A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is
In our framework, an object is made up of K distinct parts or units, and we parse a test instance by inferring the K parts, where each part occupies a distinct location in the feature space, and the instance features at this location, manifest as an active subset of part templates shared across all instances. We recognize test instances by comparing its active templates and the relative geometry of its part locations against those of the presented few-shot instances. We propose an end-to-endarXiv:2207.07110v1 fatcat:ls7ld5pzonacdmh2tx66di45si
more »... ining method to learn part templates on-top of a convolutional backbone. To combat visual distortions such as orientation, pose and size, we learn multi-scale templates, and at test-time parse and match instances across these scales. We show that our method is competitive with the state-of-the-art, and by virtue of parsing enjoys interpretability as well.
The authors would like to thank Ruizhao Zhu for helpful discussions. ... . • Pengkai and Venkatesh are with the Department of Electrical and Computer Engineering at Boston University. ...arXiv:2204.08090v1 fatcat:3dealbb62zdwtnymninemwrfci
We propose a novel Generalized Zero-Shot learning (GZSL) method that is agnostic to both unseen images and unseen semantic vectors during training. Prior works in this context propose to map high-dimensional visual features to the semantic domain, we believe contributes to the semantic gap. To bridge the gap, we propose a novel low-dimensional embedding of visual instances that is "visually semantic." Analogous to semantic data that quantifies the existence of an attribute in the presentedarXiv:1811.07993v2 fatcat:apeqd4gl7rcr7egap4j74vctdm
more »... nce, components of our visual embedding quantifies existence of a prototypical part-type in the presented instance. In parallel, as a thought experiment, we quantify the impact of noisy semantic data by utilizing a novel visual oracle to visually supervise a learner. These factors, namely semantic noise, visual-semantic gap and label noise lead us to propose a new graphical model for inference with pairwise interactions between label, semantic data, and inputs. We tabulate results on a number of benchmark datasets demonstrating significant improvement in accuracy over state-of-the-art under both semantic and visual supervision.
Networked embedded devices (IoTs) of limited CPU, memory and power resources are revolutionizing data gathering, remote monitoring and planning in many consumer and business applications. Nevertheless, resource limitations place a significant burden on their service life and operation, warranting cost-aware methods that are capable of distributively screening redundancies in device information and transmitting informative data. We propose to train a decentralized gated network that, given andblp:conf/aistats/ZhuAFJS19 fatcat:kqj4k2neefdfrbat74t2rutgoa
more »... erved instance at test-time, allows for activation of select devices to transmit information to a central node, which then performs inference. We analyze our proposed gradient descent algorithm for Gaussian features and establish convergence guarantees under good initialization. We conduct experiments on a number of real-world datasets arising in IoT applications and show that our model results in over 1.5X service life with negligible accuracy degradation relative to a performance achievable by a neural network.
Zero-shot detection, namely, localizing both seen and unseen objects, increasingly gains importance for largescale applications, with large number of object classes, since, collecting sufficient annotated data with ground truth bounding boxes is simply not scalable. While vanilla deep neural networks deliver high performance for objects available during training, unseen object detection degrades significantly. At a fundamental level, while vanilla detectors are capable of proposing boundingdoi:10.1109/cvpr42600.2020.01171 dblp:conf/cvpr/ZhuWS20 fatcat:tpn3cdxt5jdslb52p7ze6hb44y
more »... s, which include unseen objects, they are often incapable of assigning highconfidence to unseen objects, due to the inherent precision/recall tradeoffs that requires rejecting background objects. We propose a novel detection algorithm "Don't Even Look Once (DELO)," that synthesizes visual features for unseen objects and augments existing training algorithms to incorporate unseen object detection. Our proposed scheme is evaluated on PascalVOC and MSCOCO, and we demonstrate significant improvements in test accuracy over vanilla and other state-of-art zero-shot detectors.
As we move towards large-scale object detection, it is unrealistic to expect annotated training data, in the form of bounding box annotations around objects, for all object classes at sufficient scale, and so methods capable of unseen object detection are required. We propose a novel zero-shot method based on training an end-to-end model that fuses semantic attribute prediction with visual features to propose object bounding boxes for seen and unseen classes. While we utilize semantic featuresdoi:10.1109/tcsvt.2019.2899569 fatcat:aoqb6egmerhnjpufnlisofwkh4
more »... uring training, our method is agnostic to semantic information for unseen classes at test-time. Our method retains the efficiency and effectiveness of YOLOv2 for objects seen during training, while improving its performance for novel and unseen objects. The ability of state-of-art detection methods to learn discriminative object features to reject background proposals also limits their performance for unseen objects. We posit that, to detect unseen objects, we must incorporate semantic information into the visual domain so that the learned visual features reflect this information and leads to improved recall rates for unseen objects. We test our method on PASCAL VOC and MS COCO dataset and observed significant improvements on the average precision of unseen classes.
Zero-shot detection, namely, localizing both seen and unseen objects, increasingly gains importance for large-scale applications, with large number of object classes, since, collecting sufficient annotated data with ground truth bounding boxes is simply not scalable. While vanilla deep neural networks deliver high performance for objects available during training, unseen object detection degrades significantly. At a fundamental level, while vanilla detectors are capable of proposing boundingarXiv:1911.07933v3 fatcat:c4nesge7bzfknm55zkzs66ywp4
more »... es, which include unseen objects, they are often incapable of assigning high-confidence to unseen objects, due to the inherent precision/recall tradeoffs that requires rejecting background objects. We propose a novel detection algorithm Dont Even Look Once (DELO), that synthesizes visual features for unseen objects and augments existing training algorithms to incorporate unseen object detection. Our proposed scheme is evaluated on Pascal VOC and MSCOCO, and we demonstrate significant improvements in test accuracy over vanilla and other state-of-art zero-shot detectors
MILCOM 2009 - 2009 IEEE Military Communications Conference
Carrier Sense Multiple Access (CSMA) is the conventional medium access method used in wireless Ad Hoc networks. It can be enhanced as a Space-Division based CSMA (namely, SD-CSMA) by using multi-antenna techniques, which allows concurrent link communications in the network. In this paper, we present a MAC design that can adaptively switch between CSMA and SD-CSMA in multi-antenna based Ad Hoc Networks. We first investigate the CSMA mode and the SD-CSMA mode by considering several practicaldoi:10.1109/milcom.2009.5379728 fatcat:jqjznzefe5ejbbun64oypjdvdq
more »... raints from MIMO physical layer, which include imperfect channel estimation, modulation and coding scheme, MMSE detection, power constraint and packet length. We then propose a selection criterion that adaptively switches between CSMA and SD-CSMA based on current wireless channels among nodes. Simulation results verify that the proposed design can outperform both the CSMA mode and the SD-CSMA mode in terms of aggregate throughput.
The efficient targeting of drugs to tumor cell and subsequent rapid drug release remain primary challenges in the development of nanomedicines for cancer therapy. Here, we constructed a glucose transporter 1 (GLUT1)-targeting and tumor cell microenvironment-sensitive drug release Glucose-PEG-PAMAM-s-s-Camptothecin-Cy7 (GPCC) conjugate to tackle the dilemma. The conjugate was characterized by a small particle size, spherical shape, and glutathione (GSH)-sensitive drug release. In vitro tumordoi:10.1080/10717544.2017.1419511 pmid:29282992 pmcid:PMC6058575 fatcat:bq74xnszhncijg23efwkqaoosa
more »... eting was explored in monolayer (2D) and multilayer tumor spheroid (3D) HepG2 cancer cell models (GLUT1 þ ). The cellular uptake of GPCC was higher than that in the control groups and that in normal L02 cells (GLUT1 À ), likely due to the conjugated glucose moiety. Moreover, the GPCC conjugate exhibited stronger cytotoxicity, higher S arrest and enhanced apoptosis and necrosis rate in HepG2 cells than control groups but not L02 cells. However, the cytotoxicity of GPCC was lower than that of free CPT, which could be explained by the slower release of CPT from the GPCC compared with free CPT. Additional in vivo tumor targeting experiments demonstrated the superior tumor-targeting ability of the GPCC conjugate, which significantly accumulated in tumor meanwhile minimize in normal tissues compared with control groups. The GPCC conjugate showed better pharmacokinetic properties, enabling a prolonged circulation time and increased camptothecin area under the curve (AUC). These features contributed to better therapeutic efficacy and lower toxicity in H22 hepatocarcinoma tumor-bearing mice. The GLUT1-targeting, GSH-sensitive GPCC conjugate provides an efficient, safe and economic approach for tumor cell targeted drug delivery. ARTICLE HISTORY
Cerasus Campanulata is one of several species belonging to the Prunoideae focke, a subfamily of the flowering plant Rosaceae. We investigated the details of its chloroplast genome which may reveal its genus independent of morphological determination. Here, we determined the complete chloroplast (cp) genome sequence of C. campanulata and performed sequence analysis to reveal the presence of 18 forward repeats, 20 palindrome repeats, 2 complement repeats, 4 reverse repeats and 93 simple sequencedoi:10.32604/phyton.2020.08831 fatcat:w6xaba54trdezhz67et6fakp3a
more »... epeats (SSRs). We additionally performed a comparative study of C. campanulata and seven other Prunoideae focke species. Then, maximum parsimony (MP) and maximum likelihood (ML) phylogenetic analyses were carried out in the little part of Rosaceae, respectively. The results strongly support a position of C. campanulata as a member of the Cerasus in the Rosaceae family. Moreover, the complete cp genome can be used for plant phylogenetic and evolutionary studies that will provide insight into the degree of gene conservation.
Mammalian mitochondrial NAD-dependent isocitrate dehydrogenase (NAD-IDH) catalyzes the decarboxylation of isocitrate into α-ketoglutarate in the tricarboxylic acid cycle. It exists as the α2βγ heterotetramer composed of the αβ and αγ heterodimers. Different from the αγ heterodimer that can be allosterically activated by CIT and ADP, the αβ heterodimer cannot be allosterically regulated by the activators; however, the molecular mechanism is unclear. We report here the crystal structures of thedoi:10.1074/jbc.ra119.010099 pmid:31515270 pmcid:PMC6827300 fatcat:w7d6rzmp2fbznhiud5m6jdx2o4
more »... heterodimer of human NAD-IDH with the α subunit in apo form and in Ca2+-bound, NAD-bound, and NADH-bound forms. Structural analyses and comparisons reveal that the αβ heterodimer has a similar yet more compact overall structure compared with the αγ heterodimer and contains a pseudo-allosteric site that is structurally different from the allosteric site. In particular, the β3-α3 and β12-α8 loops of the β subunit at the pseudo-allosteric site adopt significantly different conformations from those of the γ subunit at the allosteric site and hence impede the binding of the activators, explaining why the αβ heterodimer cannot be allosterically regulated by the activators. The structural data also show that NADH can compete with NAD to bind to the active site and inhibits the activity of the αβ heterodimer. These findings together with the biochemical data reveal the molecular basis for the function of the αβ heterodimer of human NAD-IDH.
Peripheral artery disease (PAD) is a serious public health issue, characterized by circulation disorder of the lower extreme that reduces the physical activity of the lower extremity muscle. The artery narrowed by atherosclerotic lesions initiates limb ischemia. In the progression of treatment, reperfusion injury is still inevitable. Ischemia-reperfusion injury induced by PAD is responsible for hypoxia and nutrient deficiency. PAD triggers hindlimb ischemia and reperfusion (I/R) cycles throughdoi:10.1155/2021/4954070 pmid:34899949 pmcid:PMC8660193 fatcat:27udsmozjzflpmarpudjgm54ge
more »... arious mechanisms, mainly including mitochondrial dysfunction and inflammation. Alternatively, mitochondrial dysfunction plays a central role. The I/R injury may cause cells' injury and even death. However, the mechanism of I/R injury and the way of cell damage or death are still unclear. We review the pathophysiology of I/R injury, which is majorly about mitochondrial dysfunction. Then, we focus on the cell damage and death during I/R injury. Further comprehension of the progress of I/R will help identify biomarkers for diagnosis and therapeutic targets to PAD. In addition, traditional Chinese medicine has played an important role in the treatment of I/R injury, and we will make a brief introduction.
AbstractThe aluminum (Al) cation Al3+ in acidic soil shows severe rhizotoxicity that inhibits plant growth and development. Most woody plants adapted to acidic soils have evolved specific strategies against Al3+ toxicity, but the underlying mechanism remains elusive. The four-carbon amino acid gamma-aminobutyric acid (GABA) has been well studied in mammals as an inhibitory neurotransmitter; GABA also controls many physiological responses during environmental or biotic stress. The woody plantdoi:10.1038/s41438-021-00517-y pmid:33790239 fatcat:qz2pwhmmibcozixpvrseku7tji
more »... rid Liriodendron (L. chinense × tulipifera) is widely cultivated in China as a horticultural tree and provides high-quality timber; studying its adaptation to high Al stress is important for harnessing its ecological and economic potential. Here, we performed quantitative iTRAQ (isobaric tags for relative and absolute quantification) to study how protein expression is altered in hybrid Liriodendron leaves subjected to Al stress. Hybrid Liriodendron shows differential accumulation of several proteins related to cell wall biosynthesis, sugar and proline metabolism, antioxidant activity, cell autophagy, protein ubiquitination degradation, and anion transport in response to Al damage. We observed that Al stress upregulated glutamate decarboxylase (GAD) and its activity, leading to increased GABA biosynthesis. Additional GABA synergistically increased Al-induced antioxidant enzyme activity to efficiently scavenge ROS, enhanced proline biosynthesis, and upregulated the expression of MATE1/2, which subsequently promoted the efflux of citrate for chelation of Al3+. We also showed similar effects of GABA on enhanced Al3+ tolerance in Arabidopsis. Thus, our findings suggest a function of GABA signaling in enhancing hybrid Liriodendron tolerance to Al stress through promoting organic acid transport and sustaining the cellular redox and osmotic balance.
Introduction Some plant species have evolved multifaceted mechanisms to mitigate adverse stress effects and improve their tolerance toward extreme environmental conditions such as salt stress (Zhu, 2002 ... ROS signaling disruption causes defects during plant development (Guillou et al., 2022; Zhang et al., 2022) , activating detoxification pathways for remediation of salinity-induced damage (Zhu, 2002) ...doi:10.3389/fpls.2022.961651 pmid:36003812 pmcid:PMC9393555 fatcat:qbuofowg4rdo5h7aciyyxkfyhe
(Zhu et al., 2018; Kumar Verma et al., 2018; Xian et al., 2018b; Jiang et al., 2018) . ... (Zhu et al., 2018) further extend this insight and propose visual part detector (VPDE-Net) and utilize high-dimensional part feature vectors as an input for semantic transfer, namely, to synthesize unseen ...arXiv:1901.09079v3 fatcat:synn6rzpkjaa3iiy2trkcih6gi
« Previous Showing results 1 — 15 out of 24 results