Fusion Function Research Articles

Deep collaborative multi-task network: A human decision process inspired model for hierarchical image classification

Hierarchical classification is significant for big data, where the original task is divided into several sub-tasks to provide multi-granularity predictions based on a tree-shape label structure. Obviously, these sub-tasks are highly correlated: results of the coarser-grained sub-tasks can reduce the candidates for the fine-grained sub-tasks, while results of the fine-grained sub-tasks provide attributes describing the coarser-grained classes. A human can integrate feedbacks from all the related sub-tasks instead of considering each sub-task independently. Therefore, we propose a deep collaborative multi-task network for hierarchical image classification. Specifically, we first extract the relationship matrix between every two sub-tasks defined by the hierarchical label structure. Then, the information of each sub-task is broadcasted to all the related sub-tasks through the relationship matrix. Finally, to combine this information, a novel fusion function based on the task evaluation and the decision uncertainty is designed. Extensive experimental results demonstrate that our model can achieve state-of-the-art performance.

A data-level fusion model for unsupervised attribute selection in multi-source homogeneous data

Information fusion refers to derive an overall precise description of data by using certain fusion technique for utilizing the complementary information from multiple sources of data, which can facilitate effective decision-making, prediction and classification, etc. Multi-source homogeneous data, characterizing the data type of variables in the sample in form of one type (i.e., numerical or categorical) in different information sources, which widely exists in many practical applications. This paper concentrates on efficient fusion of multi-source homogeneous data with a data-level fusion model which involves the consolidation of multiple information sources and unsupervised attribute selection of the fused data. A unified description and modeling method of a multi-source homogeneous information system is introduced. The neighborhood rough sets model is used to construct the neighborhood granular structure, which uses the idea of granular computing to build methods of uncertainty measures. Given the uncertainty of fusing multiple information sources, Sup–Inf fusion functions are developed based on the proposed uncertainty measures, which can fuse the multi-source homogeneous information system into a single-source information system. Finally, an unsupervised attribute selection approach is employed to eliminate redundant attribute of the single-source information system. Theoretical analysis and comprehensive experiments on several datasets demonstrate the feasibility and superiority of our method.

SVA-SSD: saliency visual attention single shot detector for building detection in low contrast high-resolution satellite images.

Building detection in high-resolution satellite images has received great attention, as it is important to increase the accuracy of urban planning. The building boundary detection in the desert environment is a real challenge due to the nature of low contrast images in the desert environment. The traditional computer vision algorithms for building boundary detection lack scalability, robustness, and accuracy. On the other hand, deep learning detection algorithms have not been applied to such low contrast satellite images. So, there is a real need to employ deep learning algorithms for building detection tasks in low contrast high-resolution images. In this paper, we propose a novel building detection method based on a single-shot multi-box (SSD) detector. We develop the state-of-the-art SSD detection algorithm based on three approaches. First, we propose data-augmentation techniques to overcome the low contrast images’ appearance. Second, we develop the SSD backbone using a novel saliency visual attention mechanism. Moreover, we investigate several pre-trained networks performance and several fusion functions to increase the performance of the SSD backbone. The third approach is based on optimizing the anchor-boxes sizes which are used in the detection stage to increase the performance of the SSD head. During our experiments, we have prepared a new dataset for buildings inside Riyadh City, Saudi Arabia that consists of 3878 buildings. We have compared our proposed approach vs other approaches in the literature. The proposed system has achieved the highest average precision, recall, F1-score, and IOU performance. Our proposed method has achieved a fast average prediction time with the lowest variance for our testing set. Our experimental results are very promising and can be generalized to other object detection tasks in low contrast images.

Fusion and proton transfer in the B10+Al27 system at sub-barrier energies

Background: Significant enhancement in sub-barrier fusion cross sections by several orders of magnitude, in the frame of one-dimensional barrier penetration models, is observed for a variety of nuclear systems. To understand this, one needs to find the fundamental degrees of freedom relevant in the sub-barrier fusion region by incorporating nonelastic channels such as coupling of the excited target and/or projectile nuclei, particle transfer channels, threshold variation of nuclear potential, etc.Purpose: We probe what degrees of freedoms are associated with the $^{10}\mathrm{B}+^{27}\mathrm{Al}$ reaction at energies $\ensuremath{\approx}18%$ below to $\ensuremath{\approx}12%$ above the Coulomb barrier.Method: An online $\ensuremath{\gamma}$-ray spectroscopy technique is used to determine the populated excitation functions in the $^{10}\mathrm{B}+^{27}\mathrm{Al}$ reaction. The reaction cross sections obtained from these $\ensuremath{\gamma}$ rays are compared with the statistical model to explore the mechanism of fusion reactions.Results: At energies near and below the Coulomb barrier, two isotopes of sulfur ($^{34}\mathrm{S}$ and $^{32}\mathrm{S}$) were found to reproduce cross section within the frame of the pace2 model based on Hauser-Feshbach calculations, whereas isotopes of chlorine ($^{35}\mathrm{Cl}$), sulfur ($^{35}\mathrm{S}$), and phosphorous ($^{32}\mathrm{P}$) agree well but at above-barrier energies. Isotopes of Ar, Cl, S, P, and Si show significant enhancement over an entire band of energies. This enhancement (suppression) in the experimental cross section at energies above (below) the barrier points towards the involvement of some nuclear reaction mechanism other than fusion-evaporation. Distorted-wave Born approximation calculations strongly support the population of $^{28}\mathrm{Si}$ residue via the one-proton transfer process. The total fusion cross section is found to follow the theory after the inclusion of coupling of low lying excited states of projectile and target. Calculations concerning the astrophysical $S$ factor and the logarithmic derivative factor support the idea of no fusion hindrance at the studied energies. The universal fusion function benchmark shows consistency with previous data for lithium (or Li isotopes) induced systems.Conclusions: The inclusion of inelastic couplings associated with both target and projectile is important to understand the behavior of total fusion cross section. The formation of $^{28}\mathrm{Si}$ via the one-proton transfer reaction supports the present experimental technique to make inclusive transfer measurements. The calculations with the universal fusion function present a slight suppression in the higher energy region, indicating possible breakup effects in this region. The present calculations show that till $18%$ below the barrier there is no fusion hindrance. On extrapolating the theoretical values we further observed that the present system has no fusion hindrance. No connection of $^{10}\mathrm{B}+^{27}\mathrm{Al}$ fusion with astrophysics has been made, which raises the question to think of cosmic ray reactions and perhaps reactions in the early solar system, but how the fusion reaction would fit into the picture is not clear at all.

Subunit exchange among endolysosomal tethering complexes is linked to contact site formation at the vacuole.

The hexameric HOPS (homotypic fusion and protein sorting) complex is a conserved tethering complex at the lysosome-like vacuole, where it mediates tethering and promotes all fusion events involving this organelle. The Vps39 subunit of this complex also engages in a membrane contact site between the vacuole and the mitochondria, called vCLAMP. Additionally, four subunits of HOPS are also part of the endosomal CORVET tethering complex. Here, we analyzed the partition of HOPS and CORVET subunits between the different complexes by tracing their localization and function. We find that Vps39 has a specific role in vCLAMP formation beyond tethering, and that vCLAMPs and HOPS compete for the same pool of Vps39. In agreement, we find that the CORVET subunit Vps3 can take the position of Vps39 in HOPS. This endogenous pool of a Vps3-hybrid complex is affected by Vps3 or Vps39 levels, suggesting that HOPS and CORVET assembly is dynamic. Our data shed light on how individual subunits of tethering complexes such as Vps39 can participate in other functions, while maintaining the remaining subcomplex available for its function in tethering and fusion.

Application of radial basis function in gravity field modeling and data fusion

Application of radial basis function in gravity field modeling and data fusion

Theoretical study of dynamic effects on fusion cross sections for reactions S32,34,36+Pb204,206,208

The universal fusion function prescription and the empirical coupled-channel (ECC) model are used to analyze the dynamic effects on fusion cross sections for reactions $^{32,34,36}\mathrm{S}+^{204,206,208}\mathrm{Pb}$. An examination of the reduced fusion functions shows that the total dynamic effects on fusion cross sections in $^{36}\mathrm{S}+^{204,206,208}\mathrm{Pb}$ are almost the same as those in $^{34}\mathrm{S}+^{204,206,208}\mathrm{Pb}$. Furthermore, at sub-barrier energies of the reactions $^{32,34,36}\mathrm{S}+^{208}\mathrm{Pb}$, the largest enhancement of reduced fusion function is observed in the reaction $^{32}\mathrm{S}+^{208}\mathrm{Pb}$. Then the enhancement of fusion cross sections due to couplings to inelastic excitation channels and neutron transfer channels with positive $Q$ values is investigated by using the ECC model. The results show that the experimental data of the reactions $^{34,36}\mathrm{S}+^{204,206,208}\mathrm{Pb}$ and $^{32}\mathrm{S}+^{208}\mathrm{Pb}$ are reproduced well by the ECC model. In addition, it can be found that the effect of coupling to positive $Q$-value neutron transfer channels is necessary to reproduce the sub-barrier fusion data of the reaction $^{32}\mathrm{S}+^{208}\mathrm{Pb}$. Furthermore, for the two reactions $^{32}\mathrm{S}+^{204}\mathrm{Pb}$ and $^{32}\mathrm{S}+^{206}\mathrm{Pb}$, the effects of coupling to positive $Q$-value neutron transfer channels are predicted to be significant. The fusion cross sections for these two reactions are also predicted.

A Fully Automatic, Interpretable and Adaptive Machine Learning Approach to Map Burned Area from Remote Sensing

The paper proposes a fully automatic algorithm approach to map burned areas from remote sensing characterized by human interpretable mapping criteria and explainable results. This approach is partially knowledge-driven and partially data-driven. It exploits active fire points to train the fusion function of factors deemed influential in determining the evidence of burned conditions from reflectance values of multispectral Sentinel-2 (S2) data. The fusion function is used to compute a map of seeds (burned pixels) that are adaptively expanded by applying a Region Growing (RG) algorithm to generate the final burned area map. The fusion function is an Ordered Weighted Averaging (OWA) operator, learnt through the application of a machine learning (ML) algorithm from a set of highly reliable fire points. Its semantics are characterized by two measures, the degrees of pessimism/optimism and democracy/monarchy. The former allows the prediction of the results of the fusion as affected by more false positives (commission errors) than false negatives (omission errors) in the case of pessimism, or vice versa; the latter foresees if there are only a few highly influential factors or many low influential ones that determine the result. The prediction on the degree of pessimism/optimism allows the expansion of the seeds to be appropriately tuned by selecting the most suited growing layer for the RG algorithm thus adapting the algorithm to the context. The paper illustrates the application of the automatic method in four study areas in southern Europe to map burned areas for the 2017 fire season. Thematic accuracy at each site was assessed by comparison to reference perimeters to prove the adaptability of the approach to the context; estimated average accuracy metrics are omission error = 0.057, commission error = 0.068, Dice coefficient = 0.94 and relative bias = 0.0046.

One-neutron transfer, complete fusion, and incomplete fusion from the Be9 + Au197 reaction

In this work, one-neutron transfer (pickup and stripping), complete and incomplete fusion cross sections for the $^{9}\mathrm{Be}+^{197}\mathrm{Au}$ system were measured over a wide range of energies around the Coulomb barrier by the offline $\ensuremath{\gamma}$-ray detection method. Coupled-channel calculations were used to determine the elastic, inelastic, and transfer cross sections. Coupled reaction channel calculations were performed to derive the one-neutron stripping and pickup cross sections. Three-body continuum discretized coupled-channel calculations were used to determine the effect of the breakup channel on the other reaction mechanisms. The reduced complete and total fusion were found to be hindered above and enhanced below the Coulomb barrier compared with the universal fusion function due to the breakup plus transfer effects.

Convolutional Neural Network Using for Multi-Sensor 3D Object Detection

The purpose of this article is to detect 3D objects inside the independent vehicle with great accuracy. The method proposed a Multi-View 3D System (MV3D) framework which encodes the sparse 3d-point cloud with a compact multi-view image, using LIDAR satellite image and RGB pictures as inputs, and predicts 3D boundary boxes. The network comprises two sub-networks: one for creating 3D artifacts and one for multi-visual fusion functionality. Propose an autonomous 3D object tracking approach to manipulate sparse and dense knowledge about romanticizing and geometry in stereo images. The Stereo R-CNN strategy applies Faster R-CNNs to stereo inputs such that objects are simultaneously identified and linked in conservative and liberal images. Such charts were then combined and fed into a 3D proposal generator to generate accurate 3D proposals for vehicles. In the second step, the refining network extended the features of the proposal regions further and carried through the classification, regression of a 3D package box, and guidance estimates, to predict vehicle location and heading in 3D area and add additional branches after the stereo region Proposal Network (RPN).

β-Asarone Attenuates Aβ-Induced Neuronal Damage in PC12 Cells Overexpressing APPswe by Restoring Autophagic Flux.

Alzheimer’s disease (AD) is a neurodegenerative disorder characterized by progressive memory damage and cognitive dysfunction. Studies have shown that defective autophagic flux is associated with neuronal dysfunction. Modulating autophagic activity represents a potential method of combating AD. In Chinese medicine, Acori Tatarinowii Rhizoma is used to treat dementia and amnesia. β-Asarone, an active component of this rhizome can protect PC12 cells from Aβ-induced injury and modulate expression of autophagy factors. However, its cytoprotective mechanisms have yet to be discerned. It is unclear whether β-asarone affects autophagic flux and, if it does, whether this effect can alleviate Aβ cell damage. In the present study, we constructed APPswe-overexpressing PC12 cell line as a cell model of Aβ-induced damage and assessed expression of autophagic flux-related proteins as well as the number and morphology of autophagosomes and autolysosomes. Our results show that β-asarone decreases the expression levels of Beclin-1, p62, LC3-Ⅱ, and Aβ1-42. β-Asarone reduced the number of autophagosomes and increased the number of autolysosomes, as determined by confocal laser scanning microscopy and transmission electron microscopy. Our results suggest that β-asarone can protect PC12 cells from Aβ-induced damage by promoting autophagic flux, which may be achieved by enhancing autophagosome-lysosome fusion and/or lysosome function.

Mfn2 localization in the ER is necessary for its bioenergetic function and neuritic development

Mfn2 is a mitochondrial fusion protein with bioenergetic functions implicated in the pathophysiology of neuronal and metabolic disorders. Understanding the bioenergetic mechanism of Mfn2 may aid in designing therapeutic approaches for these disorders. Here we show using endoplasmic reticulum (ER) or mitochondria-targeted Mfn2 that Mfn2 stimulation of the mitochondrial metabolism requires its localization in the ER, which is independent of its fusion function. ER-located Mfn2 interacts with mitochondrial Mfn1/2 to tether the ER and mitochondria together, allowing Ca2+ transfer from the ER to mitochondria to enhance mitochondrial bioenergetics. The physiological relevance of these findings is shown during neurite outgrowth, when there is an increase in Mfn2-dependent ER-mitochondria contact that is necessary for correct neuronal arbor growth. Reduced neuritic growth in Mfn2 KO neurons is recovered by the expression of ER-targeted Mfn2 or an artificial ER-mitochondria tether, indicating that manipulation of ER-mitochondria contacts could be used to treat pathologic conditions involving Mfn2.

The role of the PZP domain of AF10 in acute leukemia driven by AF10 translocations

Chromosomal translocations of the AF10 (or MLLT10) gene are frequently found in acute leukemias. Here, we show that the PZP domain of AF10 (AF10PZP), which is consistently impaired or deleted in leukemogenic AF10 translocations, plays a critical role in blocking malignant transformation. Incorporation of functional AF10PZP into the leukemogenic CALM-AF10 fusion prevents the transforming activity of the fusion in bone marrow-derived hematopoietic stem and progenitor cells in vitro and in vivo and abrogates CALM-AF10-mediated leukemogenesis in vivo. Crystallographic, biochemical and mutagenesis studies reveal that AF10PZP binds to the nucleosome core particle through multivalent contacts with the histone H3 tail and DNA and associates with chromatin in cells, colocalizing with active methylation marks and discriminating against the repressive H3K27me3 mark. AF10PZP promotes nuclear localization of CALM-AF10 and is required for association with chromatin. Our data indicate that the disruption of AF10PZP function in the CALM-AF10 fusion directly leads to transformation, whereas the inclusion of AF10PZP downregulates Hoxa genes and reverses cellular transformation. Our findings highlight the molecular mechanism by which AF10 targets chromatin and suggest a model for the AF10PZP-dependent CALM-AF10-mediated leukemogenesis.

To design an intelligent analysis platform for the disabled population data

To design demand-oriented intelligent analysis platform framework for the disabled population data from overall management to security. DATAI-WebEx, active learning, Browser/Server architecture, role-role-based access control, Bayesian network, GIS analysis technology, cluster analysis, regression analysis and other intelligent technologies were used in this study, which provided the functions of multi-source heterogeneous disabled population data fusion, intelligent analysis, secure access and data sharing. The disability data warehouse and intelligent analysis platform can realize the structured and unstructured information disabled population data alignment and data fusion. Also, it can provide disability risk module clustering, disability risk factor identification, disabled distribution analysis, disability scale dynamic trajectory prediction, early warning, disability grade development. Moreover, it can provide a guarantee for the safe and convenient access of sensitive data with the support of "classified boxes", and realize the safe sharing of data of the disabled population. The disability data warehouse and intelligent analysis platform can provide the services of "comprehensive fusion-intelligent mining-safe sharing".

Abstract LB188: RNA-seq and chIP-seq profiling identifies genes and pathways dysregulated by hey1-ncoa2 fusion and shed a light on mesenchymal chondrosarcoma tumorigenesis

Abstract RNA-seq and chIP-seq profiling identifies genes and pathways dysregulated by hey1-ncoa2 fusion and shed a light on mesenchymal chondrosarcoma tumorigenesis Wenqing Qi1, Wojciech Rosikiewicz2, Zhaohong Yin1, Beisi Xu2, Shibiao Wan2, Yiping Fan2, Gang Wu1,2 and Lu Wang1 1. Department of Pathology, St. Jude Children's Research Hospital, Memphis, TN 2. Center for Applied Bioinformatics, St. Jude Children's Research Hospital, Memphis, TN Correspondence and requests for materials should be addressed to Lu Wang (email: lu.wang2@stjude.org) Mesenchymal chondrosarcoma is a rare and often aggressive cancer that accounts for 2-10% of all chondrosarcomas. Genetically, this tumor is characterized by the recurrent HEY1-NCOA2 fusion. However, the function of HEY1-NCOA2 fusion in mesenchymal chondrosarcoma remains to be elucidated. In this study, we transduced iPSC-derived mesenchymal stem cells (MSCs) with inducible expression HEY1, NCOA2 and HEY1-NCOA2 (HN) construct, respectively. The iPSC-MSCs expressing HEY1-NCOA2 (HN_MSCs) showed significantly accelerated growth rates and higher percentage of cells in S phase compared with cells expressing wild-type HEY1 or NCOA2 (HEY1_MSCs or NCOA2_MSCs). Intriguingly, in 3-D spheroidal growth model, we observed both accelerated cell proliferation and distinct morphological features of the HN_MSCs in contrast to the control cells. To identify direct target genes and the pathways modulated by HEY1-NCOA2, stably transduced iPSC-MSCs with and without the expression of the exogenous HN, HEY1 and NCOA2, were subjected to Chromatin Immunoprecipitation Sequencing and RNA-seq analyses. ChIP-seq data showed a largely overlap in DNA-binding peaks observed in HN_MSCs and HEY1_MSCs (~90% of HN binding peaks). In contrast, very few (0.1%) overlaps observed between HN_MSCs and NCOA2_MSCs. In addition, ~25% of HN binding peaks contain the canonical HEY1 binding E-box motif (CACGTG). Through RNA-seq analysis, HN_MSC and HEY1-MSC can be robustly separated by gene expression profiling, and GSEA pathway enrichment analysis revealed that genes encoding cell cycle related targets of E2F transcription factors and genes involved in the G2/M checkpoint as in progression through the cell division cycle are significantly enriched in genes highly expressed in HN_MSC. Differential expression analysis identified 316 genes that were significantly up-regulated by HN but down-regulated by HEY1, of which CCND1 is one of the most differentially expressed genes. When correlating with ChIP-seq analysis, 59 of the 316 genes showed DNA-binding peaks with the canonical HEY1-binding motif, including HES1 and SOX4 which were reported to be involved in chondrogenesis. We further verified the significant upregulation of CCND1, HES1 and SOX4 in HN_MSCs by Q-RT-PCR. Taken together, our study demonstrated that HN is oncogenic by significantly accelerating cell proliferation. RNA-seq and ChIP-seq profiling revealed that HEY1-NCOA2 directly binds to HEY1 target genes through HEY1-binding motif to dysregulate their expression and consequently rewire the downstream pathways. The combined RNA-seq and ChIP-seq analysis suggested that HES1, SOX4 and CCND1 are critical targets of the transcriptional dysregulation mediated by HN and may play key roles in mesenchymal chondrosarcoma tumorigenesis. Citation Format: Wenqing Qi. RNA-seq and chIP-seq profiling identifies genes and pathways dysregulated by hey1-ncoa2 fusion and shed a light on mesenchymal chondrosarcoma tumorigenesis [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr LB188.

Study of breakup induced fusion mechanisms for 9Be+181Ta system at around barrier energies

The excitation functions for total, complete and incomplete fusion (TF, CF and ICF) mechanisms have been investigated for 9Be + [Formula: see text]Ta system at near and above barrier energy region. The calculations are performed by employing the code FRESCO based on continuum discretized coupled channel (CDCC) method. An enhancement is observed in the TF excitation function with respect to the data when breakup couplings are taken into account. Further, the CF and ICF probabilities, defined as the ratio of CF (ICF) cross-sections to the TF cross-sections, are also calculated. The dominance of breakup induced fusion is noticed at near barrier energies.

Health-related quality of life and anxiety associated with childhood intermittent exotropia before and after surgical correction

BackgroundIntermittent exotropia (IXT) is the most common form of exotropia in children. In addition to cosmetic effects and loss of stereoscopic function, IXT may negatively impact the psychological well-being of children and their parents. The purpose of this study was to assess the patient-reported outcomes of Chinese children with IXT before and after strabismus surgery.MethodsThe records of children with IXT who underwent strabismus surgery at the Zhongshan Ophthalmic Center of Sun Yat-sen University, China over the period from January 1, 2016 to December 31, 2018 were prospectively recruited. All children underwent ophthalmic and orthoptic examinations, including the prism and alternate cover test, fusion function by synoptophore, stereoacuity and Newcastle control score. Two patient-reported outcome measures were used: the intermittent Exotropia Questionnaire (IXTQ) to measure disease-specific health-related quality of life (HRQOL) and the Hospital Anxiety and Depression Scale (HADS) to measure anxiety and depression. Patient-reported outcome measurements were made before and after surgery with responses from children and their parents.ResultsA total of 389 children were eligible for inclusion (47.8% male, 52.2% female, mean + SD age = 8.17 ± 2.81). Preoperative IXTQ scores in both children (48.21 ± 26.2) and their parents (44.6 ± 25.68) were significantly correlated with near stereoacuity (P = 0.029 and P = 0.015, respectively). The angle of deviation at near vision showed a negative linear relationship with visual function (P = 0.026) and psychological (P = 0.019) scores as well as opinions regarding surgery (P = 0.024). HADS scores (anxiety scale score: 11 ± 2.92, depression scale score: 10.44 ± 2.9) were also related to near stereoacuity (P < 0.05). After surgery, both children’s (74.83 ± 16.59) and parents’ (68.57 ± 17.06) IXTQ scores significantly improved (p<0.01). Children’s IXTQ scores were related to the angle of deviation at distance, and their psychological and visual function scores showed a negative relationship with the angle of deviation at near vision (P < 0.05).ConclusionChildren and parents’ HRQOL and HADS were associated with near stereoacuity. Parents usually attend more readily to the angle of deviation at near in their IXT children. HRQOL improved significantly after surgery and can be used as one of the indices for preoperative evaluation but is not recommended as a criterion for surgical intervention.

DEEPLIO: DEEP LIDAR INERTIAL SENSOR FUSION FOR ODOMETRY ESTIMATION

Abstract. Having a good estimate of the position and orientation of a mobile agent is essential for many application domains such as robotics, autonomous driving, and virtual and augmented reality. In particular, when using LiDAR and IMU sensors as the inputs, most existing methods still use classical filter-based fusion methods to achieve this task. In this work, we propose DeepLIO, a modular, end-to-end learning-based fusion framework for odometry estimation using LiDAR and IMU sensors. For this task, our network learns an appropriate fusion function by considering different modalities of its input latent feature vectors. We also formulate a loss function, where we combine both global and local pose information over an input sequence to improve the accuracy of the network predictions. Furthermore, we design three sub-networks with different modules and architectures derived from DeepLIO to analyze the effect of each sensory input on the task of odometry estimation. Experiments on the benchmark dataset demonstrate that DeepLIO outperforms existing learning-based and model-based methods regarding orientation estimation and shows a marginal position accuracy difference.

Apigenin ameliorates hepatic lipid accumulation by activating the autophagy-mitochondria pathway.

Apigenin, a flavonoid isolated from plants, provides protection against non-alcoholic fatty liver disease. However, the mechanism by which apigenin decreases lipid accumulation in the liver is unclear. In this study, we investigated the molecular mechanism underlying the beneficial effect of apigenin on the hepatic deregulation of lipid metabolism. Oleic acid (OA)-induced lipid accumulation in human hepatoma cells (Huh7 cells) was used as an in vitro model. Western blot analysis was used for evaluating protein expression. Oil red O staining, Nile red staining, and conventional assay kits were used to assess the level of lipids. Immunocytochemistry was performed to observe mitochondrial morphology. Seahorse XF analyzer was used to measure mitochondrial bioenergetics. Treatment with OA induced lipid accumulation in Huh7 cells, which was attenuated by apigenin. Mechanistically, treatment with apigenin increased the expression of autophagy-related proteins including Beclin1, autophagy related gene 5 (ATG5), ATG7, and LC3II, and the formation of autophagolysosomes, leading to an increase in intracellular levels of fatty acids. Inhibition of autophagy by bafilomycin A1 or chloroquine abolished the protection of apigenin in OA-induced lipid accumulation. Apigenin up-regulated the protein expression related to the β-oxidation pathway including acyl-CoA synthetase long chain family member 1, carnitine palmitoyltransferase 1α, acyl-CoA oxidase 1, peroxisome proliferator activated receptor (PPAR) α, and PPARγ coactivator 1-α. Moreover, apigenin increased the mitochondrial network structure and mitochondrial function by increasing the protein expression related to the process of mitochondria fusion and mitochondrial function. Collectively, our findings suggest that apigenin ameliorates hepatic lipid accumulation by activating the autophagy-mitochondrial pathway.

Dual-granularity weighted ensemble clustering

Ensemble clustering is one of the research hotspots of data mining in recent years. The selection of high-quality and large-diversity base clustering results plays a key role in the quality of the final result. Traditional ensemble clustering selection algorithms usually treat each base clustering result as a whole which ignores the difference between the clusters in the same clustering result. It may cause the validity of the final clustering result to be affected. Aiming at this problem, inspired by the measurement method of uncertainty in the rough set theory, a dual-granularity weighted ensemble clustering model is proposed. The main contribution of this paper is shown as follows: (1) the evaluation of the reliability of clusters is transformed into an uncertainty measurement problem in the rough set; (2) in a finer-grained level, a sample local similarity measurement method is designed; (3) a weighted co-association matrix elements generation method based on global cluster reliability and local sample pair similarity is proposed, then the fusion function is used to get the final clustering result. Experimental results show that the proposed method is not sensitive to the size and diversity of base clustering members which has good robustness and stability. The final result obtained by this model is closer to the actual distribution of data sets.