Search

A large training dataset with high-quality annotations is necessary for building an accurate and generalisable deep learning system, which can be difficult and expensive to prepare in medical applications. We present a novel deep-learning-based system, requiring no annotator but weak annotation from a diagnosis report, for accurate and generalisable performance in detecting multiple head disorders from CT scans, including ischaemia, haemorrhage, tumours, and skull fractures.Our system was developed on 104 597 head CT scans from the Chinese PLA General Hospital, with associated textual diagnosis reports. Without expert annotation, we used keyword matching on the reports to automatically generate disorder labels for each scan. The labels were inaccurate because of the unreliable annotator-free strategy and inexact because of scan-level annotation. We proposed RoLo, a novel weakly supervised learning algorithm, with a noise-tolerant mechanism and a multi-instance learning strategy to address these issues. RoLo was tested on retrospective (2357 scans from the Chinese PLA General Hospital), prospective (650 scans from the Chinese PLA General Hospital), cross-centre (1525 scans from the Brain Hospital of Hunan Province), cross-equipment (1484 scans from the Chinese PLA General Hospital), and cross-nation (CQ500 public dataset from India) test datasets. Four radiologists were tested on the prospective test dataset before and after viewing system recommendations to assess whether the system could improve diagnostic performance.The area under the receiver operating characteristic curve for detecting the four disorder types was 0·976 (95% CI 0·976-0·976) for retrospective, 0·975 (0·974-0·976) for prospective, 0·965 (0·964-0·966) for cross-centre, and 0·971 (0·971-0·972) for cross-equipment test datasets, and 0·964 (0·964-0·966) for CQ500 (with only haemorrhage and fracture). The system achieved similar performance to four radiologists and helped to improve sensitivity and specificity by 0·109 (95% CI 0·086-0·131) and 0·022 (0·017-0·026), respectively.Without expert annotated data, our system achieved accurate and generalisable performance for head disorder detection. The system improved the diagnostic performance of radiologists. Because of its accuracy and generalisability, our computer-aided diganostic system could be used in clinical practice to improve the accuracy and efficiency of radiologists in different hospitals.National Key RD Program of China, National Natural Science Foundation of China, and Beijing Natural Science Foundation.

A broadband, efficient and stable yellow (Ba,Sr)Si3Al3O4N5:Yb2+ phosphor shows promising applications in single-phosphor-converted warm w-LEDs.

In this paper, the extension method is applied to the assessment of seismic damage risk in mountain tunnels. Based on various considerations such as the geological conditions of the tunnel site, the state of the tunnel, the structural situation and the earthquake, the seismic damage risk assessment index for mountainous tunnels is proposed. The range of values of different assessment indexes under each grade is quantified. The seismic damage risk of the tunnel is divided into four classes, and assessed based on the extension theory. The specific assessment process is as follows: Firstly, five tunnels affected by the Wenchuan earthquake were selected to construct the matter element to be assessed, and construct the classical domain and nodal domain; then, the entropy weight method and analytical hierarchy process (AHP) were used to determine the weights of each index; finally, the correlation function values were calculated to obtain the seismic damage risk level of the tunnels. Among them, the seismic damage risk level of the Longchi tunnel and Longxi tunnel is “high risk”, and the seismic damage risk level of the Longdongzi tunnel, Youyi tunnel and Maanshi tunnel is “moderate risk”. The five tunnels are sorted according to the eigenvalues of the seismic damage risk level as follows: Longchi tunnel > Longxi tunnel > Longdongzi tunnel > Youyi tunnel > Maanshi tunnel. Combined with the seismic damage investigation of the Wenchuan earthquake, the actual seismic damage of tunnel is consistent with the assessment results, which verifies the accuracy and effectiveness of this method. The seismic damage risk assessment model for mountain tunnels proposed in this paper has certain reference value for the future seismic damage risk assessment of tunnels.

Metal-organic frameworks (MOFs) and mechanoluminescent (ML) materials have been considered as two types of promising materials that have their own application fields. It would be amazing to endow one material with the advantages of ML and MOFs, thus broadening their applications. However, there are quite few investigations on this topic, and the ML mechanism in ML-MOFs remains unclear. In this study, we proposed a strategy for developing ML-MOFs by doping lanthanide ions into the non-centrosymmetric SBD ([Sr(μ-BDC)(DMF)]∞) MOF, and successfully synthesized a series of lanthanide-doped MOFs Ln-SBD (Ln = Tb, Dy, Sm, Eu) and Tb1− x Eu x -SBD ( x = 0.2, 0.4, 0.6, 0.8) with multicolor ML. The lanthanide ions were uniformly distributed in the matrix of the SBD-MOF, and occupied the Sr site. The ML-MOFs exhibited intense multicolor ML emissions varying from green to yellow to red by changing the co-doping ratios and species of lanthanide ions. The similar ML and photoluminescence (PL) spectra indicated that the ML emission was assigned to the radiative transition from the excited states to the ground states of lanthanide ions. The radiative transition was induced by the electron bombardment process that originated from the piezoelectric effect of the non-centrosymmetric SBD host. In addition, a pioneering temperature sensing research based on ML was carried out, which is promising for realizing dual-functional detection of stress and temperature without excitation light sources. This study gives a unique insight for developing more versatile and interesting smart materials by combining the versatility of MOF with the ML emission, imparting additional values to both MOF and ML materials. Moreover, this study provides a general rule for selecting MOFs with an acentric structure as the host for ML materials.

Visual analytics for machine learning has recently evolved as one of the most exciting areas in the field of visualization. To better identify which research topics are promising and to learn how to apply relevant techniques in visual analytics, we systematically review 259 papers published in the last ten years together with representative works before 2010. We build a taxonomy, which includes three first-level categories: techniques before model building, techniques during modeling building, and techniques after model building. Each category is further characterized by representative analysis tasks, and each task is exemplified by a set of recent influential works. We also discuss and highlight research challenges and promising potential future research opportunities useful for visual analytics researchers.

Ensemble feature selection algorithms aggregate the results of multiple feature selection methods in order to select an effective subset of features. However, typically, ensemble algorithms treat each feature selection method equally and do not consider performance differences. Consequently, features selected by a relatively smaller number of methods may not be included. To address this problem, we propose an interactive feature selection method that can more effectively aggregate the results of multiple feature selection methods and iteratively improve the selected features by integrating expert knowledge. The proposed method includes a learning-from-crowds-based ensemble feature selection algorithm and a visual analysis system. The algorithm models the performance of multiple feature selection methods, calculates their reliabilities, and aggregates results. To integrate expert knowledge, the visual analysis system provides a set of ranking schemes to assist experts in understanding the results of an individual feature selection method and the roles played by the features in classification tasks. A numerical experiment conducted on four real-world datasets shows that the proposed algorithm can improve classification accuracy by 0.63%–2.85% compared to state-of-the-art ensemble feature selection algorithms. In addition, we conducted case studies on text and image data to demonstrate that the proposed visual analysis system can further improve classification accuracy by 0.28%–5.24%.

In this paper, we have designed long afterglow luminescent MgGeO3:Mn2+,Yb3+,Li+ (MGO) nanoparticles in the first (NIR-I) and second (NIR-II) biological windows. Yb3+ ions served not only as the trap center to enhance the NIR-I long afterglow emission of Mn2+ at 680 nm, but also as an emitting center to produce a NIR-II long afterglow emission at ∼1000 nm. Furthermore, we have found the addition of Li+ can greatly increase the NIR-II afterglow emission of Yb3+ and the optimal amount of Mn2+, Yb3+ and Li+ was found to be 0.1, 0.5 and 0.5 mol%, respectively. The MGO nanoparticles synthesized using sol–gel methods showed a uniform morphology with a diameter of 50–100 nm, which were suitable for applications in bioimaging. More importantly, we have found MGO nanoparticles can be effectively excited to produce long persistent NIR-I and II luminescence using soft X-rays, suggesting that low dosage soft X-rays can also serve as a more powerful and deep tissue excitation source to recharge MGO nanoparticles. Furthermore, the MGO nanoparticles can also be re-excited to produce photo-stimulated emission under the irradiation of 650 and 808 nm NIR lasers. The in vivo imaging results have shown that MGO nanoparticles modified with folic acid (FA) can effectively realize super long-term targeted in vivo imaging of inflammation with a high sensitivity via recharging using soft X-rays and NIR lasers, which can provide not only an accurate diagnosis of inflammation, but also long-term monitoring of possible changes in the focus of inflammation in real time.

Autophagy is a catabolic pathway involved in the regulation of bone homeostasis. We explore clinical correlation of autophagy-related key molecules to establish risk signature for predicting the prognosis, tumor microenvironment (TME), and immunotherapy response of osteosarcoma. Single cell RNA sequencing data from GSE162454 dataset distinguished malignant cells from normal cells in osteosarcoma. Autophagy-related genes (ARGs) were extracted from the established risk signature of the Molecular Signatures Database of Gene Set Enrichment Analysis (GSEA) by univariate Cox and least absolute shrinkage and selection operator (LASSO) Cox regression analysis. Overall survival (OS), TME score, abundance of infiltrating immune cells, and response to immune-checkpoint blockade (ICB) treatment in patients with different risks were compared based on risk score. Nine ARGs were identified and risk signature was constructed. In all osteosarcoma datasets, the OS was significantly longer in the high-risk patients than low-risk onset. Risk signature significantly stratified clinical outcomes, including OS, metastatic status, and survival status. Risk signature was an independent variable for predicting osteosarcoma OS and showed high accuracy. A nomogram based on risk signature and metastases was developed. The calibration curve confirmed the consistency in 1-year, 3-year, and 5-year predicted OS and the actual OS. The risk score was related to 6 kinds of T cells and macrophages, myeloid-derived suppressor cell, natural killer cell, immune score, and stromal score in TME. The risk signature helped in predicting patients’ response to anti-PD1/anti-PD-L1 treatment. The ARGs-led risk signature has important value for survival prediction, risk stratification, tumor microenvironment, and immune response evaluation of osteosarcoma.

Semi-supervised learning (SSL) provides a way to improve the performance of prediction models (e.g., classifier) via the usage of unlabeled samples. An effective and widely used method is to construct a graph that describes the relationship between labeled and unlabeled samples. Practical experience indicates that graph quality significantly affects the model performance. In this paper, we present a visual analysis method that interactively constructs a high-quality graph for better model performance. In particular, we propose an interactive graph construction method based on the large margin principle. We have developed a river visualization and a hybrid visualization that combines a scatterplot, a node-link diagram, and a bar chart, to convey the label propagation of graph-based SSL. Based on the understanding of the propagation, a user can select regions of interest to inspect and modify the graph. We conducted two case studies to showcase how our method facilitates the exploitation of labeled and unlabeled samples for improving model performance.

Fully automatic image colorization is extensively developed using a large number of examples. However, this kind of methods need further detailed revision if users require fine colorization for delicate objects. This paper proposes a novel scribble-based colorization scheme which automatically generates scribbles and propagates color information from representative pixels within homogeneous superpixels. The scribbles are placed in the regions with minimal entropy to ensure color propagation from representative pixels with high confidence. To avoid color bleeding and under colorization, colors are propagated under a robust similarity metric using a set of hierarchical features, where the low level features consist of quaternion Gabor phases and the high level features are extracted using VGG-19 pre-trained deep model. Experimental results on open dataset demonstrate that the proposed colorization scheme can achieve more consistent results with the users' purposes.

In order to effectively infer correct labels from noisy crowdsourced annotations, learning-from-crowds models have introduced expert validation. However, little research has been done on facilitating the validation procedure. In this paper, we propose an interactive method to assist experts in verifying uncertain instance labels and unreliable workers. Given the instance labels and worker reliability inferred from a learning-from-crowds model, candidate instances and workers are selected for expert validation. The influence of verified results is propagated to relevant instances and workers through the learning-from-crowds model. To facilitate the validation of annotations, we have developed a confusion visualization to indicate the confusing classes for further exploration, a constrained projection method to show the uncertain labels in context, and a scatter-plot-based visualization to illustrate worker reliability. The three visualizations are tightly integrated with the learning-from-crowds model to provide an iterative and progressive environment for data validation. Two case studies were conducted that demonstrate our approach offers an efficient method for validating and improving crowdsourced annotations.

Interactive Machine Learning (IML) is an iterative learning process that tightly couples a human with a machine learner, which is widely used by researchers and practitioners to effectively solve a wide variety of real-world application problems. Although recent years have witnessed the proliferation of IML in the field of visual analytics, most recent surveys either focus on a specific area of IML or aim to summarize a visualization field that is too generic for IML. In this paper, we systematically review the recent literature on IML and classify them into a task-oriented taxonomy built by us. We conclude the survey with a discussion of open challenges and research opportunities that we believe are inspiring for future work in IML.

Xu, H.; Chen, C.; Shao, Z.; Wang, Z.; Cai, L.; Li, Z., and Xu, C., 2021. Dynamic response of heavy-haul railway tunnels in fully weathered coastal red sandstone strata with base rock reinforcement: A numerical study. Journal of Coastal Research, 37(1), 191–202. Coconut Creek (Florida), ISSN 0749-0208.Coastal fully weathered red sandstone exhibits poor cementation, and tunnels can easily develop large deformations undercrossing such rock masses under long-term cyclic loading. This article studies dynamic response of the base rock and tunnel lining structures under the dynamic load of heavy-haul train when different base rock reinforcing methods (curtain grouting reinforcement and cement-soil compaction pile reinforcement) are adopted. A tunnel undercrossing fully weathered red sandstone strata is taken as the research object, from which the rock mechanical strength parameters are obtained for a numerical model. The dynamic response of the base rock significantly decreases with the increased depth, and the inverted arch and arch foot are the most unfavorable stress locations, to which more attention should be paid. Base rock reinforcements enhance the integrity of rock mass, and the acceleration and peak stress of the rock base and lining structure significantly reduce when reinforcement is exerted. The effect of curtain grouting proved to be better, which is indicated by the reduction of acceleration and displacement of tunnel lining as well as the low value of peak acceleration and stress of the base rock.

Stress sensing is the basis of human-machine interface, biomedical engineering, and mechanical structure detection systems. Stress sensing based on mechanoluminescence (ML) shows significant advantages of distributed detection and remote response to mechanical stimuli and is thus expected to be a key technology of next-generation tactile sensors and stress recorders. However, the instantaneous photon emission in ML materials generally requires real-time recording with a photodetector, thus limiting their application fields to real-time stress sensing. In this paper, we report a force-induced charge carrier storage (FICS) effect in deep-trap ML materials, which enables storage of the applied mechanical energy in deep traps and then release of the stored energy as photon emission under thermal stimulation. The FICS effect was confirmed in five ML materials with piezoelectric structures, efficient emission centres and deep trap distributions, and its mechanism was investigated through detailed spectroscopic characterizations. Furthermore, we demonstrated three applications of the FICS effect in electronic signature recording, falling point monitoring and vehicle collision recording, which exhibited outstanding advantages of distributed recording, long-term storage, and no need for a continuous power supply. The FICS effect reported in this paper provides not only a breakthrough for ML materials in the field of stress recording but also a new idea for developing mechanical energy storage and conversion systems., Lung disease: Mechanoluminescence: stress-sensitive phosphors hold deep memories A material that records mechanical impacts and provides optical readouts at later dates shows promise for anti-counterfeiting devices and structural damage analysis. Phosphors, such as rare earth-doped silicates, can emit light in response to physical stress because they store charge carriers in easy-to-access energy states. Rong-Jun Xie from China’s Xiamen University and colleagues now report development of phosphors that release charges to less accessible ‘deep’ energy states after being stimulated mechanically. These carriers are retained in the deep states and then released on-demand as photon emissions following thermal treatments. The team demonstrated several applications for the new phosphors including sensors that can record signature traces and composite films that attach to vehicles to monitor for potential collisions.

Aims Osteoarthritis (OA) is a common degenerative joint disease characterized by cartilage degeneration and joint inflammation. As its pathogenesis remains unclear, there are no effective treatments established. Circular RNA (circRNA), microRNA (miRNA), and other noncoding RNAs participate in OA development; however, the effects and mechanisms of circRNA and miRNA in OA remain unknown. Main methods Cartilage miRNA was examined in patients with and without OA. Key findings CircRNA-9119 and phosphatase and tensin homolog (PTEN) expression decreased in OA-affected cartilage and interleukin (IL)-1β-induced chondrocytes, and miR-26a expression significantly decreased in normal cells and tissues. CircRNA-9119 overexpression restored chondrocyte growth, whereas IL-1β treatment impaired chondrocyte growth. Annexin V-FITC & PI flow cytometry and Bcl-2/Bax ratio measurement indicated that the apoptosis of IL-1β-treated articular chondrocytes was decreased by circRNA-9119 upregulation. Bioinformatic prediction and the dual-luciferase reporter assay indicated that circRNA-9119 served as a miR-26a sponge and that miR-26a targeted the 3′-UTR of PTEN. Transfection of chondrocytes with a circRNA-9119-overexpressing vector revealed downregulation of miR-26a expression. Furthermore, circRNA-9119 overexpression induced PTEN expression. In addition, a miR-26a mimic induced IL-1β-induced chondrocyte apoptosis, and circRNA-9119 overexpression inhibited IL-1β-induced chondrocyte apoptosis. Significance CircRNA-9119 is an important regulator of IL-1β-treated chondrocytes through the miR-26a/PTEN axis, possibly contributing to OA development.

In-situ combustion (ISC) is regarded as one of the most promising and economic alternative in the development of heavy oil reservoir. Maintaining high temperature oxidation (HTO) reaction with a high cracking efficiency at relatively low fuel consumption is deemed particularly crucial for the success of ISC process. In this work, Al2O3 nano-particles as a catalyst for achieving this during ISC is investigated in three parts. First, kinetic cell experiments with ramped temperature oxidation (RTO) process were conducted on the heaviest fraction of crude oil (residue) to obtain the apparent activation energy (Ea) fingerprints for oil-oxygen reactions with and without Al2O3. Second, thermogravimetric analysis-mass spectrometry-differential scanning calorimetry (TGA-MS-DSC) was used to detect pyrolysis products from the residue and characterize its dynamic change during the process, with SEM and EDS employed to observe the morphology and verify the condition of nano-particles afterwards. Third, combustion tube experiments were performed on the whole crude to analyze the distinction between combustion behavior with and without Al2O3. The results showed addition of Al2O3 catalyst resulted in: Ea decreased and oscillated visibly, temperature of fire front increased by 60.3–83.6 °C, speed of propagation was enhanced by 37.3%, residual coke remains in the quartz sand at the end of combustion tube experiment decreased by 24.8%, production efficiency increased by 13–64% with the reduction of COx production (12.3%) and oxygen consumption (26.9%). This shows a higher ultimate oil recovery is possible with higher upgraded oil percentage during a catalyst-assisted ISC process, which bodes well for future field application.

Dark enclaves in the late Paleocene granitoids of Doumer Island, northern Antarctic Peninsula, were examined by field investigations, zircon U–Pb geochronology, mineral chemistry, whole-rock major- and trace-elements, and Sr-Nd-Li isotopes to elucidate their origins and genetic relationships with host granitoids. The enclaves and host granitoids have similar zircon U–Pb ages of 54.6–55.7 Ma and indistinguishable Nd–Sr isotopes of eNd(t) = 4.7–6.1, ISr = 0.7036–0.7039, indicating a cogenetic relationship. A pyroxene-dioritic enclave (WA16–06-3) is a cognate restite with peritectic phases that resulted from amphibole-dehydration melting at temperatures above 800 °C and at low-P (7–10 kbar). Two other enclaves (WA16–07-2 and WA15–27-3) are fragments of the chilled margins of the conduits entrained in the later insurgent felsic magma. The flat, right-leaning REE patterns are accompanied by positive Eu anomalies (Eu/Eu* = 1.07–1.16) in the dark enclaves and negative Eu anomalies (Eu/Eu* = 0.72–0.88) in the host intrusions. The accumulation of restitic plagioclase and peritectic pyroxene could have induced the positive Eu anomalies in the dark enclaves. The P–T conditions estimated from cummingtonite rims on pyroxene in the pyroxene-dioritic enclave suggest near-isothermal decompression in an environment of fast exhumation. Enclave WA16–07-2 is part of an intense swarm of enclaves near the pluton margin, and crystallized from a less evolved melt that was mechanically eroded by the host magma, indicating a transient state of magma flow. The dioritic enclaves were transported in the form of ductile crystal–melt mushes that were rotated in the insurgent viscous granitoid magma. The enclaves become rounder and sparser towards the interior of the host pluton. Dehydration melting of metabasalt could have produced these melts that are rich in δ7Li while leaving a lighter δ7Li pyroxene-rich residue at the source. The underlying cause of these events was the retreat of the subducting proto-Pacific (Phoenix) oceanic slab beneath the western Antarctic Peninsula that triggered melting of asthenospheric mantle, with basaltic magma underplating and melting a juvenile metabasaltic lower crust.