Search

Deep anomaly detection on sequential data has garnered significant attention due to the wide application scenarios. However, deep learning-based models face a critical security threat - their vulnerability to backdoor attacks. In this paper, we explore compromising deep sequential anomaly detection models by proposing a novel backdoor attack strategy. The attack approach comprises two primary steps, trigger generation and backdoor injection. Trigger generation is to derive imperceptible triggers by crafting perturbed samples from the benign normal data, of which the perturbed samples are still normal. The backdoor injection is to properly inject the backdoor triggers to comprise the model only for the samples with triggers. The experimental results demonstrate the effectiveness of our proposed attack strategy by injecting backdoors on two well-established one-class anomaly detection models., Comment: This work is accepted by the PAKDD 2024. 12 pages

Anomaly detection in sequential data has been studied for a long time because of its potential in various applications, such as detecting abnormal system behaviors from log data. Although many approaches can achieve good performance on anomalous sequence detection, how to identify the anomalous entries in sequences is still challenging due to a lack of information at the entry-level. In this work, we propose a novel framework called CFDet for fine-grained anomalous entry detection. CFDet leverages the idea of interpretable machine learning. Given a sequence that is detected as anomalous, we can consider anomalous entry detection as an interpretable machine learning task because identifying anomalous entries in the sequence is to provide an interpretation to the detection result. We make use of the deep support vector data description (Deep SVDD) approach to detect anomalous sequences and propose a novel counterfactual interpretation-based approach to identify anomalous entries in the sequences. Experimental results on three datasets show that CFDet can correctly detect anomalous entries.

Lei Wang,1 He-Xiang Cheng,2 Yan-Hui Zhou,1 Min Ma1 1Department of Otorhinolaryngology, Head and Neck Surgery, The First Affiliated Hospital, and College of Clinical Medicine of Henan University of Science and Technology, Luoyang, 471003, People’s Republic of China; 2Department of Otolaryngology, The Luoyang First People’s Hospital, Luoyang, 471000, People’s Republic of ChinaCorrespondence: Min MaDepartment of Otorhinolaryngology, Head and Neck Surgery, The First Affiliated Hospital, and College of Clinical Medicine of Henan University of Science and Technology, No. 24 Jing Hua Road, Jianxi District, Luoyang, 471003, People’s Republic of ChinaTel +8615290533346Fax +86037969823266Email minma_drma@outlook.comObjective: The aim of the present study was to investigate the D-loop gene mutation and microsatellite instability in the mitochondrial DNA (mtDNA) and the correlation with the clinical and pathological parameters in laryngeal cancer.Methods: The tumor tissues and paratumor tissues in 60 cases of laryngeal cancer were selected, and DNA was extracted from these tissues. The D-loop region in mtDNA was amplified by PCR with the gene sequence of the amplified product being detected. The gene sequence of the detected region was compared with the revised Cambridge Reference Sequence (rCRS) and the related database by using the Mitomaster software. The correlation between the D-loop gene mutation and the clinical and pathological parameters was investigated.Results: A total of 174 mutations across 38 sites were detected in 51 (85%) of samples. Most of the mutations were concentrated in the high various (HV) I region, and the main types of mutations were the substitution of a single base or insertion and deletion of a single base. There was also microsatellite instability in the D310 region. The statistical results showed that there was no correlation between the age, gender, tumor diameter, and TNM stage, and the number of the D-loo

Artificial rain is explored as a remedy for climate change caused farmland drought and bushfires. Increasing the ion density in the open air is an efficient way to generate charged nuclei from atmospheric aerosols and induce precipitation or eliminate fog. Here we report on the development of a large commercial-installation-scale atmospheric ion generator based on corona plasma discharges, experimental monitoring, and numerical modeling of the parameters and range of the atmospheric ions, as well as the application of the generated ions to produce charged aerosols and induce precipitation at the scale of a large cloud chamber. The coverage area of the ions generated by the large corona discharge installation with the 7.2 km long wire electrode and applied voltage of -90 kV is studied under prevailing weather conditions including wind direction and speed. By synergizing over 300 000 localized corona discharge points, we demonstrate a substantial decrease in the decay of ions compared to a single corona discharge point in the open air, leading to large-scale (30m-23m-90 m) ion coverage. Once aerosols combine with the generated ions, charged nuclei are produced. Higher wind speed has led to larger areas covered by the plasma-generated ions. The cloud chamber experiments (relative humidity 130-10 %) suggest that charged aerosols generated by ions with a density of 104 cm-3 can accelerate the settlement of moisture by 38 %. These results are promising for the development of large-scale installations for the effective localized control of atmospheric phenomena.

The identification of early distant metastases (DM) in patients with newly diagnosed, previously untreated nasopharyngeal carcinoma (NPC) plays an important role in selecting the most appropriate treatment approach. Here, we sought to investigate the predictive value of distinct MRI parameters for the detection of early DM.Between November 2010 and June 2011, a total of 51 newly diagnosed NPC patients were included. All of the study participants were followed until December 2014 at a single institution after completion of therapy. DM was defined as early when they were detected on pretreatment FDG-PET scans or within 6 months after initial diagnosis. The following parameters were tested for their ability to predict early DM: pretreatment FDG-PET standardized uptake value (SUV), MRI-derived AJCC tumor staging, tumor volume, and dynamic contrast-enhanced (DCE) values. The DCE-derived ve was defined as the volume fraction of the extravascular, extracellular space.Compared with patients without early DM, patients with early DM had higher SUV, tumor volume, DCE mean (median) ve, ve skewness, ve kurtosis, and the largest mean ve selected among sequential slices (P < 0.05). No differences were identified when early DM were defined only according to the results of pretreatment FDG-PET. Among different quantitative DCE parameters, the mean ve had the highest area under curve (AUC, 0.765). However, the AUCs of SUV, tumor volume, mean ve, ve skewness, ve kurtosis, or the largest mean ve selected among the sequential slices did not differ significantly from one another (P = 0.82).Taken together, our results suggest that DCE-derived ve may be a useful parameter in combination with SUV and tumor volume for predicting early DM. Dynamic contrast-enhanced MRI may be complementary to FDG-PET for selecting the most appropriate treatment approach in NPC patients.

The identification of early distant metastases (DM) in patients with newly diagnosed, previously untreated nasopharyngeal carcinoma (NPC) plays an important role in selecting the most appropriate treatment approach. Here, we sought to investigate the predictive value of distinct MRI parameters for the detection of early DM.Between November 2010 and June 2011, a total of 51 newly diagnosed NPC patients were included. All of the study participants were followed until December 2014 at a single institution after completion of therapy. DM was defined as early when they were detected on pretreatment FDG-PET scans or within 6 months after initial diagnosis. The following parameters were tested for their ability to predict early DM: pretreatment FDG-PET standardized uptake value (SUV), MRI-derived AJCC tumor staging, tumor volume, and dynamic contrast-enhanced (DCE) values. The DCE-derived ve was defined as the volume fraction of the extravascular, extracellular space.Compared with patients without early DM, patients with early DM had higher SUV, tumor volume, DCE mean (median) ve, ve skewness, ve kurtosis, and the largest mean ve selected among sequential slices (P < 0.05). No differences were identified when early DM were defined only according to the results of pretreatment FDG-PET. Among different quantitative DCE parameters, the mean ve had the highest area under curve (AUC, 0.765). However, the AUCs of SUV, tumor volume, mean ve, ve skewness, ve kurtosis, or the largest mean ve selected among the sequential slices did not differ significantly from one another (P = 0.82).Taken together, our results suggest that DCE-derived ve may be a useful parameter in combination with SUV and tumor volume for predicting early DM. Dynamic contrast-enhanced MRI may be complementary to FDG-PET for selecting the most appropriate treatment approach in NPC patients.

We demonstrated the surface functionalization of a highly three-dimensional, superhydrophilic wicking substrate using light to immobilize functional biomolecules for sensor or microarray applications. We showed here that the three-dimensional substrate was compatible with photo-attachment and the performance of functionalization was greatly improved due to both increased surface capacity and reduced substrate reflectivity. In addition, photo-attachment circumvents the problems induced by wicking effect that was typically encountered on superhydrophilic three-dimensional substrates, thus reducing the difficulty of producing miniaturized sites on such substrate. We have investigated various aspects of photo-attachment process on the nanowire substrate, including the role of different buffers, the effect of wavelength as well as how changing probe structure may affect the functionalization process. We demonstrated that substrate fabrication and functionalization can be achieved with processes compatible with microelectronics processes, hence reducing the cost of array fabrication. Such functionalization method coupled with the high capacity surface makes the substrate an ideal candidate for sensor or microarray for sensitive detection of target analytes., National University of Singapore (Graduate School for Integrative Sciences and Engineering scholarship), GLOBALFOUNDRIES Singapore Pte. Ltd., Singapore-MIT Alliance

Thesis (M. Eng.)--Massachusetts Institute of Technology, Dept. of Materials Science and Engineering, 2010., Cataloged from PDF version of thesis., Includes bibliographical references (p. 72-83)., In the last decade, the field of single cell mechanics has emerged with the development of high resolution experimental and computational methods, providing significant amount of information about individual cells instead of the averaged characteristics provided by classical assays from large populations of cells. These single cell mechanical properties correlate closely with the intracellular organelle arrangement and organization, which are determined by load bearing cytoskeleton network comprised of biommolecules. This thesis will assess the feasibility of a high throughput single cell force spectroscopy using an atomic force microscopy (AFM)-based platform. A conventional AFM set-up employs a single cantilever probe for force measurement by using laser to detect the deflection of the cantilever structure, and usually can only handle one cell at a time. To improve the throughput of the device, a modified scheme to make use of cantilever based array is proposed and studied in this project. In addition, to complement the use of AFM array, a novel cell chip design is also presented for the fine positioning of cells in coordination with AFM cantilevers. The advantages and challenges of the system are analyzed too. To assess the feasibility of developing this technology, the commercialization possibility is discussed with intellectual property research, market analysis, cost modeling and supply chain positioning. Conclusion about this technology and its market prospect is drawn at the end of the thesis., by He Cheng., M.Eng.

Thesis (M. Eng.)--Massachusetts Institute of Technology, Dept. of Materials Science and Engineering, 2010., Cataloged from PDF version of thesis., Includes bibliographical references (p. 72-83)., In the last decade, the field of single cell mechanics has emerged with the development of high resolution experimental and computational methods, providing significant amount of information about individual cells instead of the averaged characteristics provided by classical assays from large populations of cells. These single cell mechanical properties correlate closely with the intracellular organelle arrangement and organization, which are determined by load bearing cytoskeleton network comprised of biommolecules. This thesis will assess the feasibility of a high throughput single cell force spectroscopy using an atomic force microscopy (AFM)-based platform. A conventional AFM set-up employs a single cantilever probe for force measurement by using laser to detect the deflection of the cantilever structure, and usually can only handle one cell at a time. To improve the throughput of the device, a modified scheme to make use of cantilever based array is proposed and studied in this project. In addition, to complement the use of AFM array, a novel cell chip design is also presented for the fine positioning of cells in coordination with AFM cantilevers. The advantages and challenges of the system are analyzed too. To assess the feasibility of developing this technology, the commercialization possibility is discussed with intellectual property research, market analysis, cost modeling and supply chain positioning. Conclusion about this technology and its market prospect is drawn at the end of the thesis., by He Cheng., M.Eng.

Thesis (M. Eng.)--Massachusetts Institute of Technology, Dept. of Materials Science and Engineering, 2010., Cataloged from PDF version of thesis., Includes bibliographical references (p. 72-83)., In the last decade, the field of single cell mechanics has emerged with the development of high resolution experimental and computational methods, providing significant amount of information about individual cells instead of the averaged characteristics provided by classical assays from large populations of cells. These single cell mechanical properties correlate closely with the intracellular organelle arrangement and organization, which are determined by load bearing cytoskeleton network comprised of biommolecules. This thesis will assess the feasibility of a high throughput single cell force spectroscopy using an atomic force microscopy (AFM)-based platform. A conventional AFM set-up employs a single cantilever probe for force measurement by using laser to detect the deflection of the cantilever structure, and usually can only handle one cell at a time. To improve the throughput of the device, a modified scheme to make use of cantilever based array is proposed and studied in this project. In addition, to complement the use of AFM array, a novel cell chip design is also presented for the fine positioning of cells in coordination with AFM cantilevers. The advantages and challenges of the system are analyzed too. To assess the feasibility of developing this technology, the commercialization possibility is discussed with intellectual property research, market analysis, cost modeling and supply chain positioning. Conclusion about this technology and its market prospect is drawn at the end of the thesis., by He Cheng., M.Eng.

Thesis (M. Eng.)--Massachusetts Institute of Technology, Dept. of Materials Science and Engineering, 2010., Cataloged from PDF version of thesis., Includes bibliographical references (p. 72-83)., In the last decade, the field of single cell mechanics has emerged with the development of high resolution experimental and computational methods, providing significant amount of information about individual cells instead of the averaged characteristics provided by classical assays from large populations of cells. These single cell mechanical properties correlate closely with the intracellular organelle arrangement and organization, which are determined by load bearing cytoskeleton network comprised of biommolecules. This thesis will assess the feasibility of a high throughput single cell force spectroscopy using an atomic force microscopy (AFM)-based platform. A conventional AFM set-up employs a single cantilever probe for force measurement by using laser to detect the deflection of the cantilever structure, and usually can only handle one cell at a time. To improve the throughput of the device, a modified scheme to make use of cantilever based array is proposed and studied in this project. In addition, to complement the use of AFM array, a novel cell chip design is also presented for the fine positioning of cells in coordination with AFM cantilevers. The advantages and challenges of the system are analyzed too. To assess the feasibility of developing this technology, the commercialization possibility is discussed with intellectual property research, market analysis, cost modeling and supply chain positioning. Conclusion about this technology and its market prospect is drawn at the end of the thesis., by He Cheng., M.Eng.

Thesis (M. Eng.)--Massachusetts Institute of Technology, Dept. of Materials Science and Engineering, 2010., Cataloged from PDF version of thesis., Includes bibliographical references (p. 72-83)., In the last decade, the field of single cell mechanics has emerged with the development of high resolution experimental and computational methods, providing significant amount of information about individual cells instead of the averaged characteristics provided by classical assays from large populations of cells. These single cell mechanical properties correlate closely with the intracellular organelle arrangement and organization, which are determined by load bearing cytoskeleton network comprised of biommolecules. This thesis will assess the feasibility of a high throughput single cell force spectroscopy using an atomic force microscopy (AFM)-based platform. A conventional AFM set-up employs a single cantilever probe for force measurement by using laser to detect the deflection of the cantilever structure, and usually can only handle one cell at a time. To improve the throughput of the device, a modified scheme to make use of cantilever based array is proposed and studied in this project. In addition, to complement the use of AFM array, a novel cell chip design is also presented for the fine positioning of cells in coordination with AFM cantilevers. The advantages and challenges of the system are analyzed too. To assess the feasibility of developing this technology, the commercialization possibility is discussed with intellectual property research, market analysis, cost modeling and supply chain positioning. Conclusion about this technology and its market prospect is drawn at the end of the thesis., by He Cheng., M.Eng.