Search

Collaborative Filtering (CF) algorithms have been widely used to provide personalized recommendations in e-commerce websites and social network applications. Among them, Matrix Factorization (MF) is one of the most popular and efficient techniques. However, most MF-based recommender models only rely on the past transaction information of users, so there is inevitably a data sparsity problem. In this article, we propose a novel recommender model based on matrix factorization and semantic similarity measure. Firstly, we propose a new semantic similarity measure based on semantic information in the Linked Open Data (LOD) knowledge base, which is a hybrid measure based on feature and distance metrics. Then, we make an improvement on the traditional MF model to deal with data sparsity. Specifically, the MF process has been extended from both the user and item sides with implicit feedback information and semantic similar items, respectively. Experiments on two real datasets show that our proposed semantic similarity measure and recommender model are superior to the state-of-the-art approaches in recommendation performance.

Mesoporous alumina modified by silica (MA-S) with good thermal stability was prepared by a self-assembly method. The morphology and microstructure of as-prepared samples were characterized by transmission electron microscopy (TEM), X-ray diffraction (XRD), and so on. Though adding 5 wt% SiO2 in Al2O3 framework, the MA-S exhibited high thermal stability and large specific surface area (87 m2 g−1) at 1200 °C. The results indicated that SiO2 can suppress the phase transformation from γ-Al2O3 to α-Al2O3 and postpone the crystallization temperature at same time. Therefore, the as-prepared materials possess a great potential for application at high temperature.

Deep learning (DL) techniques have been widely used in recommender systems for user modeling and matching function learning based on historical interaction matrix. However, existing DL-based recommendation methods usually perform static user preference modeling by using historical interacted items of the user. In this article, we present a time-aware deep CF framework which contains two stages: dynamic user preference modeling based on attention mechanism and matching score prediction based on DL. In the first stage, short-term user preferences are modeled by the time-aware attention mechanism that fully considered the predicted item, the recent interacted items and their interaction time. The resulting short-term preferences are combined with long-term preferences for dynamic user preference modeling. In the second stage, high-order user-item feature interactions are learned by two types of DL models, Deep Matrix Factorization (DMF) and Multiple-Layer Perception (MLP), and the feature interaction vectors of the two models are fused in the last layer of the model to predict the matching score. Extensive experiments on five datasets indicate that our method is superior to the existing time-aware and DL-based recommendation methods in top-k recommendations significantly and consistently.

Modern data centers coordinate hundreds of thousands of heterogeneous tasks aiming at providing highly reliable cloud computing services. Failure prediction is of vital importance in the analysis of cloud reliability. Recently, a novel kernel learning method called relevance vector machine (RVM) has been widely applied to solve nonlinear predicting problems and has been verified to perform well in many situations. However, it remains a great challenge for existing approaches to acquire the optimal RVM parameters. In this research, an artificial immune system is introduced into a Quantum-inspired Binary Gravitational Search Algorithm (QBGSA) in order to improve the convergence rate of standard QBGSA. In addition, a hybrid model of RVM with improved QBGSA called IQBGSA-RVM is proposed that aims to predict the failure time of cloud services. To evaluate the effectiveness of IQBGSA-RVM in failure prediction, its predicting performance is compared with that of the following algorithms, all of which employs RVM: chaotic genetic algorithms, binary gravitational search algorithms, binary particle swarm optimization, quantum-inspired binary particle swarm optimization and standard QBGSA. The experimental results show that the IQBGSA-RVM model is either comparable to the other models or it outperforms them, to say the least.

The traditional collaborative filtering (CF) method based on static user preference modeling and linear matching function learning severely limits the recommendation performance. To solve the above problem, in this article, we adopt dynamic user preference modeling and nonlinear matching function learning in the CF recommendation. For dynamic user preference modeling, a two-layer neural attention network is used, which fully considers the predicted item, the recent historical interacted items and their interaction time to estimate the contribution weight of each interacted item in user preferences modeling. For nonlinear matching function learning, we add a single hidden layer neural network on top of the traditional matrix factorization (MF) model, which can significantly improve the feature interaction learning capabilities of the model with only a few additional parameters. Extensive experiments show that our method significantly outperforms the state-of-the-art CF methods and the key technologies we proposed in this research have a positive effect on improving the recommendation performance.

This paper presents a novel reconfigurable spherical (rS) joint resulting from the line dependence of screws and their embodiment geometry. Subsequently, the structure and phase change of the rS joint are elaborated. With this new invention, two novel metamorphic mechanisms, that are, the Schatz-derived metamorphic mechanism and the R(rS)R-R(rS)R metamorphic mechanism, are obtained by inserting the rS joint. The mechanisms entail a reconfiguration process that occurs in two manners: the phase change of the rS joint and the bifurcation of the mechanism. The Schatz-derived metamorphic mechanism entails reconfiguration by phase change of the rS joint and the bifurcation, the R(rS)R-R(rS)R metamorphic mechanism reconfigures the configuration based on phase change of the two rS joints. Remarkably, the former realizes the switch between two assembly configurations without resorting to disassembly and reassembly, and maintains the characteristics of the original Schatz mechanism while the rS joint situates in phase R. The latter changes the mobility and bridges gap between a spatial single-loop mechanism and a structure. With this new reconfigurable joint, various reconfigurable mechanisms could be generated by inserting the reconfigurable joint into the classical ones.

In this paper, we propose a deep collaborative filtering recommendation model, which consists of an attention-based representation learning component and a multi-input matching function learning component. This model takes interaction matrix based on implicit feedback as data source to construct representations of long-term user preferences and item latent features. In the representation learning, a time-aware attention network is used, which uses the embedding vectors of the predicted item, recent historical interaction items, and the interaction time of recent historical interaction items to estimate the weights of different historical interaction items to short-term user preferences modeling. Then, the dynamic user preference representation can be obtained by combining short-term preferences with long-term preferences. In the matching function learning, a multi-input deep learning model is used. Its input includes not only the dynamic user preference representation and the item latent feature representation, but also the linear interaction between the two representations, so that the model has more powerful feature interactions learning ability. Experimental results on four datasets from different domains show that our method is largely superior to the state-of-the-art collaborative filtering methods, and the novel techniques we propose in this paper are effective in improving recommendation performance.

Herein, N-doped carbon coated vanadium nitride nanowires (VN@NC NWs) with inner porous structure were synthesized by coating a thin layer of a carbon precursor of polydopamine (PDA) on ultralong V2O5 NWs followed by nitridation. The amorphous carbon coating could ensure the one-dimensional (1D) structure, as well as produce the inner pores during nitriding progress. The formational 1D porous architecture, high-crystallized VN nanoparticles and homogeneous N-doped carbon coating synergistically improved the materials' impedance match and thus enhanced their microwave absorption performances. The minimum reflection loss (RLmin) of the optimal samples/paraffin hybrid with 20 wt% was −60.2 dB with an effective absorption bandwidth (EAB, RL

Origami engineering as an interdisciplinary subject brings thriving progress of mechanisms innovation. In this paper, a construction approach from origami to multiple spherical-integrated mechanisms is proposed and subtly sets up the bridge between origami and Bennett linkage, Bricard linkage. Subsequently, the attention is drawn to a special origami-inspired integrated 8R kinematotropic metamorphic mechanism. The paper applies screw theory to illustrating the reconfigurability of singular configurations with distinct parametric constraints in the 8R mechanism. Two overconstrained linkages, a 4R linkage and a 6R linkage, are evolved under peculiar geometrical constraints respectively. Furthermore, the kaleidocycle and deployable properties of the 8R mechanism are revealed. The paper hence not only presents an interesting work extracted from origami but also sheds light on the promising investigation about the intersection of distinct types of reconfigurable mechanisms. Finally, by using the kinematotropic metamorphic 8R mechanism as the reconfigurable trunk, a novel quadruped robot is built and its characteristics will be investigated in future research.

A novel quinone-modified metal-organic frameworks NH2-MIL-101(Fe) was synthesized using a simple chemical method under mild condition. The introduced 2-anthraquinone sulfonate (AQS) can be covalently modified with NH2-MIL-101(Fe) and acts as a redox mediator to enhance the degradation of bisphenol A (BPA) via persulfate activation. The obtained AQS-NH-MIL-101(Fe) was characterized by Fourier transform infrared spectroscopy, X-ray diffraction, scanning electron microscopy, X-ray photoelectron spectra, cyclic voltammetry and electrochemical impedance spectroscopy. AQS-NH-MIL-101(Fe) exhibited better catalytic performance compared with NH2-MIL-101(Fe) and NH2-MIL-101(Fe) with free AQS (NH2-MIL-101(Fe)/AQS). That is, AQS-NH-MIL-101(Fe) was proved to be the most effective in that more than 97.7% of BPA was removed. The degradation rate constants (k) of AQS-NH-MIL-101(Fe) was 9-fold higher than that of NH2-MIL-101(Fe) and 7-fold higher than NH2-MIL-101(Fe)/AQS, indicating that AQS is a great electron-transfer mediator when modified with NH2-MIL-101(Fe). Based on the above results, the possible mechanism of catalytic reaction has been investigated in view of the trapping experiments. In addition, the AQS-NH-MIL-101(Fe) catalyst exhibited excellent stability and can be used several times without significant deterioration in performance.

Many abundant and inexpensive materials are being developed for the synthesis of carbon dots. However, the structures of most raw materials are complicated, which inevitably bring difficulties for the optimization of their photocatalytic ability, especially for the electron transfer properties. Herein, petroleum coke was selected as raw material to prepare carbon dots (CDs). The types and contents of the functional groups C O and/or S O) on CDs surface were finely tuned by facile chemical oxidation, enhancing both of the electron-accepting (11.9 and 3.5 times) and electron-donating abilities (1.7 and 1.4 times). The photocatalytic efficiency of the optimized carbon dots (C-120) in the photooxidation of 1,4-dihydro-2,6-dimethylpyridine-3,5-dicarboxylate (1,4-DHP) is 6.6 times higher than that of CDs with few functionalized groups. The present study demonstrates a feasible and effective strategy to solve the poor electron transfer difficulty of CDs and also helps to understand the regulating strategies and mechanisms of surface structure.

Fe-based metal–organic frameworks (MOFs) including MIL-101(Fe), MIL-100(Fe), MIL-53(Fe), and MIL-88B(Fe) prepared via a facile solvothermal process were introduced as both adsorbents and catalysts to generate powerful radicals from persulfate for acid orange 7 (AO7) removal in aqueous solution. Various catalysts were described and characterized by X-ray diffraction, Fourier transform infrared spectroscopy, scanning electron microscopy and X-ray photoelectron spectra. Because of the high specific surface area of the materials, we studied the adsorption isotherms of the four MILs by the fitting of Langmuir adsorption isotherm. Meanwhile, the catalytic activities in persulfate oxidation system were investigated. The results showed that the sequence of the materials ability in the combination of adsorption and degradation was MIL-101(Fe) > MIL-100(Fe) > MIL-53(Fe) > MIL-88B(Fe), which had a close connection with the activity of metal ion in active site of the catalysts and their different cages in size. Moreover, the reactive species in MILs/persulfate system were identified as sulfate radicals and hydroxyl radicals. The reaction mechanism for persulfate activation over MILs was also studied.

Photooxidation plays an important role in organic synthesis. However, deep oxidation and degradation could reduce the product selectivity and the final yield, which is a crucial issue in the photooxidation reactions. Polymeric carbon nitride (C3N4) shows great potentials in the photocatalysis, but most improved methods make C3N4 show strong oxidation ability, which limits their applications in the selective oxidation synthesis. In order to overcome the limits and enhance the photooxidation efficiency, terephthalic acid was used to change the monomer structure of carbon nitride (P-CN). The physical and electrochemical properties of P-CN were studied with TEM, SEM, XRD, XPS, BET, NMR, ESR and theoretical calculations. The conduction band (CB) position of P-CN is reduced while the valance band (VB) is almost unchanged, which benefits for enhancing visible light response, and preventing the deep oxidation or degradation. At the same time, the structure defects (unreacted amino in the copolymerization process of g-C3N4) are reduced. As a result of these enhanced properties, the photocatalytic efficiency of P-CN in the photooxidation reaction of 1,4-dihydro-2,6-dimethylpyridine-3,5-dicarboxylate (1,4-DHP) is improved 17-fold compared with g-C3N4. More importantly, the pure product could be obtained by simple filtration without any purification, and there is no obvious deep oxidation or degradation.

Brain computer interface (BCI) is a novel technology that translates human intention into command to control external device. Common spatial pattern (CSP) algorithm is most frequently applied for feature engineering in motor imagery (MI) based BCI system. How to select the most suitable spatial channels, temporal & frequency parameters for different people before CSP is still a challenging issue which greatly affects the performance of MI based BCI system. In this paper, we introduce an adaptive multi-domain feature joint optimization framework. Specifically, random forest (RF) and composite kernel support vector machine (CKSVM) algorithms are used to measure the significances of different spatial channels and local temporal-frequency segments. An ant colony optimization (ACO) based scheme is proposed to search the most suitable spatial channels and temporal-frequency segments. We evaluated the effectiveness of the proposed algorithm on public BCI competition III data set IVa and two self-collected MI EEG datasets. For BCI competition III data set IVa, our method outperforms some other close related algorithms in the literature. For the two self-collected datasets, compared to the traditional manual parameter setting, the classification performance is proven to significantly improve (more than 15%) adopting our adaptive multi-domain parameters. Since our proposed method can simultaneously and automatically optimize subject-specific features in the entire spatial-temporal-frequency domains, the most discriminative CSP features can be selected and the performance of MI EEG classification is significantly improved. Thus, our research is a useful complement to the BCI field.

High-throughput and low-cost methods for the fabrication of 2D patterned structures are highly desirable. Herein we proposed a new colloidal lithography method using 2D microgel colloidal crystal (CC) as template. Unlike the previously developed methods in which the microgel spheres act as lithography mask, herein they act as nanoreactors to convert a precursor into the target product. As an example, highly ordered 2D CC of poly(N-isopropylacrylamide-co-acrylic acid) (P(NIPAM-AA)) microgel was prepared on charge-reversible silicon wafer. The template was first treated with trimethylchlorosilane (TMCS) and then loaded with Fe(NO3)3. Finally the polymer was removed and Fe(NO3)3 was converted in situ to Fe2O3 by calcination, and an ordered Fe2O3 array was obtained. The composition left after calcination was confirmed to be Fe2O3 by XPS. TMCS treatment was demonstrated to be necessary to obtain arrays of discrete patches, instead of continuous film. 0.2 mM Fe(NO3)3 solution was found to be optimal for Fe(NO3)3 loading. A calcination temperature of 600 °C was high enough to remove the polymeric materials and convert Fe(NO3)3 to Fe2O3. The method not only allows constructing ordered structures in a simple and cost-efficient way, but also adjusting the parameters of the pattern by adjusting the parameters of the template.

Recently, deep learning techniques have been widely used in recommendation tasks and have attained record performance. However, the input quality of the deep learning model has a great influence on the recommendation performance. In this work, an efficient and effective input optimization method is proposed. Specifically, we propose an integrated recommendation framework based on two-stage deep learning. In the first stage, with user and item features as the original input, a low-cost marginalized stacked denoising auto-encoder (mSDA) model is used to learn the latent factors of users and items. In the second stage, the resulting latent factors are combined and used as input vector to the DNN model for fast and accurate prediction. Using the latent factor vector as the input to the deep learning-based recommendation model not only captures the high-order feature interaction, but also reduces the burden of the hidden layer, and also avoids the model training falling into local optimum. Extensive experiments with real-world datasets show that the proposed model shows much better performance than the state-of-the-art recommendation methods in terms of prediction accuracy, parameter space and training speed.

Combination of three-dimensional (3D) nanoarchitectures and good electrical conductive performance is charming for the application of graphene-based composite. Herein, 3D lamellar composite of vanadium nitride nanowires encapsulated in reduced graphene oxide (VN NWs-rGO), in which interconnected VN NWs as skeletons with a porous structure were encapsulated in N-doped rGO nanosheets to form the lamellar structure, was synthesized by combination of freeze-casting and nitridation techniques. The as-prepared lamellar composite exhibited excellent EM wave absorbing properties: the min reflection loss (RL) value of the VN NWs-rGO/paraffin hybrid with a filler loading of 15 wt% was −41.5 dB at a thickness of 1.5 mm, with an effective absorption bandwidth (EAB, RL

Broadly neutralizing antibodies have been shown promise as prophylactic and therapeutic agents to provide passive protection against human immunodeficiency virus type 1 (HIV-1) infection. Such protein based microbicides are traditionally produced using mammalian cell expression system, which is not able to satisfy the increasing demand of these proteins in large scale. In this report, two of HIV-1 broadly neutralizing antibodies, b12 and VRC01, were successfully expressed in Sf9 insect cells by co-infection with baculoviruses respectively expressing the light and heavy chain of the antibodies. The purified antibodies are fully assembled as H2L2 (two heavy chains plus two light chains) heterodimer linked by covalent bonds. The b12 and VRC01 generated from insect cells reacted well to HIV-1 gp120, and their antigen binding ability is comparable to the mammalian cell-derived b12 as determined by ELISA and flow cytometry. Our data suggest that baculovirus/insect cell expression system can be utilized as an alternative to the mammalian expression system for the rapid production of HIV-1 broadly neutralizing antibodies.