Characterization

The use of advanced, including in situ and operando, microscopy and spectroscopy techniques has enabled detailed characterization of battery materials during synthesis and cycling. Data-driven approaches, especially guided by computational spectroscopy, enable robust and accelerated information extraction from such microscopy and spectroscopy data. In this talk, we discuss the featurization of x-ray absorption near edge structure (XANES) for extraction of local structure and electronic properties in NMC battery cathode materials [1], and the extension thereof towards the detection of antisite defect and oxygen vacancy from multimodal EELS/XANES. We also discuss the use of data-driven approaches to accelerate the acquisition [2] and analysis [3] of XANES mapping data. Finally, we discuss the use of AI to extract relevant labeled microscopy and spectroscopy data from scientific literature [4], and recent extensions involving the use of large language models and multi-model contrastive learning.[1] Y. Chen, C. Chen, I. Hwang, M. J. Davis, W. Yang, C.J. Sun, G. Lee, D. McReynolds, D. Allan, J. M. Arias, S. P. Ong, and M. K. Y. Chan, “Robust Machine Learning Inference from X-ray Absorption Near Edge Spectra through Featurization,” Chemistry of Materials, 36, 5, 2304–2313 (2024).[2] S. Tetef, A. Pattammattel, Y. S. Chu, M. K. Y. Chan#, G. T. Seidler, “Accelerating nano-XANES imaging via feature selection,” Digital Discovery 3(1), 201-209 (2024).[3] S. Tetef, A. Pattammattel, Y. S. Chu, M. K. Y. Chan#, G. T. Seidler, “Manifold Projection Image Segmentation for Nano-XANES Imaging,” APL Machine Learning 1, 046119 (2023).[4] E. Schwenker, W. Jiang, T. Spreadbury, N. Ferrier, O. Cossairt, M. K. Y. Chan, “EXSCLAIM! -- Harnessing materials science literature for labeled microscopy datasets,” Patterns 4, 100843 (2023).Corresponding author information

RETRACTED: Wang et al. Panoramic Manifold Projection (Panoramap) for Single-Cell Data Dimensionality Reduction and Visualization. Int. J. Mol. Sci. 2022, 23, 7775.

The Int [...].

Transferable manifold projection embedded dictionary learning for multimode process monitoring

Transferable manifold projection embedded dictionary learning for multimode process monitoring

Manifold projection image segmentation for nano-XANES imaging

As spectral imaging techniques are becoming more prominent in science, advanced image segmentation algorithms are required to identify appropriate domains in these images. We present a version of image segmentation called manifold projection image segmentation (MPIS) that is generally applicable to a broad range of systems without the need for training because MPIS uses unsupervised machine learning with a few physically motivated hyperparameters. We apply MPIS to nanoscale x-ray absorption near edge structure (XANES) imaging, where XANES spectra are collected with nanometer spatial resolution. We show the superiority of manifold projection over linear transformations, such as the commonly used principal component analysis (PCA). Moreover, MPIS maintains accuracy while reducing computation time and sensitivity to noise compared to the standard nano-XANES imaging analysis procedure. Finally, we demonstrate how multimodal information, such as x-ray fluorescence data and spatial location of pixels, can be incorporated into the MPIS framework. We propose that MPIS is adaptable for any spectral imaging technique, including scanning transmission x-ray microscopy, where the length scale of domains is larger than the resolution of the experiment.

Multimodal fusion diagnosis of Alzheimer’s disease based on FDG-PET generation

Multimodal fusion diagnosis of Alzheimer’s disease based on FDG-PET generation

FGDA-GS: Fast guided decision attack based on gradient signs for skeletal action recognition

FGDA-GS: Fast guided decision attack based on gradient signs for skeletal action recognition

Stability of longitudinal sediment waves formed by turbidity currents: linear and weakly nonlinear perspectives

We perform the linear and weakly nonlinear stability analyses of longitudinal sediment waves triggered by the interaction of turbidity currents with an erodible bed. The mathematical framework is based on the two-dimensional flow equations, advection–diffusion equation for sediment concentration and Exner equation for bed evolution. Using the standard linearization, the linear analysis offers the resonant wavenumber that maximizes the growth rate of sediment waves. We study the influence of the key parameters, such as the gravitational parameter, longitudinal bed slope, Rouse number, shear Reynolds number, relative roughness number and erosion coefficient, on the growth rate, resonant wavenumber and phase velocity. The results reveal that the longitudinal sediment waves migrate both upstream and downstream. By plotting the gravitational parameter against the dimensionless wavenumber, we obtain a stability diagram that accurately captures the experimental data plots within the unstable zone. Additionally, we explore the perturbation fields of velocity components, pressure and sediment concentration. Employing the centre manifold projection, the weakly nonlinear analysis provides the equilibrium amplitude of sediment waves and its sensitivity to the key parameters. The predicted wavelength and amplitude of sediment waves are comparable with field observations in the Atlantic Ocean, the Sea of Japan and the Pacific Ocean.

Accounting for diverse feature-types improves patient stratification on tabular clinical datasets

Accounting for diverse feature-types improves patient stratification on tabular clinical datasets

MP50-11 INTEGRATIVE MULTI-OMICS PROFILING IN PENILE SQUAMOUS CELL CARCINOMA

You have accessJournal of UrologyCME1 Apr 2023MP50-11 INTEGRATIVE MULTI-OMICS PROFILING IN PENILE SQUAMOUS CELL CARCINOMA Timothy Shaw, Alyssa Obermayer, Dalia Ercan, Paul Stewart, Jad Chahoud, Peter Johnstone, Philippe Spiess, Steven Eschrich, and Daniel Grass Timothy ShawTimothy Shaw More articles by this author , Alyssa ObermayerAlyssa Obermayer More articles by this author , Dalia ErcanDalia Ercan More articles by this author , Paul StewartPaul Stewart More articles by this author , Jad ChahoudJad Chahoud More articles by this author , Peter JohnstonePeter Johnstone More articles by this author , Philippe SpiessPhilippe Spiess More articles by this author , Steven EschrichSteven Eschrich More articles by this author , and Daniel GrassDaniel Grass More articles by this author View All Author Informationhttps://doi.org/10.1097/JU.0000000000003298.11AboutPDF ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsReprints ShareFacebookLinked InTwitterEmail Abstract INTRODUCTION AND OBJECTIVE: Penile squamous cell carcinoma (PSCC) is rare and about 50% of cases are related to HPV. Relapse after definitive therapy or first-line chemotherapy portends poor survival, thus we hypothesized novel targeting strategies in PSCC may be revealed by relating molecular pathways and patient survival. METHODS: We performed RNA-Seq on 37 PSCC tumors (HPV+=16, HPV-=16, unknown=5) from patients at our institution; 16 samples also underwent bulk proteomics with mass spectrometry. PATH-SURVEIOR is a framework for analyzing and visualizing “on-the-fly” associations of pathways with clinical endpoints. We applied PATH-SURVEIOR to perform single sample gene set enrichment (ssGSEA) on the PSCC RNA and protein expression data. Pathway scores and gene expression were dichotomized by the median to group samples as above/below for each pathway or individual gene expression. Multi-omics integration of pathways related to HPV status was assessed with DRPPM-EASY. TCGA data was downloaded from cBioportal. Uniform manifold projection (UMAP) was performed with the umap R package. RESULTS: We first performed an unsupervised analysis on PSCC and TCGA RNA data and UMAP found similarities with PSCC and other SCCs. Next, we associated pathways/genes and overall survival (OS) through univariable Cox regression and filtered results by hazard ratio (HR>1) and p-value<0.05, followed by HR ranking to define pathways/genes related with high-risk patients. Genes related to OS risk in PSCC included higher levels of DLL4, MET, and MSLN. Also, a cisplatin resistance gene set was linked to poor OS in a subset of PSCC patients. Systematic analysis of gene sets found expression associated with HPV as one prognostic feature based on an unsupervised pathway-level survival analysis; projection of this gene-set feature in the TCGA manifold space revealed an HPV-related cluster that included PSCC and other HPV-related cancers. By integrative GSEA of RNA and proteomics data, we found E2F and DNA damage repair pathways were enriched in HPV+ tumors, whereas HPV-tumors had features of EMT and inflammation. Whereas most RNAs and proteins aligned in direction of expression, several genes (e.g., TP53, MAGI3) were discordant to protein expression in HPV+ tumors, which may imply HPV-mediated translational regulation of protein subsets. CONCLUSIONS: Integrative analysis relating multi-omics data to patient survival from PSCC and TCGA samples identified features consistent with other HPV-related cancers, as well as potential novel targets to consider in PSCC. Source of Funding: Congressionally Directed Medical Research Programs: Rare Cancer Research Programs © 2023 by American Urological Association Education and Research, Inc.FiguresReferencesRelatedDetails Volume 209Issue Supplement 4April 2023Page: e689 Advertisement Copyright & Permissions© 2023 by American Urological Association Education and Research, Inc.MetricsAuthor Information Timothy Shaw More articles by this author Alyssa Obermayer More articles by this author Dalia Ercan More articles by this author Paul Stewart More articles by this author Jad Chahoud More articles by this author Peter Johnstone More articles by this author Philippe Spiess More articles by this author Steven Eschrich More articles by this author Daniel Grass More articles by this author Expand All Advertisement PDF downloadLoading ...

Multi-temporal grass hyperspectral classification via full pixel decomposition spectral manifold projection and boosting active learning model

ABSTRACT Grassland detection is essential for sustainable development, particularly in ecological and animal husbandry. Hyperspectral can retrieve the functional characteristics of vegetation species, nevertheless, the efficiency of grassland communities with rich species at different periods needs to be improved. In this study, the multi-temporal grass hyperspectral classification via full pixel decomposition spectral manifold projection and boosting active learning (FPD-SMP-BAL) model is proposed to relieve the above issue. Considering the plentiful spectral information of hyperspectral, we proposed a novel preprocessing method of spectral manifold projection based on full pixel decomposition (FPD-SMP), which reveals the internal nonlinear structure of global hyperspectral and alleviate the data redundancy. Furthermore, we innovatively designed the boosting active learning (BAL) classification model with an effective optimization strategy, which can decrease the cost of sample labeling by the minimal loss value. An on-site hyperspectral acquisition platform was assembled, and a dataset of 4,200 samples was established from the perspective of vegetation species composition and multi-stages to verify the validity of the model. Experimental results show that the proposed FPD-SMP-BAL framework reaches an overall accuracy of 94.94%, which is superior to some state-of-the-art methods in the classification of multi-temporal grass hyperspectral images, and also provides a reference for related research of grassland dynamic monitoring.

ScTopoGAN: unsupervised manifold alignment of single-cell data

Abstract Motivation Single-cell technologies allow deep characterization of different molecular aspects of cells. Integrating these modalities provides a comprehensive view of cellular identity. Current integration methods rely on overlapping features or cells to link datasets measuring different modalities, limiting their application to experiments where different molecular layers are profiled in different subsets of cells. Results We present scTopoGAN, a method for unsupervised manifold alignment of single-cell datasets with non-overlapping cells or features. We use topological autoencoders (topoAE) to obtain latent representations of each modality separately. A topology-guided Generative Adversarial Network then aligns these latent representations into a common space. We show that scTopoGAN outperforms state-of-the-art manifold alignment methods in complete unsupervised settings. Interestingly, the topoAE for individual modalities also showed better performance in preserving the original structure of the data in the low-dimensional representations when compared to other manifold projection methods. Taken together, we show that the concept of topology preservation might be a powerful tool to align multiple single modality datasets, unleashing the potential of multi-omic interpretations of cells. Availability and implementation Implementation available on GitHub (https://github.com/AkashCiel/scTopoGAN). All datasets used in this study are publicly available.

Purifying Adversarial Images Using Adversarial Autoencoder With Conditional Normalizing Flows

We present a <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">target-agnostic</i> adversarial autoencoder with conditional normalizing flows specifically designed to, given any <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">unlabeled</i> image dataset, purify adversarial samples into clean images, i.e., remove adversarial noise from the images while preserving their visual quality. In our model interpretation, samples are processed by manifold projection in which the encoder brings the sample back into a posterior data distribution in latent space so that the sample is less likely to be irregular to the learned representation of any target classifier. Normalizing flows conditioned on top of our hybrid network structure and walk-back training are used to deal with common drawbacks of generative model and autoencoder-based approaches: not only the trade-off between compression loss and over-fitting on training data but also the <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">structural model dependency</i> on dataset classes and labels. Experiments demonstrated that our proposed model is preferable to existing target-agnostic adversarial defense methods particularly for large and unlabeled image datasets.

Deep Manifold Harmonic Network With Dual Attention for Brain Disorder Classification.

Numerous studies have shown that accurate analysis of neurological disorders contributes to the early diagnosis of brain disorders and provides a window to diagnose psychiatric disorders due to brain atrophy. The emergence of geometric deep learning approaches provides a new way to characterize geometric variations on brain networks. However, brain network data suffer from high heterogeneity and noise. Consequently, geometric deep learning methods struggle to identify discriminative and clinically meaningful representations from complex brain networks, resulting in poor diagnostic accuracy. Hence, the primary challenge in the diagnosis of brain diseases is to enhance the identification of discriminative features. To this end, this paper presents a dual-attention deep manifold harmonic discrimination (DA-DMHD) method for early diagnosis of neurodegenerative diseases. Here, a low-dimensional manifold projection is first learned to comprehensively exploit the geometric features of the brain network. Further, attention blocks with discrimination are proposed to learn a representation, which facilitates learning of group-dependent discriminant matrices to guide downstream analysis of group-specific references. Our proposed DA-DMHD model is evaluated on two independent datasets, ADNI and ADHD-200. Experimental results demonstrate that the model can tackle the hard-to-capture challenge of heterogeneous brain network topological differences and obtain excellent classifying performance in both accuracy and robustness compared with several existing state-of-the-art methods.

LDA-MIG Detectors for Maritime Targets in Nonhomogeneous Sea Clutter

This paper deals with the problem of detecting maritime targets embedded in nonhomogeneous sea clutter, where limited number of secondary data is available due to the heterogeneity of sea clutter. A class of linear discriminant analysis (LDA)-based matrix information geometry (MIG) detectors is proposed in the supervised scenario. As customary, Hermitian positive-definite (HPD) matrices are used to model the observational sample data, and the clutter covariance matrix of received dataset is estimated as geometric mean of the secondary HPD matrices. Given a set of training HPD matrices with class labels, that are elements of a higher-dimensional HPD matrix manifold, the LDA manifold projection learns a mapping from the higher-dimensional HPD matrix manifold to a lower-dimensional one subject to maximum discrimination. In the current study, the LDA manifold projection, with the cost function maximizing between-class distance while minimizing within-class distance, is formulated as an optimization problem in the Stiefel manifold. Four robust LDA-MIG detectors corresponding to different geometric measures are proposed. Numerical results based on both simulated radar clutter with interferences and real IPIX radar data show the advantage of the proposed LDA-MIG detectors against their counterparts without using LDA as well as the state-of-art maritime target detection methods in nonhomogeneous sea clutter.

Computerized analysis of pulmonary sounds using uniform manifold projection

Computerized analysis of pulmonary sounds using uniform manifold projection

Hamiltonian dynamics and targeted energy transfer of a grounded bistable nonlinear energy sink

A two-degree-of-freedom nonlinear mechanical system describing a damped linear oscillator weakly coupled to a lightweight grounded bistable nonlinear energy sink is proposed to investigate the targeted energy transfer (TET) under asymmetric local nonlinear potentials. The bifurcations of equilibria are discussed to reflect the bistable and asymmetric characteristics of the system. Through projections of isoenergetic manifolds, the topological features of the Hamiltonian dynamics under 1 : 1 internal resonance have been investigated, which induces conditions for energy localization and exchange. In the general case, the topological features of the Hamiltonian dynamics related to the nonlinear normal modes (NNMs) are classified by the Poincar e ́ map, the relations between the NNMs and the Hamiltonian dynamics have been interpreted through the frequency energy plots. The results show that introducing the asymmetric characteristics may enhance the TET and realize the complete TET. • The structure of a grounded NES combines both the characteristics of bistability and asymmetry to trigger complete TET is practical in engineering. • The capability of the grounded bistable NES to absorb the energy from the LO is evaluated by the transformed variable under 1:1 internal resonance conditions. And the dynamical properties of NNMs and the mechanism of these dynamical properties are studied by the Poincare map and the frequency energy plots (FEPs) in detail. • The introduction of asymmetry can optimize TET, which makes it possible to completely transfer the input energy to the NES and enhance the vibration absorbing characteristic of the grounded bistable NES.

Neurons on amoebae

We apply methods of machine-learning, such as neural networks, manifold learning and image processing, in order to study 2-dimensional amoebae in algebraic geometry and string theory. With the help of embedding manifold projection, we recover complicated conditions obtained from so-called lopsidedness. For certain cases it could even reach ∼ 99 % accuracy, in particular for the lopsided amoeba of F 0 with positive coefficients which we place primary focus. Using weights and biases, we also find good approximations to determine the genus for an amoeba at lower computational cost. In general, the models could easily predict the genus with over 90% accuracies. With similar techniques, we also investigate the membership problem, and image processing of the amoebae directly.

Panoramic Manifold Projection (Panoramap) for Single-Cell Data Dimensionality Reduction and Visualization.

Nonlinear dimensionality reduction (NLDR) methods such as t-Distributed Stochastic Neighbour Embedding (t-SNE) and Uniform Manifold Approximation and Projection (UMAP) have been widely used for biological data exploration, especially in single-cell analysis. However, the existing methods have drawbacks in preserving data’s geometric and topological structures. A high-dimensional data analysis method, called Panoramic manifold projection (Panoramap), was developed as an enhanced deep learning framework for structure-preserving NLDR. Panoramap enhances deep neural networks by using cross-layer geometry-preserving constraints. The constraints constitute the loss for deep manifold learning and serve as geometric regularizers for NLDR network training. Therefore, Panoramap has better performance in preserving global structures of the original data. Here, we apply Panoramap to single-cell datasets and show that Panoramap excels at delineating the cell type lineage/hierarchy and can reveal rare cell types. Panoramap can facilitate trajectory inference and has the potential to aid in the early diagnosis of tumors. Panoramap gives improved and more biologically plausible visualization and interpretation of single-cell data. Panoramap can be readily used in single-cell research domains and other research fields that involve high dimensional data analysis.