Search NASA⌕ Search

SEARCH · Search NASA

Results for “sparse data representation”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 163 records · Page 9

Bayesian differential programming for robust systems identification under uncertainty

This paper presents a machine learning framework for Bayesian systems identification from noisy, sparse and irregular observations of nonlinear dynamical systems. The proposed method takes advantage of recent developments in differentiable programming to propagate gradient information through ordinary differential equation solvers and perform Bayesian inference with respect to unknown model parameters using Hamiltonian Monte Carlo sampling. This allows an efficient inference of the posterior distributions over plausible models with quantified uncertainty, while the use of sparsity-promoting priors enables the discovery of interpretable and parsimonious representations for the underlying latent dynamics. A series of numerical studies is presented to demonstrate the effectiveness of the proposed methods, including nonlinear oscillators, predator–prey systems and examples from systems biology. Taken together, our findings put forth a flexible and robust workflow for data-driven model discovery under uncertainty. All codes and data accompanying this article are available at https://bit.ly/34FOJMj .

Science & Technology - Other Topics↗

TH-CD-206-09: Learning-Based MRI-CT Prostate Registration Using Spare Patch-Deformation Dictionary

Purpose: To enable MRI-guided prostate radiotherapy, MRI-CT deformable registration is required to map the MRI-defined tumor and key organ contours onto the CT images. Due to the intrinsic differences in grey-level intensity characteristics between MRI and CT images, the integration of MRI into CT-based radiotherapy is very challenging. We are developing a learning-based registration approach to address this technical challenge. Methods: We propose to estimate the deformation between MRI and CT images in a patch-wise fashion by using the sparse representation technique. Specifically, we assume that two image patches should follow the same deformation if their patch-wise appearance patterns are similar. We first extract a set of key points in the new CT image. Then, for each key point, we adaptively construct a coupled dictionary from the training MRI-CT images, where each coupled element includes both appearance and deformation of the same image patch. After calculating the sparse coefficients in representing the patch appearance of each key point based on the constructed dictionary, we can predict the deformation for this point by applying the same sparse coefficients to the respective deformations in the dictionary. Results: This registration technique was validated with 10 prostate-cancer patients’ data and its performance was compared with the commonly used free-form-deformation-based registration. Several landmarks in both images were identified to evaluate the accuracy of our approach. Overall, the averaged target registration error of the intensity-based registration and the proposed method was 3.8±0.4 mm and 1.9±0.3 mm, respectively. Conclusion: We have developed a novel prostate MR-CT registration approach based on patch-deformation dictionary, demonstrated its clinical feasibility, and validated its accuracy. This technique will either reduce or compensate for the effect of patient-specific treatment variation measured during the course of radiotherapy, is therefore well-suited for a number of MRI-guided adaptive radiotherapy, and potentially enhance prostate radiotherapy treatment outcome.

60 APPLIED LIFE SCIENCES↗

Global field reconstruction from sparse sensors with Veronoi tessellation-assisted deep learning

Achieving accurate and robust global situational awareness of a complex time-evolving field from a limited number of sensors has been a longstanding challenge. This reconstruction problem is especially difficult when sensors are sparsely positioned in a seemingly random or unorganized manner, which is often encountered in a range of scientific and engineering problems. Moreover, these sensors can be in motion and can become online or offline over time. The key leverage in addressing this scientific issue is the wealth of data accumulated from the sensors. As a solution to this problem, we propose a data-driven spatial field recovery technique founded on a structured grid-based deep-learning approach for arbitrary positioned sensors of any numbers. It should be noted that the naïve use of machine learning becomes prohibitively expensive for global field reconstruction and is furthermore not adaptable to an arbitrary number of sensors. In the present work, we consider the use of Voronoi tessellation to obtain a structured-grid representation from sensor locations enabling the computationally tractable use of convolutional neural networks. One of the central features of the present method is its compatibility with deep-learning based super-resolution reconstruction techniques for structured sensor data that are established for image processing. The proposed reconstruction technique is demonstrated for unsteady wake flow, geophysical data, and three-dimensional turbulence. The current framework is able to handle an arbitrary number of moving sensors, and thereby overcomes a major limitation with existing reconstruction methods. The presented technique opens a new pathway towards the practical use of neural networks for real-time global field estimation.

Fukami, Kai↗

Global Assimilation of Loon Stratospheric Balloon Observations

Project Loon has an overall goal of providing worldwide internet coverage using a network of long-durationsuper-pressure balloons. Since 2013, Loon has launched over 1600 balloons from multiple tropical and middlelatitude locations. These GPS tracked balloon trajectories provide lower stratospheric wind information overthe oceans and remote land areas where traditional radiosonde soundings are sparse, thus providing uniquecoverage of lower stratospheric winds. To fully investigate these Loon winds we: 1) compare the Loon windsto winds produced by a global data assimilation system (DAS: NASA GEOS) and 2) assimilate the Loon windsinto the same comprehensive DAS. Results show that in middle latitudes the Loon winds and DAS winds agreewell, and the Loon wind assimilation has only a minor impact on the forecasts. However, in the Tropics, thereis often a substantial difference between the assimilated winds and the observed Loon winds, of 8 m/s or morein magnitude. In these cases, assimilating the Loon winds significantly improves the meteorological analysesand subsequently the forecasts of the Loon winds. By highlighting cases where the Loon and DAS winds differ,these results can lead to improved understanding of stratospheric winds, especially in the tropics, as well asaiding analyses of the representation of dynamical forcing mechanisms in the GEOS model.

Coy, Lawrence↗

TH-E-18A-01: Developments in Monte Carlo Methods for Medical Imaging

Monte Carlo simulation methods are widely used in medical physics research and are starting to be implemented in clinical applications such as radiation therapy planning systems. Monte Carlo simulations offer the capability to accurately estimate quantities of interest that are challenging to measure experimentally while taking into account the realistic anatomy of an individual patient. Traditionally, practical application of Monte Carlo simulation codes in diagnostic imaging was limited by the need for large computational resources or long execution times. However, recent advancements in high-performance computing hardware, combined with a new generation of Monte Carlo simulation algorithms and novel postprocessing methods, are allowing for the computation of relevant imaging parameters of interest such as patient organ doses and scatter-to-primaryratios in radiographic projections in just a few seconds using affordable computational resources. Programmable Graphics Processing Units (GPUs), for example, provide a convenient, affordable platform for parallelized Monte Carlo executions that yield simulation times on the order of 10{sup 7} xray/ s. Even with GPU acceleration, however, Monte Carlo simulation times can be prohibitive for routine clinical practice. To reduce simulation times further, variance reduction techniques can be used to alter the probabilistic models underlying the x-ray tracking process, resulting in lower variance in the results without biasing the estimates. Other complementary strategies for further reductions in computation time are denoising of the Monte Carlo estimates and estimating (scoring) the quantity of interest at a sparse set of sampling locations (e.g. at a small number of detector pixels in a scatter simulation) followed by interpolation. Beyond reduction of the computational resources required for performing Monte Carlo simulations in medical imaging, the use of accurate representations of patient anatomy is crucial to the virtual generation of medical images and accurate estimation of radiation dose and other imaging parameters. For this, detailed computational phantoms of the patient anatomy must be utilized and implemented within the radiation transport code. Computational phantoms presently come in one of three format types, and in one of four morphometric categories. Format types include stylized (mathematical equation-based), voxel (segmented CT/MR images), and hybrid (NURBS and polygon mesh surfaces). Morphometric categories include reference (small library of phantoms by age at 50th height/weight percentile), patient-dependent (larger library of phantoms at various combinations of height/weight percentiles), patient-sculpted (phantoms altered to match the patient's unique outer body contour), and finally, patient-specific (an exact representation of the patient with respect to both body contour and internal anatomy). The existence and availability of these phantoms represents a very important advance for the simulation of realistic medical imaging applications using Monte Carlo methods. New Monte Carlo simulation codes need to be thoroughly validated before they can be used to perform novel research. Ideally, the validation process would involve comparison of results with those of an experimental measurement, but accurate replication of experimental conditions can be very challenging. It is very common to validate new Monte Carlo simulations by replicating previously published simulation results of similar experiments. This process, however, is commonly problematic due to the lack of sufficient information in the published reports of previous work so as to be able to replicate the simulation in detail. To aid in this process, the AAPM Task Group 195 prepared a report in which six different imaging research experiments commonly performed using Monte Carlo simulations are described and their results provided. The simulation conditions of all six cases are provided in full detail, with all necessary data on material composition, source, geometry, scoring and other parameters provided. The results of these simulations when performed with the four most common publicly available Monte Carlo packages are also provided in tabular form. The Task Group 195 Report will be useful for researchers needing to validate their Monte Carlo work, and for trainees needing to learn Monte Carlo simulation methods. In this symposium we will review the recent advancements in highperformance computing hardware enabling the reduction in computational resources needed for Monte Carlo simulations in medical imaging. We will review variance reduction techniques commonly applied in Monte Carlo simulations of medical imaging systems and present implementation strategies for efficient combination of these techniques with GPU acceleration. Trade-offs involved in Monte Carlo acceleration by means of denoising and “sparse sampling” will be discussed. A method for rapid scatter correction in cone-beam CT (<5 min/scan) will be presented as an illustration of the simulation speeds achievable with optimized Monte Carlo simulations. We will also discuss the development, availability, and capability of the various combinations of computational phantoms for Monte Carlo simulation of medical imaging systems. Finally, we will review some examples of experimental validation of Monte Carlo simulations and will present the AAPM Task Group 195 Report. Learning Objectives: Describe the advances in hardware available for performing Monte Carlo simulations in high performance computing environments. Explain variance reduction, denoising and sparse sampling techniques available for reduction of computational time needed for Monte Carlo simulations of medical imaging. List and compare the computational anthropomorphic phantoms currently available for more accurate assessment of medical imaging parameters in Monte Carlo simulations. Describe experimental methods used for validation of Monte Carlo simulations in medical imaging. Describe the AAPM Task Group 195 Report and its use for validation and teaching of Monte Carlo simulations in medical imaging.

60 APPLIED LIFE SCIENCES↗

Robust Group Subspace Recovery: A New Approach for Multi-Modality Data Fusion

Robust Subspace Recovery (RoSuRe) algorithm was recently introduced as a principled and numerically efficient algorithm that unfolds underlying Unions of Subspaces (UoS) structure, present in the data. The union of Subspaces (UoS) is capable of identifying more complex trends in data sets than simple linear models. In this work, we build on and extend RoSuRe to prospect the structure of different data modalities individually. We propose a novel multi-modal data fusion approach based on group sparsity which we refer to as Robust Group Subspace Recovery (RoGSuRe). Relying on a bi-sparsity pursuit paradigm and non-smooth optimization techniques, the introduced framework learns a new joint representation of the time series from different data modalities, respecting an underlying UoS model. We subsequently integrate the obtained structures to form a unified subspace structure. The proposed approach exploits the structural dependencies between the different modalities data to cluster the associated target objects. The resulting fusion of the unlabeled sensors’ data from experiments on audio and magnetic data has shown that our method is competitive with other state of the art subspace clustering methods. The resulting UoS structure is employed to classify newly observed data points, highlighting the abstraction capacity of the proposed method.

47 OTHER INSTRUMENTATION↗

Scaling neural simulations in STACS

Abstract As modern neuroscience tools acquire more details about the brain, the need to move towards biological-scale neural simulations continues to grow. However, effective simulations at scale remain a challenge. Beyond just the tooling required to enable parallel execution, there is also the unique structure of the synaptic interconnectivity, which is globally sparse but has relatively high connection density and non-local interactions per neuron. There are also various practicalities to consider in high performance computing applications, such as the need for serializing neural networks to support potentially long-running simulations that require checkpoint-restart. Although acceleration on neuromorphic hardware is also a possibility, development in this space can be difficult as hardware support tends to vary between platforms and software support for larger scale models also tends to be limited. In this paper, we focus our attention on Simulation Tool for Asynchronous Cortical Streams (STACS), a spiking neural network simulator that leverages the Charm++ parallel programming framework, with the goal of supporting biological-scale simulations as well as interoperability between platforms. Central to these goals is the implementation of scalable data structures suitable for efficiently distributing a network across parallel partitions. Here, we discuss a straightforward extension of a parallel data format with a history of use in graph partitioners, which also serves as a portable intermediate representation for different neuromorphic backends. We perform scaling studies on the Summit supercomputer, examining the capabilities of STACS in terms of network build and storage, partitioning, and execution. We highlight how a suitably partitioned, spatially dependent synaptic structure introduces a communication workload well-suited to the multicast communication supported by Charm++. We evaluate the strong and weak scaling behavior for networks on the order of millions of neurons and billions of synapses, and show that STACS achieves competitive levels of parallel efficiency.

59 BASIC BIOLOGICAL SCIENCES↗

Global Assimilation of EOS-Aura Data as a Means of Mapping Ozone Distribution in the Lower Stratosphere and Troposphere

Ozone in the lower stratosphere and the troposphere plays an important role in forcing the climate. However, the global ozone distribution in this region is not well known because of the sparse distribution of in-situ data and the poor sensitivity of satellite based observations to the lowermost of the atmosphere. The Ozone Monitoring Instrument (OMI) and Microwave Limb Sounder (MLS) instruments on EOS-Aura provide information on the total ozone column and the stratospheric ozone profile. This data has been assimilated into NASA s Global Earth Observing System, Version 5 (GEOS-5) data assimilation system (DAS). We will discuss the results of assimilating three years of OMI and MLS data into GEOS-5. This data was assimilated alongside meteorological observations from both conventional sources and satellite instruments. Previous studies have shown that combining observations from these instruments through the Trajectory Tropospheric Ozone Residual methodology (TTOR) or using data assimilation can yield useful, yet low biased, estimates of the tropospheric ozone budget. We show that the assimilated ozone fields in this updated version of GEOS-5 exhibit an excellent agreement with ozone sonde and High Resolution Dynamics Limb Sounder (HIRDLS) data in the lower stratosphere in terms of spatial and temporal variability as well as integrated ozone abundances. Good representation of small-scale vertical features follows from combining the MLS data with the assimilated meteorological fields. We then demonstrate how this information can be used to calculate the Stratosphere - Troposphere Exchange of ozone and its contribution to the tropospheric ozone column in GEOS-5. Evaluations of tropospheric ozone distributions from the assimilation will be made by comparisons with sonde and other in-situ observations.

Wargan, Krzysztof↗

A production-tagged aerosol module for Earth system models, OsloAero5.3 – extensions and updates for CAM5.3-Oslo

Abstract. We document model updates and present and discuss modeling and validation results from a further developed production-tagged aerosol module, OsloAero5.3, for use in Earth system models. The aerosol module has in this study been implemented and applied in CAM5.3-Oslo. This model is based on CAM5.3–CESM1.2 and its own predecessor model version CAM4-Oslo. OsloAero5.3 has improved treatment of emissions, aerosol chemistry, particle life cycle, and aerosol–cloud interactions compared to its predecessor OsloAero4.0 in CAM4-Oslo. The main new features consist of improved aerosol sources; the module now explicitly accounts for aerosol particle nucleation and secondary organic aerosol production, with new emissions schemes also for sea salt, dimethyl sulfide (DMS), and marine primary organics. Mineral dust emissions are updated as well, adopting the formulation of CESM1.2. The improved model representation of aerosol–cloud interactions now resolves heterogeneous ice nucleation based on black carbon (BC) and mineral dust calculated by the model and treats the activation of cloud condensation nuclei (CCN) as in CAM5.3. Compared to OsloAero4.0 in CAM4-Oslo, the black carbon (BC) mass concentrations are less excessive aloft, with a better fit to observations. Near-surface mass concentrations of BC and sea salt aerosols are also less biased, while sulfate and mineral dust are slightly more biased. Although appearing quite similar for CAM5.3-Oslo and CAM4-Oslo, the validation results for organic matter (OM) are inconclusive, since both of the respective versions of OsloAero are equipped with a limited number of OM tracers for the sake of computational efficiency. Any information about the assumed mass ratios of OM to organic carbon (OC) for different types of OM sources is lost in the transport module. Assuming that observed OC concentrations scaled by 1.4 are representative for the modeled OM concentrations, CAM5.3-Oslo with OsloAero5.3 is slightly inferior for the very sparsely available observation data. Comparing clear-sky column-integrated optical properties with data from ground-based remote sensing, we find a negative bias in optical depth globally; however, it is not as strong as in CAM4-Oslo, but has positive biases in some areas typically dominated by mineral dust emissions. Aerosol absorption has a larger negative bias than the optical depth globally. This is reflected in a lower positive bias in areas where mineral dust is the main contributor to absorption. Globally, the low bias in absorption is smaller than in CAM4-Oslo. The Ångström parameter exhibits small biases both globally and regionally, suggesting that the aerosol particle sizes are reasonably well represented. Cloud-top droplet number concentrations over oceans are generally underestimated compared to satellite retrievals, but seem to be overestimated downwind of major emissions of dust and biomass burning sources. Finally, we find small changes in direct radiative forcing at the top of the atmosphere, while the cloud radiative forcing due to anthropogenic aerosols is now more negative than in CAM4-Oslo, being on the strong side compared to the multi-model estimate in IPCC AR5. Although not all validation results in this study show improvement for the present CAM5.3-Oslo version, the extended and updated aerosol module OsloAero5.3 is more advanced and applicable than its predecessor OsloAero4.0, as it includes new parameterizations that more readily facilitate sensitivity and process studies and use in climate and Earth system model studies in general.

54 ENVIRONMENTAL SCIENCES↗

Implementing and Benchmarking the Locally Competitive Algorithm on the Loihi 2 Neuromorphic Processor

Neuromorphic processors have garnered considerable interest in recent years for their potential in enabling energy-efficient and high-speed computing. The Locally Competative Algorithm (LCA) has been utilized for power efficient sparse coding on neuromophic processors, including the first Loihi processor \cite{appletospikes, loihi1}. With the Loihi 2 processor enabling custom neuron models and graded spike communication, more complex implementations of LCA are possible \cite{loihi2}. We present a new implementation of LCA designed for the Loihi 2 processor and perform an initial set of benchmarks comparing it to LCA on CPU and GPU devices. In these experiments LCA on Loihi 2 is faster and orders of magnitude more efficient, while maintaining similar reconstruction quality. We find this performance improvement increases as the LCA parameters are tuned towards greater representation sparsity. Our study highlights the potential of neuromorphic processors, particularly Loihi 2, in enabling intelligent,autonomous, real-time processing on small robots, satellite where there are strict SWaP (small, lightweighr, and low-power) requirement. By demonstrating the superior performance of LCA on Loihi 2 compared to conventional computing device, our study suggests that Loihi 2 could be a valuable tool in advancing these types of applications. Overall, our study highlights the potential of neuromorphic processors for efficient and accurate data processing on resource-constrained devices.

Parpart, Gavin G.↗

The Influence of Prescribed Boundary Conditions on Near-Surface Temperature over the Arctic in the MERRA-2 Atmospheric Model

An accurate historical record of evolving Arctic conditions is integral to furthering our understanding of climate processes and to providing a foundation for predicting future climate scenarios in northern high latitudes. Atmospheric reanalyses are seen as an important source of information on the recent past for the data-sparse Arctic region. An assessment of near-surface Arctic air temperatures finds significant discrepancies among the various modern reanalyses. An important point is the treatment of surface boundary conditions: specifically, the sea ice cover and sea surface temperatures (SSTs) over the Arctic Ocean. Reanalyses use different methodologies and data sources for SSTs and sea ice concentration boundary forcing. Notably, the Modern Era Retrospective analysis for Research and Applications, version 2 (MERRA-2) and the European Centre for Medium-Range Weather Forecasts Interim Re-Analysis (ERA-Interim) both use boundary forcing derived from the Operational Sea Surface Temperature and Sea Ice Analysis (OSTIA) over an extended, overlapping period of time. This allows for an examination of differences between the two systems while both concurrently employ the same fractional sea ice coverage. To further understand these differences, an ensemble of AMIP-style simulations using the MERRA-2 atmospheric model - but without data assimilation - shows considerable differences in Arctic temperatures as compared to reanalyses, particularly in autumn and winter months. Results from the AMIP simulations suggest that the surface representation over sea ice used in the MERRA-2 model provides an intrinsic warm bias and obfuscates Arctic Amplification, an established feature present in observations and reanalyses. An additional ensemble of AMIP-style simulations using the MERRA-2 atmospheric model was performed using boundary conditions derived from the ERA-Interim reanalysis. An in-depth comparison of surface temperatures over the Arctic from the two reanalyses and two AMIP-style ensembles will be presented, along with an assessment of the effects of the varying Arctic temperature time series on the atmospheric general circulation and energy budget.

Collow, Allison B. Marquardt↗

Latent code-based fusion: A Volterra neural network approach

We propose a deep structure encoder using Volterra Neural Networks (VNNs) to seek a latent representation of multi-modal data whose features are jointly captured by a union of subspaces. The so-called self-representation embedding of the latent codes leads to a simplified fusion which is driven by a similarly constructed decoding. The Volterra Filter architecture achieved reduction in parameter complexity is primarily due to controlled non-linearities being introduced by the higher-order convolutions in lieu of generalized activation functions. Experimental results on two different datasets have shown a significant improvement in the clustering performance for VNNs auto-encoder over conventional Convolutional Neural Networks (CNNs) auto-encoder. In addition, we also show that the proposed approach demonstrates a much-improved sample complexity over CNN-based auto-encoder with a robust classification performance.

97 MATHEMATICS AND COMPUTING↗

Inference of phase field fracture models

The phase field approach to modeling fracture uses a diffuse damage field to represent cracks. This representation mollifies singularities that arise in computations with sharp interface models and some of the resultant difficulties in the mathematical and numerical treatment of fracture. Phase field fracture models have proven effective at representing crack propagation, branching, and merging. Specific formulations, beginning with brittle fracture, have also been shown to converge to classical solutions. Extensions to cover the range of material failure, including ductile and cohesive fracture, lead to an array of possible models. There exists a large body of literature focusing on this class of models and on the impact of model form on the predicted crack evolution. However, there have not been systematic studies into how optimal models may be chosen. Here, we take a first step in this direction by developing formal methods for identification of the best parsimonious model of phase field fracture given full-field data on the damage and deformation fields. We consider some of the main models that have been used for the degradation of elastic response due to damage and its propagation. Our approach builds upon Variational System Identification (VSI), a weak form variant of the Sparse Identification of Nonlinear Dynamics (SINDy). Furthermore, in this first communication we focus on synthetically generated data but we also consider central issues associated with the use of experimental full-field data, such as data sparsity and noise.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Stability-driven nonnegative matrix factorization to interpret spatial gene expression and build local gene networks

Spatial gene expression patterns enable the detection of local covariability and are extremely useful for identifying local gene interactions during normal development. The abundance of spatial expression data in recent years has led to the modeling and analysis of regulatory networks. The inherent complexity of such data makes it a challenge to extract biological information. We developed staNMF, a method that combines a scalable implementation of nonnegative matrix factorization (NMF) with a new stability-driven model selection criterion. When applied to a set of Drosophila early embryonic spatial gene expression images, one of the largest datasets of its kind, staNMF identified 21 principal patterns (PP). Providing a compact yet biologically interpretable representation of Drosophila expression patterns, PP are comparable to a fate map generated experimentally by laser ablation and show exceptional promise as a data-driven alternative to manual annotations. Our analysis mapped genes to cell-fate programs and assigned putative biological roles to uncharacterized genes. Finally, we used the PP to generate local transcription factor regulatory networks. Spatially local correlation networks were constructed for six PP that span along the embryonic anterior-posterior axis. Using a two-tail 5% cutoff on correlation, we reproduced 10 of the 11 links in the well-studied gap gene network. In conclusion, the performance of PP with the Drosophila data suggests that staNMF provides informative decompositions and constitutes a useful computational lens through which to extract biological insight from complex and often noisy gene expression data.

59 BASIC BIOLOGICAL SCIENCES↗

Knowledge Oriented Graph Unified Transformer (KOGUT) v0.1

KOGUT — Knowledge Oriented Graph Unified Transformer KOGUT implements the Relational Graph Transformer (RelGT) architecture for knowledge graph link prediction in biological domains, with a primary focus on microbial growth media prediction. While the original RelGT (arXiv:2505.10960) targets relational tables, time series, and multi-table databases, KOGUT adapts this architecture for heterogeneous biological knowledge graphs, providing first-in-class AI predictive models for microbial cultivation. Key Adaptations Beyond Original RelGT: - Knowledge Graph Focus: Applied to biological KGs with semantic node types (taxa, chemicals, media, phenotypes, environments) versus generic relational database tables, trained on the KG-Microbe knowledge graph (1.3M entities, 2.9M edges, 24 relation types). - Multimodal Node Encoding: Integrates node labels, categories, descriptions, and synonyms from KG metadata through learned embedding layers—adapting relational column features to graph node attributes with textual semantics. - Extended K-Hop Subgraph Strategy: Optimized neighborhood sampling (3-hop default, configurable up to 200 nodes) tuned for sparse biological networks, building on the original local-global attention framework with biological relation preservation. - Biolink Predicate Preservation: Type-specific transformations for 24 biological edge semantics (occurs_in, consumes, produces, has_phenotype, subclass_of) beyond standard relational foreign keys, enabling multi-relation link prediction. - Inductive Learning Support: Enables zero-shot predictions for novel taxa through feature-based embeddings (temperature, oxygen requirements, gram stain, cell shape), extending the original transductive relational benchmark scope to uncultured microorganisms. CheapSOTA Performance Optimizations (This Distribution): - VQ-EMA Centroid Attention: Vector quantization with exponential moving average for improved global context modeling (+5-10% MRR improvement). - HDF5 Precomputed Data Loading: One-time preprocessing of k-hop subgraphs to eliminate redundant graph traversals (2-5× training speedup). - Distributed Data Parallel Training: Multi-GPU support for scaling to larger knowledge graphs (tested on 4× NVIDIA A100 GPUs at NERSC Perlmutter). - Mixed Precision Training: Automatic mixed precision (AMP) for memory efficiency and faster training. Advantages Over Standard Knowledge Graph Embedding Models: Combines RelGT's proven multi-element tokenization (features, type, hop, structure) with graph-native biological representations, enabling interpretable link prediction across heterogeneous entities that standard embedding models (TransE, RotatE, ComplEx) and table-based transformers cannot directly model. Achieves near-perfect performance on microbial growth media prediction (MRR: 0.9966, Precision@1: 0.9932, Hit@10: 1.0000) while maintaining explainability through attention-based reasoning over biological pathways. Training Data: - KG-Microbe merged knowledge graph: 1,379,337 nodes, 2,960,472 edges - 24 biological relation types including taxonomic hierarchies, metabolic interactions, phenotype associations, and environmental relationships - Primary prediction task: Growth media suitability for microbial taxa (biolink:occurs_in, 50K edges) - Multi-relation capability: Predicts links for any of the 24 relation types, including chemical consumption/production, phenotype associations, and taxonomic classification Citation: Original RelGT Architecture: Dwivedi et al., "Relational Graph Transformer", arXiv:2505.10960, 2025 KOGUT Implementation: Knowledge Oriented Graph Unified Transformer for Microbial Growth Media Prediction Developed at Lawrence Berkeley National Laboratory (LBNL) Trained on NERSC Perlmutter supercomputer

Joachimiak, Marcin [Lawrence Berkeley National Lab↗

A Space-Filling Visualization Technique for Multivariate Small World Graphs

We introduce an information visualization technique, known as GreenCurve, for large sparse graphs that exhibit small world properties. Our fractal-based design approach uses spatial cues to approximate the node connections and thus eliminates the links between the nodes in the visualization. The paper describes a sophisticated algorithm to order the neighboring nodes of a large sparse graph by solving the Fiedler vector of its graph Laplacian, and then fold the graph nodes into a space-filling fractal curve based on the Fiedler vector. The result is a highly compact visualization that gives a succinct overview of the graph with guaranteed visibility of every graph node. We show in the paper that the GreenCurve technology is (1) theoretically sustainable by introducing an error estimation metric to measure the fidelity of the new graph representation, (2) empirically rigorous by conducting a usability study to investigate its strengths and weaknesses against the traditional graph layout, and (3) pragmatically feasible by applying it to analyze stressed conditions of the large scale electric power grid on the west coast.

99 GENERAL AND MISCELLANEOUS↗

Remote Sensing of Lineage Functional Types for Modeling and Monitoring Biodiversity

Hyperspectral remote sensing has the potential to continuously scale plant function and plant diversity information from landscape to global extents. Numerous studies have indicated that VSWIR (400-2500 nm) reflectance properties of vegetation capture evolutionarily conserved biochemical, structural, and other functional attributes of plant species. Spectral properties conserved in plants provide the opportunity to both 1) aggregate species into lineages with improved classification accuracy and 2) link those lineages directly to plant traits. Full realization of this goal will enable parameterization of Land Surface Models (LSMs) with remotely sensed information, e.g., canopy nitrogen, and better representations of biodiversity and functional diversity in biogeographic studies. In this study, we use hyperspectral AVIRIS data from the 2013 HyspIRI campaign over the Southern Sierra Nevada, California flight box to investigate the potential for incorporating evolutionary thinking into landcover classification. We link the airborne hyperspectral data with vegetation plot data from roughly 1372 surveys and a phylogeny representing 1361 species. We aggregate species into lineages ranging from species level groups down to similar number of Plant Functional Types as often used in LSMs. We assessed the ability of Random Forest and Partial Least Squares Discriminant Analysis to discriminate across these different phylogenetic scales and determine the optimal number of lineages to classify. Although there are some temporal and spatial differences in our training data, our best approaches achieved moderate classification accuracy (Kappa > 0.65). Given an optimal number of lineages, we explored approaches to improve classifications including machine learning and unmixing approaches. This work suggests that lineage-based methods may be a promising way to leverage the huge amounts of data that will come from high resolution and high return interval hyperspectral data planned for the Surface Biology and Geology mission with sparsely sampled existing ground-based ecological data.

Hyperspectral↗