Search

HoBi-like pestivirus, an emerging species within the Pestivirus genus, is an important pathogen associated with a variety of clinical manifestations of ruminants, especially cattle. HoBi-like pestiviruses were identified in several countries and from different hosts, and raised concerns with regard to their acute and persistent infections, which is implicated in economic losses for cattle farmers. However, the transmission path, codon usage bias, and host adaptation of the virus has not been studied. Hence, we performed the analysis the spatio-temporal transmission based on the available 5'-UTR sequences of HoBi-like pestivirus, and then conducted codon analysis of the complete coding sequence of the virus. The results show the virus appeared in 1952 (95% HPD: 1905-1985) and may have originated in India. In addition, Italy is the hub for the spread of the virus. Moreover, six potential recombination events and two complex recombination events were discovered. Analysis of codon usage patterns revealed that the effective number of codon (ENC) values with an average of 50.85, and the codon usage bias is greatly affected by natural selection, which is different from the previous BVDV-1, 2. Finally, codon adaptation index (CAI) analysis shows that pigs may be the potential origin species of the HoBi-like pestivirus. These findings will contribute to more effective control of the spread of the virus, extend the knowledge about the genetic and evolutionary features of HoBi-like viruses and provide some information for vaccine research.

Aging-related inflammation is tightly linked with the development of osteoarthritis (OA). As the pro-inflammatory cytokine, IL-1β has been associated with physical dysfunction and frailty. It is still elusive whether and how IL-1β blockade improves the outcome of OA. Here we develop a cationic solid lipid nanoparticles (SLNs) system that effectively mediate non-viral delivery of plasmid DNA (pDNA) into cells. Compared with other DNA transfer technologies including lipofetamin 2000, SLNs-pDNA system is less toxic and exerts identical effectiveness on DNA transfer. Loaded with integrin β1 overexpression pDNA, the SLNs-pDNA mainly localized in cytoplasm and enforced expression of integrin β1 in rat chondrocytes. Moreover, upon exposure to IL-1β stimulation, SLNs-pDNA treatment attenuates the apoptosis rat chondrocytes and augments tissue repair. Our data thus demonstrate that SLNs-pDNA functions as a potential therapeutic nanomedicine in the treatment of osteoarthritis.

In the existing robot-assisted rehabilitation training process, the interactive control between the patient and the rehabilitation robot is mainly based on sensing the patient's active “movement”, failing to consider the patient's “psychological” level of active participation, which has certain limitations. To solve this problem, we propose a facial emotion recognition system for exoskeleton rehabilitation robot. This system takes the frustration, excitement and boredom of the patient as the target emotion, and combines LNUS feature points with Gabors wavelet for feature extraction. The radial basis kernel SVM target emotion classifier recognizes the facial emotions of patients, and finally uses the recognized emotions to select the appropriate training difficulty. The experimental results showed that the average recognition rate of the patient's emotions reaches 86.3%, which can effectively recognize the three emotions of the patients, providing a basis for selecting the appropriate training difficulty in the later stage.

A novel virus, designated mink Torque teno virus (mink TTV), was identified in mink. Mink TTV possesses a typical genomic organization that is consistent with those of reported anelloviruses. Phylogenetic analysis indicates that mink TTV is genetically most closely related to, but distinct from, giant panda anellovirus and swine TTV strains. This virus differs significantly from currently known members of the family Anelloviridae and represents a new species. Host-specific adaptation analysis indicates that TTVs are highly adapted to humans and primates, and observed incongruence suggests that the emergence of mink TTV was likely the result of a recent host switch from giant panda into mink. Co-infection detection indicates that mink TTV can potentially enhance the pathogenic effects of Aleutian mink disease virus and mink circovirus to cause disease.

Gastric cancer (GC) originates from the stomach and is a prevalent human malignancy. Dysfunction of death associated protein kinase 1 (DAPK1) has been identified as a major regulator involved in the development and progression of GC. However, there's limited data regarding the regulatory mechanism of GC. Herein, we investigated role of DAPK1 in natural killer (NK) cell killing ability and immune evasion of GC cells and mediated pathway. Samples from GC-related gene expression profile and clinical samples from 67 patients with GC were collected to determine the expression of DAPK1, IκB kinase β (IKKβ), programmed death receptor-ligand 1 (PD-L1), and photomorphogenesis 9 (COP9) signalosome 5 (CSN5). The binding affinity among DAPK1, IKKβ, CSN5, and PD-L1 was characterized to verify the underlying mechanism. GC lines were transfected with overexpressed plasmid or siRNA to determine the effect of DAPK1/IKKβ/CSN5/PD-L1 axis on NK cell killing ability and immune evasion of GC cells. GC cells and tissues presented low expression of DAPK1 and high expression of IKKβ, CSN5 and PD-L1. IKKβ, negatively regulated by DAPK1, was capable of activating CSN5 and upregulating PD-L1 expression. Overexpression of DAPK1 promoted NK cell killing ability and reduced immune evasion, coupled with reduction of NK cell apoptosis and increases in levels of TNF-α, IFN-γ, CD107a, and Granzyme B cytokines. The tumor-suppressing properties of DAPK1 through downregulation of IKKβ/CSN5/PD-L1 axis in GC were further confirmed in vivo. In summary, overexpression of DAPK1 promoted the NK cell killing ability and restrained immune evasion of GC cells, providing a potential therapeutic strategy for GC treatment by modulating immune evasion.

Engineering Sb-based anode materials is the key to enhance their electrochemical performance for sodium ion batteries (SIBs) by solving the issues of the rapid capacity decay and poor rate capability. In this work, a nanoporous NiSb alloy (np-NiSb) with a three-dimensionally interconnected ligament-channel structure was synthesized by a facile dealloying strategy. As an anode for SIBs, the np-NiSb alloy exhibits excellent cycling performance, rate capability and stability with a reversible capacity of 334.6 mAh g − 1 at 0.2 A g − 1 after 100 cycles, 155.6 mAh g − 1 at 20 A g − 1 and a capacity retention rate of 97% after 100 cycles at 1 A g − 1. The nanoporous structure and the introduction of inactive Ni effectively tolerate the dramatic volume changes during the charge/discharge processes, restraining the pulverization of np-NiSb. The unique ligament-channel network structure with an average size of about 30 nm significantly shortens the ion transmission distance, ensuring the fast charge transfer at high rates. Operando X-ray diffraction reveals the sodiation/desodiation mechanism of the np-NiSb anode during the discharge/charge processes. In addition, on-line differential electrochemical mass spectrometry further explores the reaction mechanism of np-NiSb. This work highlights constructing nanoporous Sb-based alloys as an effective strategy to improve the performance of SIBs.

In the electro-catalytic degradation process of phenol wastewater, bubbles and mass transfer limitation will result in the decrease in wastewater degradation efficiency, a long electrolysis time and a high energy consumption. Self-made Ti/IrO 2 -Ta 2 O 5 anodes and a high-gravity electro-catalytic reactor were used to improve them. The Ti/IrO 2 -Ta 2 O 5 anode was prepared with a thermal decomposition method and characterized by scanning electron microscopy (SEM). Under optimum conditions, the removal efficiencies of phenol, total organic carbon (TOC) and chemical oxygen demand (COD) respectively reached 94.77%, 50.96% and 41.2% after 2 h electrolysis in the high-gravity field, which were respectively 10.93%, 16.72% and 24.84% higher than those in the normal gravity field. For about the same removal efficiencies, the electrolysis time and energy consumed in the high-gravity field were 33.3% and 15.4% lower than those consumed in the normal gravity field, respectively. The degradation pathway of phenol detected by high performance liquid chromatography (HPLC) was unchanged in the high-gravity field, but the degradation rate of phenol increased. The Ti/IrO 2 -Ta 2 O 5 anode provided good stability because the removal efficiencies of phenol and TOC decreased slightly and the surface morphology of the coating was almost unchanged when it had been used in electrolysis for 11 months, about 1,200 h, in the high-gravity field. Results indicated that the phenol wastewater degradation efficiency was improved, the time was shortened, and the energy consumption was reduced in the high-gravity field.

Background and objectives The knee is innervated by the femoral, obturator, and sciatic nerves. An infrainguinal fascia iliaca compartment block (FICB) is often used as a technique for pain management after hip and knee arthroplasty. This approach blocks the femoral nerve, lateral femoral cutaneous nerve, and obturator nerve. Previous studies show suprainguinal FICB achieves improved postoperative analgesia compared with infrainguinal FICB after hip fracture. However, the analgesic effects of suprainguinal or infrainguinal FICB on the obturator nerve after total knee arthroplasty (TKA) remain to be established. This study compared the efficacy of suprainguinal versus infrainguinal FICB for the blockade of the obturator nerve using electromyography and quantification of total opioid consumption during the 24 hours after TKA. Methods This prospective, randomized controlled clinical study enrolled 74 patients scheduled to undergo TKA. Patients were randomized 1:1 to receive suprainguinal FICB (group S) or infrainguinal FICB (group I) with 30 mL of 0.375% ropivacaine. The primary endpoint was the mean amplitude of the adductor longus compound muscle action potential (CMAP) at 0 (before the block), 10, 20, and 30 minutes after FICB. The secondary endpoint was total opioid consumption during the 24 hours after TKA. Results Data from 62 patients were included in the analysis. The mean amplitude of the adductor longus CMAP was significantly lower in group S compared with group I (repeated-measures analysis of variance; F=4.73, P=0.034). At 24 hours after TKA, mean (SD) total opioid consumption was significantly lower in group S, compared with group I (131.5±76.8 vs. 201.5±85.1 μg) (P=0.001). Conclusions Suprainguinal FICB significantly increased the incidence of successful obturator nerve block and significantly decreased fentanyl consumption 24 hours after TKA compared with infrainguinal FICB.

Background Prefoldin (PFDN) subunits have recently been found to function importantly in various tumor types, while the role of PFDN subunit 1 (PFDN1) in gastric cancer (GC) remains largely unknown. Herein, we aimed to investigate the clinical significance, the biological role and the underlying mechanism of PFDN1 in GC development. Materials and Methods PFDN1 expression levels were measured in human GC specimens by quantitative real-time PCR (qRT-PCR), Western blot and immunohistochemistry. Furthermore, the effects of aberrant PFDN1 expression on GC cells behavior were assessed by wound-healing assay and transwell assay in vitro, and metastasis assay in nude mice, as well as Wnt/β-catenin signaling-induced epithelial–mesenchymal transition (EMT)-related markers by qRT-PCR and Western blot. Results PFDN1 levels were significantly upregulated in GC tissues compared with those in matched adjacent normal tissues. PFDN1 upregulation correlated strongly with clinical metastasis and unfavorable prognosis for GC patients. In vitro and in vivo studies revealed that PFDN1 facilitated GC cell migration, invasion and metastasis. Mechanically, PFDN1 modulated GC cell behavior by activating Wnt/β-catenin signaling-mediated EMT. Conclusion These results suggested a central role of PFDN1 in GC metastatic development via the Wnt/β-catenin pathway, thus providing a potential therapeutic target for patients with GC.

Criminal justice has always been a focus of China’s legal reform, and a serious problem in China’s criminal justice system is police torture and coercive confessions. Almost all of the wrongful convictions that China has identified involve torture of some sort and involuntary confessions. To prevent wrongful convictions, China has implemented exclusionary rules against illegally obtained evidence since 2010, among other mechanisms. Part I of this article provides a historical overview of the problem of torture and anti-torture efforts in China. Part II traces the evolution of exclusionary rules in China and discusses various factors leading to their evolution. Part III evaluates the degree to which exclusionary rules have been implemented in practice based on empirical findings, and analyses the impact of such implementation on both judicial authorities and defence lawyers. Part IV introduces and comments on a series of legislative efforts to address difficulties in implementing the exclusionary rules, including the recently released implementing guideline. Part V puts forward some proposals for further reforms for stricter implementation of exclusionary rules.

After years of efforts, the exclusionary rules have finally found their foothold in Chinese statutory law. However, the initial reform fervour has given way to a difficult slog of changing actual practice since the rules were officially established. This article is based on a comprehensive empirical survey of the implementation of the exclusionary rules. The author found that eforts to implement the new rules are encountering significant challenges. It is usually difficult to get claims of torture raised in court proceedings, to obtain genuine investigations by the procuratorate, and to get illegally obtained evidence excluded in court. In the cases resulting in exclusion of illegally obtained evidence, very few led to acquittal. There were multiple contributing factors to courts’ failure to implement the rules. The most important reason is the current legislation did not provide clear and readily applicable guidance for practitioners to implement the new rules. Other reasons include the absence o...

The exclusion of illegally obtained evidence has long been the focus of theoretical research and legislative reform in China. After years of efforts, the exclusionary rule has finally found a foothold in Chinese statute. However, after the exclusionary rule has been officially established, the initial fervour for reform has given way to a difficult slog of changing actual practice. This article is based on a comprehensive empirical survey on the implementation of the exclusionary rule conducted by the author as the primary investigator. This article will address three key issues that stood out in the empirical surveys: the definition and scope of illegally obtained confession, proof of illegally obtained confession, and suppression hearing. In addressing each issue, the author will follow the similar structure: first share the empirical findings on the implementation of the exclusionary rule across the country, then examine the contributing factors causing the failure of implementation and identify the existing challenges that China encountered in implementing the new rules, and finally put forward some potential solutions to these problems based on comparative study and China’s special situation.

From the point of telecommunication service providing, the paper analyzed three critical factors which help Huawei become the largest provider in Burma. They are models of operation, modern project management method, and corporation culture. Burmese project is only a microcosm of Huawei’s overseas business. It is also a perfect interpretation of the overseas development history of a Chinese private enterprise that is full of fighting spirit.

Based on the statistical data, the forecast for housing price of Xi’an city is made by identifying 13 different kinds of indicators. The multi-variable regression model and SPSS are used to analyze the data in linear and non-linear way respectively. R2 for both methods are over than 0.9. So we can get the similar conclusion that housing price in Xi’an will not increase dramatically recently and keep stable.

This article examines the legislation and practice of compulsory treatment in China. Part I traces the Chinese history of criminal commitment law, explains the research methodology, and highlights some general empirical findings. Part II provides a comprehensive empirical analysis of compulsory treatment law in China, it covers both substantial issues such as criteria of compulsory treatment and procedural issues such as the commitment hearing, enforcement, and discharge of compulsory treatment. It also explores the compulsory treatment law from the human rights protection perspective. Our primary objective is to present the empirical findings to enable the legislative and other involved government agencies to make informed decisions about the future evolution of Chinese law in this area.

China has adopted a series of reforms to encourage witnesses to testify in court in amending its Criminal Procedure Law (CPL) in 2012. However, empirical surveys have revealed that these and subsequent reforms have failed to enhance the rate of witness appearances in criminal trials. This article provides an overview of the history of Chinese criminal procedure law addressing live witness issues in Part I. Part II analyzes the reasons that the reforms have been unsuccessful, suggesting that they were overly focused on witness's motivations and missed institutional resistance to live testimony. Since police officers were not required by law to testify in court until 2010, and police are considered a special subgroup of witness in China, police witnesses are examined separately in Part III. In Part IV, recommendations for further reform are put forward to remedy past failures.

With the rapid depletion of fossil reserves, unstable prices of fossil commodities and the negative effect of fossil fuel on the environment, the demand for clean energy power generation increases. Proton exchange membrane (PEM) fuel cells are a prominent technology for power generation in a variety of applications such as fuel cell vehicles, portable power supplies, stationary power stations and submarines. With their high power density, low emission, low operating temperature, high efficiency, noiseless operation and simplicity of design and operation, PEM fuel cells have a number of attractive advantages over traditional power sources. In this paper, the degradation mechanism of the catalyst layer of the membrane electrode assembly and the core component of the proton exchange membrane fuel cell are summarized. Besides, the degradation mechanism of the Pt-based catalyst, carbon support and Nafion ionomer are obtained. Considering the agglomeration and growth of Pt catalyst, Pt loss and migration, active site pollution and other micro processes, the degradation of catalyst is analyzed in depth. Furthermore, the carbon corrosion process is studied under the normal operation, fuel starvation of the anode and partial hydrogen coverage of the anode processes. In addition, the chemical degradation of Nafion ionomer and the effect of the length of the side chain to the electrode structure are discussed. Eventually, the improved strategies for the failure of the catalyst layer of the membrane electrode assembly are proposed, in the hope to provide useful guidance for the development of long-life proton exchange membrane fuel cells in commercialization.

Municipal solid waste (MSW) leachate contains various refractory pollutants that pose potential threats to both surface water and groundwater. This paper established a novel catalytic oxidation process for leachate treatment, in which OH is generated in situ by pumping both H2 and O2 in the presence of Pd catalyst and Fe2+. Volatile fatty acids in the leachate were removed almost completely by aeration and/or mechanical mixing. In this approach, a maximum COD removal of 56.7% can be achieved after 4 h when 200 mg/L Fe2+ and 1250 mg/L Pd/Al2O3 (pH 3.0) are used as catalysts. After oxidation, the BOD/COD ratio in the proposed process increased from 0.03 to 0.25, indicating that the biodegradability of the leachate was improved. By comparing the efficiency on COD removal and economical aspect of the proposed Pd-based in-situ process with traditional Fenton, electro-Fenton and UV-Fenton for leachate treatments, the proposed Pd-based in-situ process has potential economic advantages over other advanced oxidation processes while the COD removal efficiency was maintained.

In this article, we investigate the performance bottleneck of existing deep learning (DL) systems and propose DLBooster to improve the running efficiency of deploying DL applications on GPU clusters. At its core, DLBooster leverages two-level optimizations to boost the end-to-end DL workflow. On the one hand, DLBooster selectively offloads some key decoding workloads to FPGAs to provide high-performance online data preprocessing services to the computing engine. On the other hand, DLBooster reorganizes the computational workloads of training neural networks with the backpropagation algorithm and schedules them according to their dependencies to improve the utilization of GPUs at runtime. Based on our experiments, we demonstrate that compared with baselines, DLBooster can improve the image processing throughput by 1.4× – 2.5× and reduce the processing latency by 1/3 in several real-world DL applications and datasets. Moreover, DLBooster consumes less than 1 CPU core to manage FPGA devices at runtime, which is at least 90 percent less than the baselines in some cases. DLBooster shows its potential to accelerate DL workflows in the cloud.

Quasars that exhibit blue-shifted, broad absorption lines (BAL QSOs) are an important probe of black hole feedback on galaxy evolution. Yet the presence of BALs is also a complication for large, spectroscopic surveys that use quasars as cosmological probes because the BAL features can affect redshift measurements and contaminate information about the matter distribution in the Lyman-$\alpha$ forest. We present a new BAL QSO catalog for quasars in the Sloan Digital Sky Survey (SDSS) Data Release 14 (DR14). As the SDSS DR14 quasar catalog has over 500,000 quasars, we have developed an automated BAL classifier with a Convolutional Neural Network (CNN). We trained our CNN classifier on the C IV $\lambda 1549$ region of a sample of quasars with reliable human classifications, and compared the results to both a dedicated test sample and visual classifications from the earlier SDSS DR12 quasar catalog. Our CNN classifier correctly classifies over 98% of the BAL quasars in the DR12 catalog, which demonstrates comparable reliability to human classification. The disagreements are generally for quasars with lower signal-to-noise ratio spectra and/or weaker BAL features. Our new catalog includes the probability that each quasar is a BAL, the strength, blueshifts and velocity widths of the troughs, and similar information for any Si IV $\lambda 1398$ BAL troughs that may be present. We find significant BAL features in 16.8% of all quasars with $1.57 < z < 5.56$ in the SDSS DR14 quasar catalog., Comment: 15 pages, 14 figures, submitted to ApJ

Aluminum nitride (AlN) precipitates and microstructure of 4 wt.% (Si+Al) non-oriented electrical steel were investigated. The 2.0 mm thick cast strips with three different silicon/aluminum (Si/Al) ratios were produced by twin-roll casting process, then the strips were reheated, warm rolled, cold rolled and annealed. The microstructure and AlN precipitates were characterized using optical microscopy, scanning electron microscopy and transmission electron microscopy. The results showed that with the increase of Si/Al ratio, on the one hand, the casting microstructure changed from columnar grains to equiaxed grains, and the uniformity of annealed microstructure was improved; On the other hand, the number of AlN precipitates in cast strips reduced meanwhile the distribution became dispersed. By the reheat treatment, the size and distribution of the AlN precipitates can be changed. Moreover, the grain size of the annealed strips is in the range of 20–50 µm, at the same time, many AlN precipitates were located at grain boundaries. Therefore, controlling the Si/Al ratio is a simple method to obtain desired microstructure. Then AlN precipitates in non-oriented electrical steel prepared by twin-roll casting process hinder markedly the recrystallized grains growth. A compatible reheat treatment can be an approach worth exploring to control the behavior of AlN precipitates.

To explore the effect of heat shock protein 90 (HSP90) inhibitor (17-DMAG) and oxaliplatin on the proliferation and invasion of colorectal cancer.After 17-DMAG, oxaliplatin and half-dose combination of 2 drugs processing colorectal cancer SW480 and HCT116 cell lines, CCK8 assay was applied to detect cell viability. RT-PCR and Western blot were used to detect the expression level of the apoptosis-related molecules. Transwell chemokine axis experiment and Western blot were employed to detect cell invasion ability and the expression level of tumor metastasis-associated protein.The growth of SW480 and HCT116 cells was inhibited after the administration of 17-DMAG and oxaliplatin(P0.05) in dose- and time-dependent manner. Processed by 17-DMAG 100 nmol/L, oxaliplatin 50 mg/L and half-dose combination of 2 drugs, transcription level of the apoptosis inhibitory gene (Bcl-2) in SW480 and HCT116 cells was decreased, the level of apoptosis promoting gene (Bax) transcription and protein PARP-1 spliceosome expression was increased, and the above trend was more obvious when using half-dose combination of 2 drugs. Transwell chemokine axis experiments showed the penetrating relative percentage and expression level of MMP9 and integrin β3 decreased, especially for half-dose combination of 2 drugs.17-DMAG and oxaliplatin can co-inhibit the proliferation and invasion of colorectal cancer.

Query by humming (QBH) allows users to retrieve songs by humming a clip. In the previous work, the query has been regarded as a fragment of the music, so the task of QBH is considered to find a subsequence, which is most similar to the whole query, from the database. Taking into account humming errors, especially at the beginning or ending of the query, we assume that only part of the query is a subsequence of the music. Based on this assumption, we propose a local alignment framework which searches for the best match common subsequence between the query and database music. To verify the effectiveness of local alignment, two popular match algorithms, i.e. Linear Scaling and Dynamic Time Warping, are extended to identify the common subsequence. Experimental results on the 2010 MIREX-QBH corpus show that the new algorithms improve the retrieval accuracy significantly.

Locality sensitive hashing (LSH) has been introduced as an indexing structure for fast large-scale data retrieval. A significant drawback of this technique is that a large number of hash functions require lots of hash computation time. To resolve this issue, the paper presents a new index construction algorithm called boundary-expanding LSH, where the boundary of each bucket is expanded so that each point may be mapped to multiple adjacent buckets in each hash table. The proposed algorithm can map points near to the two sides of the boundary into the same bucket, which increases the collision probability of neighbors. To achieve the same search accuracy, boundary-expanding LSH needs fewer hash tables and less hash computation time, so it can provide us a higher acceleration factor. To obtain the best performance, the paper also presents a parameter optimization method. Experiments conducted on two audio databases show its efficiency.

The cellular protein tetherin tethers the HIV-1 viral particles on the cellular membrane to inhibit the replication of HIV-1. However, the HIV-1 accessory protein Vpu counteracts the antiviral function of tetherin. In this study, two retroviral vector plasmids were constructed. One inhibited the vpu gene expression; the other one over-expressed the tetherin. Both retroviral vector plasmids could be packaged in the packaging cell line PT67 to obtain the corresponding retroviruses. The retroviral vector plasmids’ functions of tetherin over-expression or vpu-RNAi were detected at the cell level. Retroviral vector plasmids were transfected to PT67 cells at different ratios from 0T3V to 3T0V, and then mixed retroviruses were harvested. The antiviral functions of mixed retroviruses were detected in HIV-1 infected TZM-bl cells. The results showed that packaged mixed retroviruses could repress the replication of HIV-1 in TZM-bl cells.

This paper presents an improved pruning method taking into account of the number of states possessed by hypotheses in some certain frames. With conventional pruning strategy, the hypotheses with a low score or a bad ranking will be discarded. However, it neglects a fact that the hypotheses several states ahead of or behind the right hypothesis in the prefix tree, which should be discarded, have similar scores and rankings with the right hypothesis. If a state is part of a partial path hypothesis, we say it is possessed by the hypothesis. So in a speech frame, we can deduce that the hypotheses which possess the most states and the hypotheses which possess the least states have little chance to be the right hypothesis. The proposed method analysis the range of the number of the states possessed by the hypotheses, and discards the hypotheses that possess too many or too few states. According to the experiments, This method could effectively improve the performance of the ASR.

Query by humming (QBH) is a technique that is used for content-based music information retrieval. It is a challenging unsolved problem due to humming errors. In this paper a novel retrieval method called note-based locality sensitive hashing (NLSH) is presented and it is combined with pitch-based locality sensitive hashing (PLSH) to screen candidate fragments. The method extracts PLSH and NLSH vectors from the database to construct two indexes. In the phase of retrieval, it automatically extracts vectors similar to the index construction and searches the indexes to obtain a list of candidates. Then recursive alignment (RA) is executed on these surviving candidates. Experiments are conducted on a database of 5,000 MIDI files with the 2010 MIREX-QBH query corpus. The results show by using the combination approach the relatively improvements of mean reciprocal rank are 29.7% (humming from anywhere) and 23.8% (humming from beginning), respectively, compared with the current state-of-the-art method.

Nearest neighbor problem has recently been a research focus, especially on large amounts of data. Locality sensitive hashing (LSH) scheme based on p-stable distributions is a good solution to the approximate nearest neighbor (ANN) problem, but points are always mapped to a poor distribution. This paper proposes a set of new hash mapping functions based on entropy for LSH. Using our new hash functions the distribution of mapped values will be approximately uniform, which is the maximum entropy distribution. This paper also provides a method on how these parameters should be adjusted to get better performance. Experimental results show that the proposed method will be more accurate with the same time consuming.

Automatic song identification has long been a research focus. In this paper, a novel structural fingerprint based hierarchical filtering method is proposed and it consists of two parts: one is the generation of fingerprint with both long structural information and low collision, and the other is an efficient searching algorithm based on a set of selective 2-level filters. Experiments conducted on a database of 10,000 songs show that our approach is fast enough and can achieve the accuracy of 99.7% on 5 second clips with the SNR at 0db comparable to the state-of-the-art.