Search

In the field of geomatics, artificial intelligence (AI) and especially machine learning (ML) are rapidly transforming the field of geomatics with respect to collecting, managing, and analyzing spatial data. Feature selection as a building block in ML is crucial because it directly impacts the performance and predictive power of a model by selecting the most critical variables and eliminating the redundant and irrelevant ones. Random forests have now been used for decades and allow for building models with high accuracy. However, finding the most expressive features from the dataset by selecting the most important features within random forests is still a challenging question. The often-used internal Gini importances of random forests are based on the amount of training examples that are divided by a feature but fail to acknowledge the magnitude of change in the target variable, leading to suboptimal selections. Shapley values are an established and unified framework for feature attribution, i.e., specifying how much each feature in a trained ML model contributes to the predictions for a given instance. Previous studies highlight the effectiveness of Shapley values for feature selection in real-world applications, while other research emphasizes certain theoretical limitations. This study provides an application-driven discussion of Shapley values for feature selection by first proposing four necessary conditions for a successful feature selection with Shapley values that are extracted from a multitude of critical research in the field. Given these valuable conditions, Shapley value feature selection is nevertheless a model averaging procedure by definition, where unimportant features can alter the final selection. Therefore, we additionally present Conditional Feature Selection (CFS) as a novel algorithm for performing feature selection that mitigates this problem and use it to evaluate the impact of model averaging in several real-world examples, covering the use of ML in geomatics. The results of this study show Shapley values as a good measure for feature selection when compared with Gini feature importances on four real-world examples, improving the RMSE by 5% when averaged over selections of all possible subset sizes. An even better selection can be achieved by CFS, improving on the Gini selection by approximately 7.5% in terms of RMSE. For random forests, Shapley value calculation can be performed in polynomial time, offering an advantage over the exponential runtime of CFS, building a trade-off to the lost accuracy in feature selection due to model averaging. [ABSTRACT FROM AUTHOR]

Land Surface Temperature (LST) is an important resource for a variety of tasks. The data are mostly free of charge and combine high spatial and temporal resolution with reliable data collection over a historical timeframe. When remote sensing is used to provide LST data, such as the MODA11 product using information from the MODIS sensors attached to NASA satellites, data acquisition can be hindered by clouds or cloud shadows, occluding the sensors' view on different areas of the world. This makes it difficult to take full advantage of the high resolution of the data. A common solution to interpolating LST data is statistical interpolation methods, such as fitting polynomials or thin plate spine interpolation. These methods have difficulties in incorporating additional knowledge about the research area and learning local dependencies that can help with the interpolation process. We propose a novel approach to interpolating remote sensing LST data in a fixed research area considering local ground-site air temperature measurements. The two-step approach consists of learning the LST from air temperature measurements, where the ground-site weather stations are located, and interpolating the remaining missing values with partial convolutions within a U-Net deep learning architecture. Our approach improves the interpolation of LST for our research area by 44% in terms of RMSE, when compared to state-of-the-art statistical methods. Due to the use of air temperature, we can provide coverage of 100%, even when no valid LST measurements were available. The resulting gapless coverage of high resolution LST data will help unlock the full potential of remote sensing LST data. [ABSTRACT FROM AUTHOR]

Remote sensing data represent one of the most important sources for automized yield prediction. High temporal and spatial resolution, historical record availability, reliability, and low cost are key factors in predicting yields around the world. Yield prediction as a machine learning task is challenging, as reliable ground truth data are difficult to obtain, especially since new data points can only be acquired once a year during harvest. Factors that influence annual yields are plentiful, and data acquisition can be expensive, as crop-related data often need to be captured by experts or specialized sensors. A solution to both problems can be provided by deep transfer learning based on remote sensing data. Satellite images are free of charge, and transfer learning allows recognition of yield-related patterns within countries where data are plentiful and transfers the knowledge to other domains, thus limiting the number of ground truth observations needed. Within this study, we examine the use of transfer learning for yield prediction, where the data preprocessing towards histograms is unique. We present a deep transfer learning framework for yield prediction and demonstrate its successful application to transfer knowledge gained from US soybean yield prediction to soybean yield prediction within Argentina. We perform a temporal alignment of the two domains and improve transfer learning by applying several transfer learning techniques, such as L 2 -SP, BSS, and layer freezing, to overcome catastrophic forgetting and negative transfer problems. Lastly, we exploit spatio-temporal patterns within the data by applying a Gaussian process. We are able to improve the performance of soybean yield prediction in Argentina by a total of 19% in terms of RMSE and 39% in terms of R 2 compared to predictions without transfer learning and Gaussian processes. This proof of concept for advanced transfer learning techniques for yield prediction and remote sensing data in the form of histograms can enable successful yield prediction, especially in emerging and developing countries, where reliable data are usually limited. [ABSTRACT FROM AUTHOR]

Spondylolysis is underdiagnosed and often missed in non-musculoskeletal abdominal CT imaging. Our aim was to assess the inter-reader agreement and diagnostic performance of a novel "Darth Vader sign" for the detection of spondylolysis in routine axial images. We performed a retrospective search in the institutional report archives through keyword strings for lumbar spondylolysis and spondylolisthesis. Abdominal CTs from 53 spondylolysis cases (41% female) and from controls (n = 6) without spine abnormalities were identified. A total of 139 single axial slices covering the lumbar spine (86 normal images, 40 with spondylolysis, 13 with degenerative spondylolisthesis without spondylolysis) were exported. Two radiology residents rated all images for the presence or absence of the "Darth Vader sign". The diagnostic accuracy for both readers, as well as the inter-reader agreement, was calculated. The "Darth Vader sign" showed an inter-reader agreement of 0.77. Using the "Darth Vader sign", spondylolysis was detected with a sensitivity and specificity of 65.0–88.2% and 96.2–99.0%, respectively. The "Darth Vader sign" shows excellent diagnostic performance at a substantial inter-reader agreement for the detection of spondylolysis. Using the "Darth Vader sign" in the CT reading routine may be an easy yet effective tool to improve the detection rate of spondylolysis in non-musculoskeletal cases and hence improve patient care. [ABSTRACT FROM AUTHOR]

X-linked deafness (DFNX) is estimated to account for up to 2% of cases of hereditary hearing loss and occurs in both syndromic and non-syndromic forms. POU3F4 is the gene most commonly associated with X-linked deafness (DFNX2, DFN3) and accounts for about 50% of the cases of X-linked non-syndromic hearing loss. This gene codes for a transcription factor of the POU family that plays a major role in the development of the middle and inner ear. The clinical features of POU3F4-related hearing loss include a pathognomonic malformation of the inner ear defined as incomplete partition of the cochlea type 3 (IP-III). Often, a perilymphatic gusher is observed upon stapedectomy during surgery, possibly as a consequence of an incomplete separation of the cochlea from the internal auditory canal. Here we present an overview of the pathogenic gene variants of POU3F4 reported in the literature and discuss the associated clinical features, including hearing loss combined with additional phenotypes such as cognitive and motor developmental delays. Research on the transcriptional targets of POU3F4 in the ear and brain is in its early stages and is expected to greatly advance our understanding of the pathophysiology of POU3F4-linked hearing loss. [ABSTRACT FROM AUTHOR]

Hypereutectic Al-high Si alloys are of immense interest for applications in the automotive, space or electronic industries, due their low weight, low thermal expansion, and excellent mechanical and tribological properties. Additionally, their production by laser powder bed fusion (LPBF) technology provides high flexibility in geometrical design and alloy composition. Since, most of the alloy properties could be improved by increasing the Si content, there is much interest in discovering the maximum that could be realized in LBPF Al-high Si alloys, without the appearance of any material failure. For this reason, in this work the production of Al-high Si alloys with extremely high silicon content of up to 70 wt.% was fundamentally investigated with respect to microstructure and mechanical properties. Highly dense (99.3%) and crack-free AlSi50 samples (5 × 5 × 5 mm3), with excellent hardness (225 HV5) and compressive strength (742 MPa), were successfully produced. Further, for the first time, AlSi70 LBPF samples of high density (98.8%) without cracks were demonstrated, using moderate scanning velocities. Simultaneously, the hardness and the compressive strength in the AlSi70 alloys were significantly improved to 350 HV5 and 935 MPa, as a result of the formation of a continuous Si network in the microstructure of the alloy. With respect to the powder source, it was found that the application of powder blends resulted in similar alloy properties as if pre-alloyed powders were used, enabling higher flexibility in prospective application-oriented alloy development. [ABSTRACT FROM AUTHOR]

Background: Interstitial lung disease (ILD) is a frequent complication of systemic sclerosis (SSc), and its early detection and treatment may prevent deterioration of lung function. Different vendors have recently made larger image matrices available as a post-processing option for computed tomography (CT), which could facilitate the diagnosis of SSc-ILD. Therefore, the objective of this study was to assess the effect of matrix size on lung image quality in patients with SSc by comparing a 1024-pixel matrix to a standard 512-pixel matrix and applying different reconstruction kernels. Methods: Lung scans of 50 patients (mean age 54 years, range 23–85 years) with SSc were reconstructed with these two different matrix sizes, after determining the most appropriate kernel in a first step. Four observers scored the images on a five-point Likert scale regarding image quality and detectability of clinically relevant findings. Results: Among the eight tested kernels, the Br59-kernel (sharp) reached the highest score (19.48 ± 3.99), although differences did not reach statistical significance. The 1024-pixel matrix scored higher than the 512-pixel matrix HRCT overall (p = 0.01) and in the subcategories sharpness (p < 0.01), depiction of bronchiole (p < 0.01) and overall image impression (p < 0.01), and lower for the detection of ground-glass opacities (GGO) (p = 0.04). No significant differences were found for detection of extent of reticulations/bronchiectasis/fibrosis (p = 0.50) and image noise (p = 0.09). Conclusions: Our results show that with the use of a sharp kernel, the 1024-pixel matrix HRCT, provides a slightly better subjective image quality in terms of assessing interstitial lung changes, whereby GGO are more visible on the 512-pixel matrix. However, it remains to be answered to what extent this is related to the improved representation of the smallest structures. [ABSTRACT FROM AUTHOR]

In this paper, we compare the transmission of a conventional monetary policy shock with that of an unexpected decrease in the term spread, which mirrors quantitative easing. Employing a time-varying vector autoregression with stochastic volatility, our results are two-fold: First, the spread shock works mainly through a boost to consumer wealth growth, while a conventional monetary policy shock affects real output growth via a broad credit/bank lending channel. Second, both shocks exhibit a distinct pattern over our sample period. More specifically, we find small output effects of a conventional monetary policy shock during the period of the global financial crisis and stronger effects in its aftermath. This might imply that when the central bank has left the policy rate unaltered for an extended period of time, a policy surprise might boost output particularly strongly. By contrast, the spread shock has affected output growth most strongly during the period of the global financial crisis and less so thereafter. This might point to diminishing effects of large-scale asset purchase programs. [ABSTRACT FROM AUTHOR]

Within the scope of the presented work the processing of AISI H11 (1.2343 or X37CrMoV5-1) tool steel powder modified by adding carbon black nanoparticles in varying concentrations by means of Laser Metal Deposition (LMD) is extensively investigated. On the basis of single weld track experiments, multi-layered cuboid-shaped samples made out of pure AISI H11 tool steel powder as well as modified tool steel powder mixtures were manufactured by applying various process parameters. The main scientific aim of the investigations was to achieve a basic understanding of the influence of the added carbon black nanoparticles on the resulting sample properties. For that purpose, the generated specimens were first analyzed with respect to relative density, inner defects, microstructure, Vickers hardness and chemical composition. Subsequently, the mechanical properties of post-heat-treated specimens were investigated, with the focus on the yield strength (Y0.2%), by means of compression tests. We prove that by adding carbon black nanoparticles to the initial AISI H11 powder, the formation of martensitic and bainitic phases, as well as the precipitation of carbides at the grain boundaries, are enhanced. As a result, a significant increase of Vickers hardness and of the compression yield strength by up to 11% can be achieved in comparison to samples made out of the unmodified AISI H11 powder. Furthermore, it can be fundamentally demonstrated that the fabrication of parts with layer-specific variable hardness can be realized by the controlled changing of the powder mixtures used during the layer-by-layer manufacturing approach. [ABSTRACT FROM AUTHOR]

Simple Summary: We investigated the feasibility and immunogenicity of an autologous Dendritic Cell (DC) vaccine pulsed with peptide neoantigens in combination with a standard of care regimen. The vaccine program took place in a serous endometrial mismatch repair (MMR) deficiency setting. We demonstrate for the first time the feasibility of producing a peptide DC vaccine in endometrial carcinoma. The safety and immunogenicity of this personalized vaccine was demonstrated by the detection of polyfunctional and durable T-cell responses. Endometrial cancer (EC) is a common gynecological malignancy and the fourth most common malignancy in European and North American women. Amongst EC, the advanced serous, p53-mutated, and pMMR subtypes have the highest risk of relapse despite optimal standard of care therapy. At present, there is no standard of care maintenance treatment to prevent relapse among these high-risk patients. Vaccines are a form of immunotherapy that can potentially increase the immunogenicity of pMMR, serous, and p53-mutated tumors to render them responsive to check point inhibitor-based immunotherapy. We demonstrate, for the first time, the feasibility of generating a personalized dendritic cell vaccine pulsed with peptide neoantigens in a patient with pMMR, p53-mutated, and serous endometrial adenocarcinoma (SEC). The personalized vaccine was administered in combination with systemic chemotherapy to treat an inoperable metastatic recurrence. This treatment association demonstrated the safety and immunogenicity of the personalized dendritic cell vaccine. Interestingly, a complete oncological response was obtained with respect to both radiological assessment and the tumor marker CA-125. [ABSTRACT FROM AUTHOR]

In recent years, the application field of laser powder bed fusion of metals and polymers extends through an increasing variability of powder compositions in the market. New powder formulations such as nanoparticle (NP) additivated powder feedstocks are available today. Interestingly, they behave differently along with the entire laser powder bed fusion (PBF-LB) process chain, from flowability over absorbance and microstructure formation to processability and final part properties. Recent studies show that supporting NPs on metal and polymer powder feedstocks enhances processability, avoids crack formation, refines grain size, increases functionality, and improves as-built part properties. Although several inter-laboratory studies (ILSs) on metal and polymer PBF-LB exist, they mainly focus on mechanical properties and primarily ignore nano-additivated feedstocks or standardized assessment of powder feedstock properties. However, those studies must obtain reliable data to validate each property metric's repeatability and reproducibility limits related to the PBF-LB process chain. We herein propose the design of a large-scale ILS to quantify the effect of nanoparticle additivation on powder characteristics, process behavior, microstructure, and part properties in PBF-LB. Besides the work and sample flow to organize the ILS, the test methods to measure the NP-additivated metal and polymer powder feedstock properties and resulting part properties are defined. A research data management (RDM) plan is designed to extract scientific results from the vast amount of material, process, and part data. The RDM focuses not only on the repeatability and reproducibility of a metric but also on the FAIR principle to include findable, accessible, interoperable, and reusable data/meta-data in additive manufacturing. The proposed ILS design gives access to principal component analysis (PCA) to compute the correlations between the material–process–microstructure–part properties. [ABSTRACT FROM AUTHOR]

High entropy or multi principal element alloys are a promising and relatively young concept for designing alloys. The idea of creating alloys without a single main alloying element opens up a wide space for possible new alloy compositions. High entropy alloys based on refractory metals such as W, Mo, Ta or Nb are of interest for future high temperature applications e.g., in the aerospace or chemical industry. However, producing refractory metal high entropy alloys by conventional metallurgical methods remains challenging. For this reason, the feasibility of laser-based additive manufacturing of the refractory metal high entropy alloy W20Mo20Ta20Nb20V20 by laser powder bed fusion (PBF-LB/M) is investigated in the present work. In-situ alloy formation from mixtures of easily available elemental powders is employed to avoid an expensive atomization of pre-alloyed powder. It is shown that PBF-LB/M of W20Mo20Ta20Nb20V20 is in general possible and that a complete fusion of the powder mixture without a significant number of undissolved particles is achievable by in-situ alloy formation during PBF-LB/M when selecting favorable process parameter combinations. The relative density of the samples with a dimension of 6 × 6 × 6 mm3 reaches, in dependence of the PBF-LB/M parameter set, 99.8%. Electron backscatter diffraction (EBSD) and transmission electron microscopy (TEM) measurements confirm the presence of a single bcc-phase. Scanning electron microscopy (SEM) images show a dendritic and/or cellular microstructure that can, to some extent, be controlled by the PBF-LB/M parameters. [ABSTRACT FROM AUTHOR]

In-situ alloy formation by Laser Powder Bed Fusion (PBF-LB/M) from mixtures of easily available elemental powders is an appealing approach for developing and qualifying new alloys for laser based additive manufacturing of metals. However, especially when dealing with high-melting elements, like W, Ta, Mo, or Nb, it is difficult to achieve a homogeneous element distribution and a complete fusion of the powder particles. The aim of this work was to understand the effects of the PBF-LB/M process parameters (laser power, scan speed, laser spot diameter) and three different single- and double-exposure strategies on the fusion of high-melting W, Ta, Mo, and Nb particles in a Ti-matrix. For this purpose, 220 samples with 10 vol.% of the high-melting particle fraction were prepared and analyzed by optical light microscopy and automated image processing, as well as by scanning electron microscopy (SEM). The results are discussed in the context of current research on the process dynamics of PBF-LB/M. Based on that process strategies to support a complete fusion of high-melting particles during in-situ alloy formation are derived. It is shown that the number of unmolten particles can be at least decreased by a factor of ten compared to the most unfavorable parameter combination. For the lower melting elements, Nb and Mo, a complete fusion without any remaining particles visible in the microsections was achieved for certain parameter combinations. The results prove the feasibility of in-situ alloy formation with high-melting alloying elements, but they also demonstrate the necessity to adjust the PBF-LB/M process strategy to achieve a complete dissolution of the alloying elements. [ABSTRACT FROM AUTHOR]

The development of sensitive biosensors, such as gallium nitride (GaN)-based quantum wells, transistors, etc., often makes it necessary to functionalize GaN surfaces with small molecules or even biomolecules, such as proteins. As a first step in surface functionalization, we have investigated silane adsorption, as well as the formation of very thin silane layers. In the next step, the immobilization of the tetrameric protein streptavidin (as well as the attachment of chemically modified iron transport protein ferritin (ferritin-biotin-rhodamine complex)) was realized on these films. The degree of functionalization of the GaN surfaces was determined by fluorescence measurements with fluorescent-labeled proteins; silane film thickness and surface roughness were estimated, and also other surface sensitive techniques were applied. The formation of a monolayer consisting of adsorbed organosilanes was accomplished on Mg-doped GaN surfaces, and also functionalization with proteins was achieved. We found that very high Mg doping reduced the amount of surface functionalized proteins. Most likely, this finding was a consequence of the lower concentration of ionizable Mg atoms in highly Mg-doped layers as a consequence of self-compensation effects. In summary, we could demonstrate the necessity of Mg doping for achieving reasonable bio-functionalization of GaN surfaces. [ABSTRACT FROM AUTHOR]

This study aims to investigate the physical and chemical characterization of six fly ash samples obtained from different municipal solid waste incinerators (MSWIs), namely grate furnaces, rotary kiln, and fluidized bed reactor, to determine their potential for CO2 and thermochemical energy storage (TCES). Representative samples were characterized via simultaneous thermal analysis (STA) in different atmospheres, i.e., N2, air, H2O, CO2, and H2O/CO2, to identify fly ash samples that can meet the minimum requirements, i.e., charging, discharging, and cycling stability, for its consideration as TCES and CO2-storage materials and to determine their energy contents. Furthermore, other techniques, such as inductively coupled plasma optical emission spectroscopy, X-ray fluorescence (XRF) spectrometry, X-ray diffraction (XRD), scanning electron microscopy, leachability tests, specific surface area measurement based on the Brunauer–Emmett–Teller method, and particle-size distribution measurement, were performed. XRF analysis showed that calcium oxide is one of the main components in fly ash, which is a potentially suitable component for TCES systems. XRD results revealed information regarding the crystal structure and phases of various elements, including that of Ca. The STA measurements showed that the samples can store thermal heat with energy contents of 50–394 kJ/kg (charging step). For one fly ash sample obtained from a grate furnace, the release of the stored thermal heat under the selected experimental conditions (discharging step) was demonstrated. The cycling stability tests were conducted thrice, and they were successful for the selected sample. One fly ash sample could store CO2 with a storage capacity of 27 kg CO2/ton based on results obtained under the selected experimental conditions in STA. Samples from rotary kiln and fluidized bed were heated up to 1150 °C in an N2 atmosphere, resulting in complete melting of samples in crucibles; however, other samples obtained from grate furnaces formed compacted powders after undergoing the same thermal treatment in STA. Samples from different grate furnaces showed similarities in their chemical and physical characterization. The leachability test according to the standard (EN 12457-4 (2002)) using water in a ratio of 10 L/S and showed that the leachate of heavy metals is below the maximum permissible values for nonhazardous materials (except for Pb), excluding the fly ash sample obtained using fluidized bed technology. The leachate contents of Cd and Mn in the fly ash samples obtained from the rotary kiln were higher than those in other samples. Characterization performed herein helped in determining the suitable fly ash samples that can be considered as potential CO2-storage and TCES materials. [ABSTRACT FROM AUTHOR]

The aim of this work is to investigate the β-Ti-phase-stabilizing effect of vanadium and iron added to Ti-6Al-4V powder by means of heterogeneous powder mixtures and in situ alloy-formation during laser powder bed fusion (L-PBF). The resulting microstructure was analyzed by metallographic methods, scanning electron microscopy (SEM), and electron backscatter diffraction (EBSD). The mechanical properties were characterized by compression tests, both prior to and after heat-treating. Energy dispersive X-ray spectroscopy showed a homogeneous element distribution, proving the feasibility of in situ alloying by LPBF. Due to the β-phase-stabilizing effect of V and Fe added to Ti-6Al-4V, instead of an α'-martensitic microstructure, an α/β-microstructure containing at least 63.8% β-phase develops. Depending on the post L-PBF heat-treatment, either an increased upsetting at failure (33.9%) compared to unmodified Ti-6Al-4V (28.8%), or an exceptional high compressive yield strength (1857 ± 35 MPa compared to 1100 MPa) were measured. The hardness of the in situ alloyed material ranges from 336 ± 7 HV0.5, in as-built condition, to 543 ± 13 HV0.5 after precipitation-hardening. Hence, the range of achievable mechanical properties in dependence of the post-L-PBF heat-treatment can be significantly expanded in comparison to unmodified Ti-6Al-4V, thus providing increased flexibility for additive manufacturing of titanium parts. [ABSTRACT FROM AUTHOR]