SEARCH · Search NASA
Results for “outliers”
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.
Know Your Space: Inlier and Outlier Construction for Calibrating Medical OOD Detectors
This software offers methods and functions for training calibrated out-of-distribution detectors for medical image classification tasks. It includes functionalities for training, synthesizing data augmentations, calibration, and out-of-distribution detection. Developed using PyTorch, this software is compatible with standard neural network architectures used for imaging data. Additionally, it provides capabilities to compute evaluation metrics for assessing the performance and quality of the detectors.
BROOD: Bilevel and Robust Optimization and Outlier Detection for Efficient Tuning of High-Energy Physics Event Generators
The parameters in Monte Carlo (MC) event generators are tuned on experimental measurements by evaluating the goodness of fit between the data and the MC predictions. The relative importance of each measurement is adjusted manually in an often time-consuming, iterative process to meet different experimental needs. In this work, we introduce several optimization formulations and algorithms with new decision criteria for streamlining and automating this process. These algorithms are designed for two formulations: bilevel optimization and robust optimization. Both formulations are applied to the datasets used in the ATLAS A14 tune and to the dedicated hadronization datasets generated by the SHERPA generator, respectively. The corresponding tuned generator parameters are compared using three metrics. We compare the quality of our automatic tunes to the published ATLAS A14 tune. Moreover, we analyze the impact of a pre-processing step that excludes data that cannot be described by the physics models used in the MC event generators.
Advancing Vision-based Feedback and Convolutional Neural Networks for Visual Outlier Detection
Machine learning has matured into a technology that has immediate applicability to the surveillance needs of nuclear material storage containers. These containers at LANL are the barrier preventing release of radioactive material to the workers, public, and environment during the storage period of the material. Annual surveillance activities can only provide coverage on a handful of containers. There is a significant need for surveillance tools to identify potential issues and precursors to containment failure that can be used during opportunistic inspections and, more generally, outside of annual surveillance activities. In this report we provide details on the advancement of our proposed embodiment that combines an automation system for taking pictures and a high-accuracy machine learning-driven object detection software. We further showcase the improvements on the software side with progress on extracting unique identification features and advances in detecting damage. The current state of the system captures subject matter expert training and a space-conscious design whose implementation is envisioned in the near future.
Revisiting Inlier and Outlier Specification for Improved Out-of-Distribution Detection
Explore the source record for details and available documents.
Unsupervised Techniques for Outlier Identification in Alloy Datasets
TMS 2023 Annual Meeting & Exhibition, San Diego, CA, March 19-23, 2023
Six Machine-Learning Methods for Predicting Hospital-Stay Duration for Patients with Sepsis: A Comparative Study
Sepsis is a life-threatening medical condition that, if not treated promptly, can result in tissue damage, organ failure, and death. According to the Centers for Disease Control, about 270,000 individuals die of sepsis in the US each year. Further, sepsis expenditures accounted for 13% of total US hospital costs in 2013, totaling more than $24 billion. Our project objectives were to determine if Machine Learning algorithms could reliably predict hospital stay duration for patients with sepsis. The data set we used has been de-identified and is freely available through the BupaR package. The data includes 1050 cases, 15214 events, and 16 types of actions related to sepsis patient care. First, we used process mining to determine how long each patient was in the hospital. Using BupaR’s functions, we created several process model graphs. These process models depict the movement of patients at a hospital and provide duration data for each patent case. Second, we identified outlier data and created two dataset versions: one with and one without outliers. We then applied the following analysis methods: Linear Regression, Random Forest, K-Nearest Neighbors, Neural Networks, XGBoost, and lightGBM. We compared the model validations for the six machine learning models using the same data-splitting method. We found that the XGBoost model had the best prediction accuracy of 73.9 percent for cases with outliers, and 79 percent for cases without outliers. We also found that the lightGBM model had the lowest mean absolute error between prediction and actual duration in days with 3.66 days for the case with outliers, and 2.4 days for the case without outliers. These two models outperformed the other four models. This work will be enhanced in the future by exploring new prediction algorithms and comparing them with the results of this study.
Photometric Redshifts with the LSST. II. The Impact of Near-infrared and Near-ultraviolet Photometry
Accurate photometric redshift (photo-z) estimates are essential to the cosmological science goals of the Vera C. Rubin Observatory Legacy Survey of Space and Time (LSST). In this work, we use simulated photometry for mock galaxy catalogs to explore how LSST photo-z estimates can be improved by the addition of near-infrared (NIR) and/or ultraviolet (UV) photometry from the Euclid, Wide-Field InfrarRed Survey Telescope (WFIRST), and/or Cosmological Advanced Survey Telescope for Optical and ultraviolet Research (CASTOR) space telescopes. Generally, we find that deeper optical photometry can reduce the standard deviation of the photo-z estimates more than adding NIR or UV filters but that additional filters are the only way to significantly lower the fraction of galaxies with catastrophically under- or overestimated photo-z. For Euclid, we find that the addition of JH 5σ photometric detections can reduce the standard deviation for galaxies with z > 1 (z > 0.3) by ∼20% (∼10%), and the fraction of outliers by ∼40% (∼25%). For WFIRST, we show how the addition of deep YJHK photometry could reduce the standard deviation by ≳50% at z > 1.5 and drastically reduce the fraction of outliers to just ∼2% overall. For CASTOR, we find that the addition of its UV- and u-band photometry could reduce the standard deviation by ∼30% and the fraction of outliers by ∼50% for galaxies with z < 0.5. We also evaluate the photo-z results within sky areas that overlap with both the NIR and UV surveys and when spectroscopic training sets built from the surveys’ small-area deep fields are used.
Evolutionary Analyses of Gene Expression Divergence in Panicum hallii : Exploring Constitutive and Plastic Responses Using Reciprocal Transplants
Abstract The evolution of gene expression is thought to be an important mechanism of local adaptation and ecological speciation. Gene expression divergence occurs through the evolution of cis- polymorphisms and through more widespread effects driven by trans-regulatory factors. Here, we explore expression and sequence divergence in a large sample of Panicum hallii accessions encompassing the species range using a reciprocal transplantation experiment. We observed widespread genotype and transplant site drivers of expression divergence, with a limited number of genes exhibiting genotype-by-site interactions. We used a modified FST–QST outlier approach (QPC analysis) to detect local adaptation. We identified 514 genes with constitutive expression divergence above and beyond the levels expected under neutral processes. However, no plastic expression responses met our multiple testing correction as QPC outliers. Constitutive QPC outlier genes were involved in a number of developmental processes and responses to abiotic environments. Leveraging earlier expression quantitative trait loci results, we found a strong enrichment of expression divergence, including for QPC outliers, in genes previously identified with cis and cis–environment interactions but found no patterns related to trans-factors. Population genetic analyses detected elevated sequence divergence of promoters and coding sequence of constitutive expression outliers but little evidence for positive selection on these proteins. Our results are consistent with a hypothesis of cis-regulatory divergence as a primary driver of expression divergence in P. hallii.
Bridging the Gap Between LLMs and LNS with Dynamic Data Format and Architecture Codesign
Deep neural networks (DNNs) have achieved tremendous success in the past few years. However, their training and inference demand exceptional computational and memory resources. Quantization has been shown as an effective approach to mitigate the cost, with the mainstream data types reduced from FP32 to FP16/BF16 and recently FP8 in the latest NVIDIA H100 GPUs. With increasingly aggressive quantization, however, the conventional floating-point formats suffer from limited precision in representing numbers around zero. Recently, NVIDIA demonstrated the potential of using a Logarithmic Number System (LNS) for the next generation of tensor cores. While LNS mitigates the hurdles in representing small numbers, in this work we observed a mismatch between LNS and the emerging Large Language Models (LLM), where LLM exhibits significant outliers when directly adopting the LNS format. In this paper, we present a data-format/architecture codesign to bright this gap. On the format side, we propose a dynamic LNS format to flexibly represent outliers at a higher precision, by exploiting asymmetry in the LNS representation and identifying outliers through a per-vector basis. On the architecture side, for demonstration, we realize the dynamic LNS format in a systolic array, which can handle the irregularity of the outliers at runtime. We implement our approach on an Alveo U280 FPGA as a prototype. Experimental results show that our design can effectively handle the outliers and resolve the mismatch between LNS and LLM, contributing to an accuracy improvement of 15.4% and 16% over the floating-point and the original LNS baselines, using four state-of-the-art LLM models. Our observation and design lay a solid foundation for the large-scale adoption of the LNS format in the next-generation deep learning hardware.
Where’s Swimmy?: Mining unique color features buried in galaxies by deep anomaly detection using Subaru Hyper Suprime-Cam data
Abstract We present the Swimmy (Subaru WIde-field Machine-learning anoMalY) survey program, a deep-learning-based search for unique sources using multicolored (grizy) imaging data from the Hyper Suprime-Cam Subaru Strategic Program (HSC-SSP). This program aims to detect unexpected, novel, and rare populations and phenomena, by utilizing the deep imaging data acquired from the wide-field coverage of the HSC-SSP. This article, as the first paper in the Swimmy series, describes an anomaly detection technique to select unique populations as “outliers” from the data-set. The model was tested with known extreme emission-line galaxies (XELGs) and quasars, which consequently confirmed that the proposed method successfully selected $\sim\!\! 60\%$–$70\%$ of the quasars and $60\%$ of the XELGs without labeled training data. In reference to the spectral information of local galaxies at z = 0.05–0.2 obtained from the Sloan Digital Sky Survey, we investigated the physical properties of the selected anomalies and compared them based on the significance of their outlier values. The results revealed that XELGs constitute notable fractions of the most anomalous galaxies, and certain galaxies manifest unique morphological features. In summary, deep anomaly detection is an effective tool that can search rare objects, and, ultimately, unknown unknowns with large data-sets. Further development of the proposed model and selection process can promote the practical applications required to achieve specific scientific goals.
Understanding Hailstone Temporal Variability and Contributing Factors over the U.S. Southern Great Plains
Abstract Hailstones are a natural hazard that pose a significant threat to property and are responsible for significant economic losses each year in the United States. Detailed understanding of their characteristics is essential to mitigate their impact. Identifying the dynamic and physical factors contributing to hail formation and hailstone sizes is of great importance to weather and climate prediction and policymakers. In this study, we have analyzed the temporal and spatial variabilities of severe hail occurrences over the U.S. southern Great Plains (SGP) states from 2004 to 2016 using two hail datasets: hail reports from the Storm Prediction Center and the newly developed radar-retrieved maximum expected size of hail (MESH). It is found that severe and significant severe hail occurrences have a considerable year-to-year temporal variability in the SGP region. The interannual variabilities have a strong correspondence with sea surface temperature anomalies over the northern Gulf of Mexico and there is no outlier. The year 2016 is identified as an outlier for the correlations with both El Niño–Southern Oscillation (ENSO) and aerosol loading. The correlations with ENSO and aerosol loading are not statistically robust to inclusion of the outlier 2016. Statistical analysis without the outlier 2016 shows that 1) aerosols that may be mainly from northern Mexico have the largest correlation with hail interannual variability among the three factors and 2) meteorological covariation does not significantly contribute to the high correlation. These analyses warrant further investigations of aerosol impacts on hail occurrence.
Machine learning enhanced predictions of ICRF heating: Overcoming numerical limitations via data curation
In this work, we present the development of robust surrogate models for Ion Cyclotron Range of Frequencies (ICRF) and High-Harmonic Fast Wave (HHFW) heating predictions in fusion plasmas. Building upon our previous efforts to achieve real-time capable models, we identify the cause of the outliers found using TORIC in certain HHFW heating scenarios. The outliers are observed to be spurious ion Bernstein wave (IBW)-like modes caused by a wavelength control algorithm designed to address challenging scenarios with high perpendicular wavenumbers. The effect arises from the modulation in the perpendicular susceptibility, which can induce sign reversal and IBW-like propagation for scenarios featuring normalized ion Larmor radius λ i ≫ 1. We use TORIC with this algorithm disabled to generate a novel HHFW-NSTX database that is free of outliers. Surrogate models trained on this database, including Random Forest Regressor (RFR), Multi-Layer Perceptrons, and Gaussian Process Regressors (GPR), demonstrate the ability to accurately predict HHFW heating profiles, with regression scores of R 2 ∈[0.93−0.99]. Additionally we demonstrate that it is possible to generalize predictions beyond training data by the use of both RFR and GPR models, enabling the prediction of scenarios previously limited to the original model. GPR models also provide uncertainty quantification, offering insights into model confidence. This work introduces a comprehensive Verification, Validation, and Uncertainty Quantification methodology for surrogate modeling, applicable not only to ICRF heating but also to other RF heating challenges and fusion physics problems. Beyond accelerated inference, these models show effective extrapolation capabilities, providing an alternative for addressing numerical challenges.
Maximizing efficiency of dataset compression for machine learning potentials with information theory
Machine learning interatomic potentials (MLIPs) balance high accuracy and lower costs compared to density functional theory calculations, but their performance often depends on the size and diversity of training datasets. Large datasets improve model accuracy and generalization but are computationally expensive to produce and train on, while smaller datasets risk discarding rare but important atomic environments and compromising MLIP accuracy/reliability. Here, we develop an information-theoretical framework to quantify the efficiency of dataset compression methods and propose an algorithm that maximizes this efficiency. By framing atomistic dataset compression as an instance of the minimum set cover (MSC) problem over atom-centered environments, our method identifies the smallest subset of structures that contains as much information as possible from the original dataset while pruning redundant information. The approach is extensively demonstrated on the GAP-20 and TM23 datasets and validated on 64 varied datasets from the ColabFit repository. Across all cases, MSC consistently retains outliers, preserves dataset diversity, and reproduces the long-tail distributions of forces even at high compression rates, outperforming other subsampling methods. Furthermore, MLIPs trained on MSC-compressed datasets exhibit reduced error for out-of-distribution data even in low-data regimes. We explain these results using an outlier analysis and show that such quantitative conclusions could not be achieved with conventional dimensionality reduction methods. The algorithm is implemented in the open-source QUESTS package and can be used for several tasks in atomistic modeling, from data subsampling, outlier detection, and training improved MLIPs at a lower cost.
A study of the central stellar populations of galaxies in SDSS-IV MaNGA: identification of a subsample with unusually young and massive stars
ABSTRACT This paper describes a search for galaxy centres with clear indications of unusual stellar populations with an initial mass function flatter than Salpeter at high stellar masses. Out of a sample of 668 face-on galaxies with stellar masses in the range $10^{10}\!-\!10^{11} \, \mathrm{M}_{\odot }$, I identify 15 galaxies with young- to intermediate-age central stellar populations with unusual stellar population gradients in the inner regions of the galaxy. In these galaxies, the 4000-Å break is either flat or rising towards the centre of the galaxy, indicating that the central regions host evolved stars, but the H α equivalent width also rises steeply in the central regions. The ionization parameter [O iii]/[O ii] is typically low in these Galactic Centres, indicating that ionizing sources are stellar rather than AGN. Wolf–Rayet features characteristic of hot young stars are often found in the spectra and these also get progressively stronger at smaller galactocentric radii. These outliers are compared to a control sample of galaxies of similar mass with young inner stellar populations, but where the gradients in H α equivalent width and 4000-Å break follow each other more closely. The outliers exhibit central Wolf–Rayet red bump excesses much more frequently, they have higher central stellar and ionized gas metallicities, and they are also more frequently detected at 20-cm radio wavelengths. I highlight one outlier where the ionized gas is clearly being strongly perturbed and blown out either by massive stars after they explode as supernovae, or by energy injection from matter falling on to a black hole.
The Most Massive White Dwarfs in the Solar Neighborhood
We present an analysis of the most massive white dwarf candidates in the Montreal White Dwarf Database 100 pc sample. We identify 25 objects that would be more massive than 1.3M ⊙ if they had pure H atmospheres and CO cores, including two outliers with unusually high photometric mass estimates near the Chandrasekhar limit. We provide follow-up spectroscopy of these two white dwarfs and show that they are indeed significantly below this limit. We expand our model calculations for CO core white dwarfs up to M = 1.334 M ⊙ , which corresponds to the high-density limit of our equation-of-state tables, ρ = 109 g cm -3 . We find many objects close to this maximum mass of our CO core models. A significant fraction of ultramassive white dwarfs are predicted to form through binary mergers. Merger populations can reveal themselves through their kinematics, magnetism, or rapid rotation rates. We identify four outliers in transverse velocity, four likely magnetic white dwarfs (one of which is also an outlier in transverse velocity), and one with rapid rotation, indicating that at least 8 of the 25 ultramassive white dwarfs in our sample are likely merger products.
Fast and efficient identification of anomalous galaxy spectra with neural density estimation
ABSTRACT Current large-scale astrophysical experiments produce unprecedented amounts of rich and diverse data. This creates a growing need for fast and flexible automated data inspection methods. Deep learning algorithms can capture and pick up subtle variations in rich data sets and are fast to apply once trained. Here, we study the applicability of an unsupervised and probabilistic deep learning framework, the probabilistic auto-encoder, to the detection of peculiar objects in galaxy spectra from the SDSS survey. Different to supervised algorithms, this algorithm is not trained to detect a specific feature or type of anomaly, instead it learns the complex and diverse distribution of galaxy spectra from training data and identifies outliers with respect to the learned distribution. We find that the algorithm assigns consistently lower probabilities (higher anomaly score) to spectra that exhibit unusual features. For example, the majority of outliers among quiescent galaxies are E+A galaxies, whose spectra combine features from old and young stellar population. Other identified outliers include LINERs, supernovae, and overlapping objects. Conditional modelling further allows us to incorporate additional information. Namely, we evaluate the probability of an object being anomalous given a certain spectral class, but other information such as metrics of data quality or estimated redshift could be incorporated as well. We make our code publicly available.
Defining blood hematology reference values in female pig-tailed macaques ( Macaca nemestrina ) using the Isolation Forest algorithm
Background: Pig-tailed macaques (PTMs) are commonly used as preclinical models to assess antiretroviral drugs for HIV prevention research. Drug toxicities and disease pathologies are often preceded by changes in blood hematology. To better assess the safety profile of pharmaceuticals, we defined normal ranges of hematological values in PTMs using an Isolation Forest (iForest) algorithm. Methods: Eighteen female PTMs were evaluated. Blood was collected 1–24 times per animal for a total of 159 samples. Complete blood counts were performed, and iForest was used to analyze the hematology data to detect outliers. Results: Median, IQR, and ranges were calculated for 13 hematology parameters. From all samples, 22 outliers were detected. These outliers were excluded from the reference index. Conclusions: Using iForest, we defined a normal range for hematology parameters in female PTMs. This reference index can be a valuable tool for future studies evaluating drug toxicities in PTMs.