Search

Text-based captchas are still widely used by many websites such as Wikipedia and Microsoft despite the emergence of many alternative captchas. Recently, the design of text-based captchas has become more and more complex to resist attacks from automatic cracking programs. However, most of the existing captcha solving methods have certain shortcomings, such as insufficient accuracy, poor generalization performance, and the need for a large number of labeled samples. This study proposes a fast captcha solver that can effectively break text-based captchas with complex security features using a small amount of labeled data. The solver was achieved by constructing a captcha transformation model based on generative adversarial networks to simplify the captcha images before character segmentation and recognition. Results showed that the proposed captcha solver achieved a high success rate of over 96% character accuracy and 74% captcha accuracy for all evaluated schemes. Moreover, the average time to process a single captcha image using a laptop GPU was only 4–8 ms. The effectiveness of this work may encourage captcha designers to reconsider a more secure human–machine distinction mechanism.

The presence of cavitation in the oil film seriously affects the bearing lubrication performance and bearing capacity. Now the research of this phenomenon mostly focuses on the model of Reynolds equation (R-E equation) or Navier-Stokes equation (N-S), the influence of the two computation models is less analyzed, and the effect of noncondensable gas (NCG) mass fraction on the bearing performance is seldom studied. In the manuscript, the cavitation mechanism is studied using the mixed model of three-dimensional N-S equation and Jakobsson-Floberg-Olsson (JFO) condition of two dimensional Reynolds equation, and the influence of rotational speed and NCG mass fraction on the cavitationoil film pressure, and bearing capacity was studied. The results show that the change trend of cavitation with the rotational speed is basically consistent for N-S equation and R-E equation. The bearing capacity calculated by N-S equation is greater than that calculated by R-E equation. The peak pressure and bearing capacity of film can be improved by increasing the NCG mass fraction of lubricant and rotational speed.

In the mechanical tribology research, reasonable and effective improvement of lubrication performance for friction pairs of sleeve bearings is one of the hot topics, surface micro texture processing technology has gradually become an important way to improve the lubrication performance of friction pairs. Taking the friction pair model of infinite wide bearing as the research object, the Reynolds equation on the fluid lubrication of the friction pair is established. According to the static characterization equation of the micro texture friction pair, the influence of the depth and the coverage rate of micro texture on the lubrication performance of the friction pair is studied. According to the theoretical analysis, in order to simulate the actual contact condition of the sleeve bearing friction pair more truly, 45 steel is selected as friction material to prepare the ring–ring contact friction pair. The micro textures of a certain shape are machined on the surface of the specimen by laser micro carving, and the friction and wear experiments are carried out by universal friction tester. Theoretical analysis and experimental results show that the lubrication performance of friction pair with reasonable distribution of micro texture is improved obviously, and the reasonable selection of micro texture depth and coverage rate can effectively improve the friction properties of friction pairs.

Mention-level relation extraction (mRE) plays an important role in extracting relational information from short texts such as those exchanged in a social network. Deep learning (DL) has made remarkable achievements; the main problem encountered with DL in mRE is a lack of training samples. In this paper, we present a design for a quick sample-marking method. First, we construct an uneven noisy labeled data (UNLD) set using a pattern matching algorithm, and then a relabeling framework is put forward for modifying the UNLD. With regard to the accuracy, the recall rates of categories with sufficient samples increased from 0.4 to nearly 1 using the relabeling framework. We have released our code and other resources for further research (https://github.com/curtainsky/UNLD).

Conventional strategies for determining phosphate concentration is limited in efficiency due to the cost, time, and labor that is required in laboratory analysis. Therefore, an on-site and rapid detection sensor for phosphate is urgently needed to characterize phosphate variability in a hydroponic system. Cobalt (Co) is a highly sensitive metal that has shown a selectivity towards phosphate to a certain extent. A disposable phosphate sensor based on the screen-printed electrode (SPE) was developed to exploit the advantages of Co-nanoparticles. A support vector machine regression model was established to predict the concentration of phosphate in the hydroponic solutions. The results showed that Co-nanoparticles improve the detection limit of the sensor in the initial state. Meanwhile, the corrosion of Co-nanoparticles leads to a serious time-drift and instability of the electrodes. On the other hand, the coefficient of variation of the disposable phosphate detection chip is 0.4992%, the sensitivity is 33 mV/decade, and the linear range is 10&minus, 1&ndash, 10&minus, 4.56 mol/L. The R2 and mean square error of the buffer-free sensor in the hydroponic solution are 0.9792 and 0.4936, respectively. In summary, the SPE modified by the Co-nanoparticles is a promising low-cost sensor for on-site and rapid measurement of the phosphate concentration in hydroponic solutions.

The target of personal information extraction (PIE) is to extract content associated with a name form Web pages. Available Web page models, which are also used widely in content extraction and automatic wrapper algorithms, include text model, document object model, and vision-based page segmentation model. Because of existing models focus on Web structure rather than semantic relevance, they are difficult to be directly used for PIE. To deal with this problem, we introduce the sequence block model (SBM), by which is easy to determine the relevance of each page block to the retrieval name. Then, we give the definition of PIE based on the SBM. Depending on the sequence correlation of SBM, we design a 4-layer seq2seq deep learning network for PIE. Experiment result shows that our new model extracts twice as much data as content extraction algorithms. And the recall rate of the network is 7% higher than the traditional model with classification algorithm.

According to the traditional lubrication theory, the smoother the friction pairs the better the friction performance. But in recent years, a large number of theoretical and experimental results have shown that the friction surface with microstructure can effectively improve the tribological performance of friction pairs. For smooth surface and micro-dimple surface with different shapes of friction pairs, the two-dimensional finite element model is established based on the Navier–Stokes calculation equation of the fluid, and the effect of hydrodynamic effects arising from micro-dimple structure with different shapes on friction performance of friction pairs is analyzed. Through friction and wear testing, the friction properties of smooth surface and micro-dimple surface friction pairs are compared and analyzed. Theoretical simulation and experimental verification show that the surface of friction pairs with micropits can obviously improve the lubrication conditions of friction pairs and effectively improve the tribological properties of friction pairs.

With the popularity of social networks, short text clustering has become a more and more important task that is widely used. Short text clustering is a challenging problem because social network short texts are characterized by irregular words, a lot of noise, and sparse features. We propose a Short Text Clustering enhanced by Semantic Matching Model (abbr. to STCSMM). The STCSMM method applies the knowledge of the tagged text similarity task dataset to the short text clustering through the semantic matching model, thereby improving the effect of short text clustering. First, we train a semantic matching network on the data set of the text similarity task, where the network contains the feature extraction layer and the vector distance calculation layer. Then, we use the learned feature extraction layer to extract short text feature and use the vector distance calculation layer replaces the commonly used distance metrics in the traditional K-means algorithm, such as cosine distance, Euclidean distance and so on. Finally, the text features obtained by feature extraction layer are applied to K-means based on vector distance calculation layer. This improved K-means clustering (STCSMM) has better performance on the microblog text clustering dataset than some existing methods such as K-means clustering with LDA, LSI or average word embedding feature vectors.

The wall climbing robot is widely used in a great many fields. There are more and more technical requirements for wall-climbing robots, such as better load, more stability and adapt to multiple environments. This paper studies a composite wall climbing robot palm for a wall climbing robot to adopt multiple environments. The palm consists of a suction cup, five hooks, and bionic setae array materials. The body of the palm is produced by 3D printing and the bionic materials are fabricated by polymer print lithography technology inspired by the gecko. The edge of the suction cup has cured a circle of bionic setae array materials to enhance the adhesion force. The bionic setae array materials on the palm for better adapt to the smooth surface and the hooks are used to attach on the rough surface. The composite palm is well tested in smooth, wet and rough surfaces. Moreover, the composite palm is fixed on the arm of the wall climbing robot. The arm of the wall climbing robot is proposed by a new type of gear transmission system.

A semantic block is treated as a unit while analyzing the webpage. First, we implement the VTPS algorithm to partition a webpage into semantic blocks. Then, we propose an algorithm to extract the spatial and content features, and then construct the feature vector for each block. Based on these vectors, the SVM learning algorithm is applied to train and classify the various theme-oriented webpage blocks. At last, the classification experiments show the efficiency of this method.

A model of spiral oil wedge hydrodynamic bearing is first established, and a numerical simulation of the fluid region is carried out by Fluent. The influence of inlet pressure and rotation speed on the cavitation area in the oil film divergence region is analyzed, and the cavitation shape, size, and location of the cavitation area are also compared and analyzed; at the same time, the influence of noncondensing gas (NCG) mass fraction on the oil film flow field is studied. The results show that the results of the numerical simulation are in agreement with the experimental results. The change of the inlet pressure and rotational speed can affect the disappearing position of the cavitation, but it has almost no influence on the appearing position of the cavitation. The increase of the inlet pressure can reduce the cavitation effectively, as well as the area of cavitation and number of large volume cavitation. But the increase of the rotational speed promotes the production of cavitation obviously and increases the number and area of cavitation and the possibility of complete cavitation. The NCG content in lubricating oil has a certain influence on the lubrication performance and bearing capacity.

Phishing website is becoming a major threat to the information security in Social Network. The attacks not only lessen the users' trust but also influence the benefit of the third party who develops the platform. In order to solve the time lag in phishing website passive detection, this paper proposes a solution to discover phishing website initiatively based on blacklist, in which the anomalies of its URL and WHOIS information are analyzed, and based on this, the heuristic rules that aim to generate suspicious URLs are made. In order to filter out noise sites in the suspicious set, a website filtering solution based on webpages image-layout is presented. We firstly propose a Ray Scan Method to generate the location feature of webpage images quickly, and then, we proposed a method of calculating the webpage layout similarity, which will be compared against the preset threshold to decide whether it will be filtered. The experimental results show that the solution successfully detects some phishing websites out before they are widely spread, and further, the webpage filtering method guarantees both high filtration ratio and high phishing website retention ratio.

Based on the finite finite difference method, the Reynolds equation, the flow equilibrium relationship of oil film gradient, and the oil film thickness equation are solved for different groove locations and depths in two-axial groove sleeve bearings. The oil film pressure, extent of the cavitation zone, and carrying capacity of two-axial groove sleeve bearings are computed at different oil groove locations and depths. With the increase of oil groove depth, the oil film pressure of two-axial groove sleeve bearings decreases noticeably, and the bearing capacity decreases gradually. It can be concluded that the maximum oil film pressure and bearing capacity of two-axial groove sleeve bearings are lower than that of the common sleeve bearing. With the groove position moving along the circumferential direction, the oil film pressure peak heights and bearing capacities decrease and then increase. The oil film pressure peak heights and load capacities reach maximum when the position of the two-axial oil groove is about 20° and 200°. It has been proven that different positions of oil grooves show different influences on the extent of the cavitation zone. The first oil groove is always located in the full oil film region, and the second oil groove may be located in the full oil film region or the cavitation region because of the change of distance from the convergent wedge of the bearing. The cavitation area firstly decreases and then increases with the movement of the groove in the circumferential direction.