Search

The blooming of live comment videos leads to the need of automatic live video comment generating task. Previous works focus on autoregressive live video comments generation and can only generate comments by giving the first word of the target comment. However, in some scenes, users need to generate comments by their given prompt keywords, which can't be solved by the traditional live video comment generation methods. In this paper, we propose a Transformer based non-autoregressive conditional masked live video comments generation model called CMVCG model. Our model considers not only the visual and textual context of the comments, but also time and color information. To predict the position of the given prompt keywords, we also introduce a keywords position predicting module. By leveraging the conditional masked language model, our model achieves non-autoregressive live video comment generation. Furthermore, we collect and introduce a large-scale real-world live video comment dataset called Bili-22 dataset. We evaluate our model in two live comment datasets and the experiment results present that our model outperforms the state-of-the-art models in most of the metrics.

Using oxygen reduction for the photocatalytic production of hydrogen peroxide (H

Machine comprehension of texts often requires external common knowledge and coreference resolution in the passage. However, most current machine reading comprehension models only incorporate external common knowledge. We propose CoSp model, which incorporates both common knowledge and specific entity linking knowledge for machine reading comprehension. It employs an attention mechanism to adaptively select relevant commonsense and lexical common knowledge from knowledge bases, then it leverages the relational-GCN for reasoning on the entity graph, which is constructed by the entity coreference and co-occurrence for each passage. Hence we obtain knowledge-aware and coreference-aware contextual word representation for answer prediction. Experimental results indicate that CoSp model offers significant and consistent improvements over BERT, outperforming competitive knowledge-aware models on ReCoRD and SQuAD1.1 benchmarks.

Reverse engineering, repackaging and misuse of Android apps are becoming more and more widespread. Many obfuscation schemes are not enough against sustainably evolving reverse engineering techniques, while others have performance problems or limitations. In this paper, we propose a novel scheme that redirects method invocation dynamically to hide the actual control flow of the program. Firstly, some pivotal methods are replaced by others unrelated methods, so that the actual semantics of the program no longer appear. Then by modifying the entry address of the methods in the obfuscated code, the methods execution automatically jumps to the corresponding target methods without any code modification. In order to accurately restore the control flow, the encrypted mapping is decrypted using dynamic passwords from the server in the standalone Java interpreter. Since the obfuscated program semantics and program execution results can not be significantly different, it is almost impossible to reveal the real logic of the program through static analysis. The scheme can also effectively increase the difficulty of dynamic analysis. The evaluation results show that the semantic redirection obfuscation scheme can well hide the actual control flow of the program and both the time and space consumption introduced are acceptable.

As the applications on smartphones and tablets are getting richer and more powerful, people are more willing to do their daily works and entertainments on mobile devices. Sharing and collaborating on these devices in a convenient and safe way has become a growing demand in modern life. However, due to the lack of a flexible access control strategy of the official online services, typical applications such as email, video and music do not support this kind of sharing scheme. In this paper, we propose TrustyShare to provide a general sharing solution based on the ARM TrustZone technology which has a extremely wide range of applications in mobile terminal devices and Internet of Things devices. Our solution allows users to share services they enjoy without the official support from these services, while avoiding the disclosure of their private credentials. In addition, credential owners have full control on how and when others can use their services through flexible access control policies. We implement our solution on OP-TEE platform, and our experimental results demonstrate that TrustyShare works well with good performance.

WireGuard, a novel VPN proposed in 2017, has been widely accepted in the industry, since it is simpler, leaner and more efficient than traditional VPNs. However, WireGuard ignores the risks of key stolen and abused on both desktop and mobile platform. These vulnerabilities might be utilized by attackers to sneak into the protected network through VPN tunnel provided by WireGuard easily without permission. It is especially obvious on mobile platforms, since they are mostly online all day and the attack scenarios are more complicated and changeable. In this paper, we propose an enhanced WireGuard Android version called SeWG based on TEE. We use TEE (Trusted Execution Environment) technology to achieve secure key storage, preventing the misuse of the private key. And we design corresponding dynamic authentication mechanisms for user mode and kernel mode respectively, to prevent malware from sneaking into the secure tunnel. Finally, we implement SeWg scheme on QSEE, the TEE platform developed by Qualcomm. Our experimental results demonstrate that SeWG can work well with high efficiency.

The fusion peptides (FP) play an essential role in fusion of viral envelope with cellular membranes. The location and properties of the FPs in the spike (S) glycoproteins of different coronaviruses (CoV) have not yet been determined. Through amino acid sequence analysis of S proteins of representative CoVs, we identified a common region as a possible FP (pFP) that shares the characteristics of FPs of class I viral fusion proteins, including high Ala/Gly content, intermediate hydrophobicity, and few charged residues. To test the hypothesis that this region contains the CoV FP, we systemically mutated every residue in the pFP of Middle East respiratory syndrome betacoronavirus (MERS-CoV) and found that 11 of the 22 residues in the pFP (from G953 to L964, except for A956) were essential for S protein-mediated cell-cell fusion and virus entry. The synthetic MERS-CoV pFP core peptide ( 955 IAGVGWTAGL 964 ) induced extensive fusion of liposome membranes, while mutant peptide failed to induce any lipid mixing. We also selectively mutated residues in pFPs of two other β-CoVs, severe acute respiratory syndrome coronavirus (SARS-CoV) and mouse hepatitis virus (MHV). Although the amino acid sequences of these two pFPs differed significantly from that of MERS-CoV and each other, most of the pFP mutants of SARS-CoV and MHV also failed to mediate membrane fusion, suggesting that these pFPs are also the functional FPs. Thus, the FPs of 3 different lineages of β-CoVs are conserved in location within the S glycoproteins and in their functions, although their amino acid sequences have diverged significantly during CoV evolution. IMPORTANCE Within the class I viral fusion proteins of many enveloped viruses, the FP is the critical mediator of fusion of the viral envelope with host cell membranes leading to virus infection. FPs from within a virus family, like influenza viruses or human immunodeficiency viruses (HIV), tend to share high amino acid sequence identity. In this study, we determined the location and amino acid sequences of the FPs of S glycoproteins of 3 β-CoVs, MERS-CoV, SARS-CoV, and MHV, and demonstrated that they were essential for mediating cell-cell fusion and virus entry. Interestingly, in marked contrast to the FPs of influenza and HIV, the primary amino acid sequences of the FPs of β-CoVs in 3 different lineages differed significantly. Thus, during evolution the FPs of β-CoVs have diverged significantly in their primary sequences while maintaining the same essential biological functions. Our findings identify a potential new target for development of drugs against CoVs.

The spike glycoprotein (S) of murine coronavirus mouse hepatitis virus (MHV) strain A59 uses murine carcinoembryonic antigen-related cell adhesion molecule 1a as its receptor for cell entry, but S protein can also be triggered in the absence of receptor by pH 8.0 alone at 37°C. The mechanism by which conformational changes of this S glycoprotein can be triggered by pH 8.0 has not yet been determined. Here, we show that MHV-A59 S protein is triggered by pH 8.0 at 37°C to induce receptor-independent syncytium (RIS) formation on 293T cells, and that the conformational changes in S proteins triggered by pH 8.0 are very similar to those triggered by receptor binding. We systemically mutated each of 15 histidine residues in S protein and found that H209 is essential for pH 8.0-triggered RIS formation, while H179, H441, H643, and H759 also play important roles in this process. Replacement of H209 with Ala had no effect on receptor binding, but in murine 17Cl.1 cells mutant H209A MHV-A59 showed delayed growth kinetics and was readily outcompeted by wild-type virus when mixed together, indicating that the H209A mutation caused a defect in virus fitness. Finally, the H209A mutation significantly increased the thermostability of S protein in its prefusion conformation, which may raise the energy barrier for conformational change of S protein required for membrane fusion and lead to a decrease in virus fitness in cell culture. Thus, MHV-A59 may have evolved to lower the stability of its S protein in order to increase virus fitness. IMPORTANCE Enveloped viruses enter cells through fusion of viral and cellular membranes, and the process is mediated by interactions between viral envelope proteins and their host receptors. In the prefusion conformation, viral envelope proteins are metastable, and activation to the fusion conformation is tightly regulated, since premature activation would lead to loss of viral infectivity. The stability of viral envelope proteins greatly influences their activation and virus fitness. Here, we report that, similar to the A82V mutation in Ebola glycoprotein, in the S glycoprotein of murine coronavirus MHV-A59, the histidine residue at position of 209 significantly affects the thermal stability of the S protein, determines whether S protein can be activated at 37°C by either pH 8.0 alone or by receptor binding, and affects viral fitness in cell culture. Thus, the spike glycoprotein of MHV-A59 has evolved to retain histidine at position 209 to optimize virus fitness.

Taxi plays an important role in public transportation. However, in real life, it usually takes a long time for passengers to wait for an unoccupied taxi, while number of no-load taxi is very high. Optimized scheme is required to handle global taxi dispatching. The paper presents a novel method of taxi dispatching based on Artificial Fish Swarm Algorithm (AFSA). To avoid slow convergence and falling into local optimum, standard AFSA is improved using strategies of optimizing visual range, step size and swimming behavior. We compare our improved algorithm with standard AFSA and Particle Swarm Optimization through the evaluation with real data of GPS trajectory of 12,000 taxis in Beijing and make performance analysis of effects of parameter configuration to dispatching algorithm. The experimental results show good effectiveness of our method.