Information Extraction and Logical Inference for Derivative Classifier Assistance.
Abstract not provided.
SEARCH · Engineering Papers
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.
Abstract not provided.
Explore the source record for details and available documents.
Identification and automatic labeling of facilities in high-resolution satellite images is a challenging task as the current thematic classification schemes and the low-level image features are not good enough to capture complex objects and their spatial relationships. In this paper we present a novel algorithm framework for automated semantic labeling of large image collections. The framework consists of various segmentation, feature extraction, vector quantization, and Latent Dirichlet Allocation modules. Initial experimental results show promise as well as the challenges in semantic classification technology development for nuclear proliferation monitoring.
PURPOSE Surgical pathology reports are critical for cancer diagnosis and management. To accurately extract information about tumor characteristics from pathology reports in near real time, we explore the impact of using domain-specific transformer models that understand cancer pathology reports. METHODS We built a pathology transformer model, Path-BigBird, by using 2.7 million pathology reports from six SEER cancer registries. We then compare different variations of Path-BigBird with two less computationally intensive methods: Hierarchical Self-Attention Network (HiSAN) classification model and an offthe-shelf clinical transformer model (Clinical BigBird). We use five pathology information extraction tasks for evaluation: site, subsite, laterality, histology, and behavior. Model performance is evaluated by using macro and micro F 1 scores. RESULTS We found that Path-BigBird and Clinical BigBird outperformed the HiSAN in all tasks. Clinical BigBird performed better on the site and laterality tasks. Versions of the Path-BigBird model performed best on the two most difficult tasks: subsite (micro F 1 score of 72.53, macro F 1 score of 35.76) and histology (micro F 1 score of 80.96, macro F 1 score of 37.94). The largest performance gains over the HiSAN model were for histology, for which a Path-BigBird model increased the micro F 1 score by 1.44 points and the macro F 1 score by 3.55 points. Overall, the results suggest that a Path-BigBird model with a vocabulary derived from wellcurated and deidentified data is the best-performing model. CONCLUSION The Path-BigBird pathology transformer model improves automated information extraction from pathology reports. Although Path-BigBird outperforms Clinical BigBird and HiSAN, these less computationally expensive models still have utility when resources are constrained.
Constellation's overarching goal is the federation of information from resources within an extreme-scale scientific collaboration to enable the scalable discovery of data and new knowledge pathways. The resource fabric is comprised of petascale supercomputers and storage systems, users, jobs, datasets and lifecycle artifacts. For an extreme-scale supercomputing center, normal operations can generate hundreds of millions of data products and metadata entries describing the resource fabric. Constellation federates the information extracted from the resources using a custom, transformative science graph network; constructs rich metadata indexes and higher-order derived metadata from the extracted information; and conducts scalable graph analytics to unravel hidden data pathways. Our implementation and deployment for a production, supercomputing facility shows that the graph can scale to more than 750 million vertices, its domain agnostic indexing can answer interesting science queries, and its analytics can aid in structural, topological and temporal analysis to identify usage hotspots.
An overview is given of the current state of automatic image pattern recognition as applied to remote sensing of the earth's resources. The framework for the discussion is provided by four important aspects of the remote sensing problem: scene information content, characterization of scene information, information extraction methods, and the net value of extractable information. Outstanding problems are surveyed, as are the prospects for future developments. The effect of increasingly complex data bases and the rapidly evolving digital computer technology are highlighted.
The process of knowledge discovery and data mining is the process of information extraction from very large databases. Its importance is described along with several techniques and considerations for selecting the most appropriate technique for extracting information from a particular data set.
Missions to Mars already under study call for very long autonomous traverses (on the order of hundreds of meters per sol) during which there will be limited contact with Earth. The onboard Autonomous Science Investigation System (OASIS) is a technology for increasing science return during rover traverses by prioritizing science opportunities. By prioritizing data for downlink onboard, it is expected that the set of images selected for downlink by OASIS will have a consistently higher scientific interest then any set of equal number of images of the same traverse obtained using random or periodic sampling. Thus, OASIS can be used to increase the science returned from a long traverse. OASIS prioritizes data and identifies science opportunities by extracting information from image data and then using this information to prioritize the data and identify potential new science opportunities. Currently, the information extracted from the images is the location of the rocks in the image and several properties of these rocks such as size, albedo, texture and shape, In this paper we give an overview of the role of rock detection in the OASIS system and describe the functionality of the rock detector. We then present results on representative images
This paper presents a method for extracting information from SAR images using the wavelet transform. It is shown experimentally that the shape of the wavelet can be chosen in order to extract different features in the image.
An approach to extracting depth information out of a stereo pair of images is described. Special camera pair configurations are shown to be effective in reducing complexity. A simple algorithm which pairs objects in the two images and allows verification is described.
In current nuclear power plants (NPPs) a large amount of condition-based data is being generated and stored to assess and monitor component health and performance. The format of this data can be either numeric (e.g., pump vibration data) or textual (e.g., condition report which assess component health). While assessing component health from numeric data can be performed with a large variety of methods, the extraction of information from textual data still remains a challenge. Natural language processing (NLP) methods are starting to be deployed in current NPPs mainly to filter out incident reports (IRs) that are not safety related by employing supervised machine learning methods. However, these methods do not really provide the quantitative information that might be contained in IRs. This paper presents an approach to extract information from textual data (e.g., from IRs, maintenance reports) that is based on NLP data analytics methods coupled with model-based system engineer (MBSE) models. NLP methods are employed to perform syntactic and semantic analyses. Syntactic analysis analyzes the grammatical structure of a sentence; such analysis includes: part of speech (POS) tagging (i.e., identification of grammatic elements of each string - e.g., nouns, verbs), named entity recognition (i.e., identification of text entities - e.g., names, dates, events), and relation extraction (e.g., coreference resolution). On the other hand, semantic analysis is designed to analyze the logic structure of a sentence. Through a specific set of rules, our methods can identify whether a sentence contains health information of a component (e.g., degraded performance, anomaly behavior) or the causal relationship between two events (i.e., a cause-effect pair). An innovative element of our approach is that semantic analysis relies on MBSE models to identify links between textual elements. MBSE are diagrams designed to represent system and component dependencies (from both a form and functional point of view). In our approach, MBSE models emulate system engineer knowledge about component/system architecture. This paper presents in detail how the integration of NLP methods and MBSE models is performed. Few analysis examples focusing on centrifugal pumps are presented.
Methods and systems for rapid automatic keyword extraction for information retrieval and analysis. Embodiments can include parsing words in an individual document by delimiters, stop words, or both in order to identify candidate keywords. Word scores for each word within the candidate keywords are then calculated based on a function of co-occurrence degree, co-occurrence frequency, or both. Based on a function of the word scores for words within the candidate keyword, a keyword score is calculated for each of the candidate keywords. A portion of the candidate keywords are then extracted as keywords based, at least in part, on the candidate keywords having the highest keyword scores.
Additional information extracted from eddy-current proximeter. Positioned over rotating shaft, measures both displacement of and torsion in shaft. Torque applied to shaft calculable from output of proximeter. Possible to extract torsion information from existing tape-recorded proximeter data.
This article evaluates three techniques developed to extract emissivity information from multispectral thermal infrared data. The techniques are the assumed Channel 6 emittance model, thermal log residuals, and alpha residuals. These techniques were applied to calibrated, atmospherically corrected thermal infrared multispectral scanner (TIMS) data acquired over Cuprite, Nevada in September 1990. Results indicate that the two new techniques (thermal log residuals and alpha residuals) provide two distinct advantages over the assumed Channel 6 emittance model. First, they permit emissivity information to be derived from all six TIMS channels. The assumed Channel 6 emittance model only permits emissivity values to be derived from five of the six TIMS channels. Second, both techniques are less susceptible to noise than the assumed Channel 6 emittance model. The disadvantage of both techniques is that laboratory data must be converted to thermal log residuals or alpha residuals to facilitate comparison with similarly processed image data. An additional advantage of the alpha residual technique is that the processed data are scene-independent unlike those obtained with the other techniques.
This SAND report summarizes the activities and outcomes of the Network and Ensemble Enabled Entity Extraction in Information Text (NEEEEIT) LDRD project, which addressed improving the accuracy of conditional random fields for named entity recognition through the use of ensemble methods.
We investigate the possibility of performing quantum tomography on a single qubit with generalized partial measurements and the technique of measurement reversal. Using concepts from statistical decision theory, we prove that, somewhat surprisingly, no information can be obtained using this scheme. It is shown that, irrespective of the measurement technique used, extraction of information from single quanta is at odds with other general principles of quantum physics.
The goal of this project was to test the applicability of information theoretic learning (feasibility study) to develop new brain computer interfaces (BCI). The difficulty to BCI comes from several aspects: (1) the effective data collection of signals related to cognition; (2) the preprocessing of these signals to extract the relevant information; (3) the pattern recognition methodology to detect reliably the signals related to cognitive states. We only addressed the two last aspects in this research. We started by evaluating an information theoretic measure of distance (Bhattacharyya distance) for BCI performance with good predictive results. We also compared several features to detect the presence of event related desynchronization (ERD) and synchronization (ERS), and concluded that at least for now the bandpass filtering is the best compromise between simplicity and performance. Finally, we implemented several classifiers for temporal - pattern recognition. We found out that the performance of temporal classifiers is superior to static classifiers but not by much. We conclude by stating that the future of BCI should be found in alternate approaches to sense, collect and process the signals created by populations of neurons. Towards this goal, cross-disciplinary teams of neuroscientists and engineers should be funded to approach BCIs from a much more principled view point.
Gravitational lensing of distant galaxies can be exploited to infer the convergence field as a function of angular position on the sky. The statistics of this field, much like that of the cosmic microwave background (CMB), can be studied to extract information about fundamental parameters in cosmology, most notably the dark energy in the universe. Unlike the CMB, the distribution of matter in the universe which determines the convergence field is highly non-Gaussian, reflecting the nonlinear processes that accompanied structure formation. Much of the cosmic information contained in the initial field is therefore unavailable to the standard power spectrum measurements. Here we propose a method for re-capturing cosmic information by using the power spectrum of a simple function of the observed (nonlinear) convergence field. We adapt the approach of Neyrinck et al. to lensing by using a modified logarithmic transform of the convergence field. The Fourier transform of the log-transformed field has modes that are nearly uncorrelated, which allows for additional cosmological information to be extracted from small-scale modes.