Min Zhang

Online reviews play a crucial role in shaping consumer decisions on e-commerce platforms and social media. Ensuring the authenticity and effectiveness of these reviews is paramount for safeguarding consumer rights and maintaining platform integrity. However, Chinese online reviews present unique challenges due to their complexity and the scarcity of annotated data, making the efficient identification of fake reviews a persistent challenge.To address these issues, this study introduces a novel pre-training model called SGAN-SAM-ERNIE, which leverages a self-attention mechanism within a semi-supervised framework. Initially, the ERNIE3.0 pre-training model is employed to derive high-quality language representations from online review data. Subsequently, the generator within the semi-supervised generative adversarial network (SGAN) generates synthetic data resembling the distribution of authentic reviews, while the self-attention mechanism is integrated to capture key information and semantic dependencies within the input sequences, thereby enhancing the model’s comprehension of the reviews. Finally, a discriminator is employed to distinguish between authentic and fake reviews.Experimental validation is conducted using both cross-language and Chinese datasets. The results demonstrate a significant performance enhancement in fake review identification, even in scenarios with limited labeled data availability. This underscores the effectiveness of the proposed SGAN-SAM-ERNIE model in tackling the challenges associated with identifying fake reviews in Chinese.

We demonstrate that the ECC-based authentication scheme for internet of drones (short for IoD-AS, published in IEEE Systems Journal, Hussain et al., 2021) Amassing the security: An ECC-based authentication scheme for Internet of Drones) fails to prevent an adversary from launching drone impersonation attacks. We also demonstrate that an adversary can compromise the session key of a target user and her/his employed drone.

In the wireless sensor networks (WSNs), users sometimes need to access real-time information from a certain sensor node. In order to prevent unauthorized users from acquiring information from the sensor node, a lot of authentication schemes suited for the WSNs condition have been proposed. Recently, Das has proposed a secure and light-weight authentication scheme for the WSNs based on three factors, claiming that it can withstand various attacks. But after a careful analysis, there are still several security problems as follows: a) The scheme has some design flaws; b)The scheme may suffer sensor node impersonation attack; c) The scheme may suffer repeat registered attack; d) When the user access the sensor node, the scheme cannot protect the identity information of the sensor node. To solve these problems, a new scheme using the secure sketch algorithm is presented in this paper. The security of improved scheme has been analyzed by ProVerif. Thorough analysis shows that the presented scheme can provide stronger security and slight lower computation at client than Das’s protocol. What’s more, it can overcome repeat registered attack and achieve sensor node’s identity anonymity protection.

In this paper, a resistive-reactive series of continuous inverse modes is proposed for extending bandwidth of the series of Inverse continuous mode (SICM). A broadband class inverse F power amplifier (PA), which operates across 2.5-4.1 GHz (48.5% bandwidth), is designed by using bare GaN HEMT to prove this theory. Experiment results indicate that the PA mentioned above is able to output 9.1-18.4 W saturation power and the drain efficiency can reach 47%-75% in the whole interested band. The average output power of this designed PA is 14.2 W and the average drain efficiency is 60.1%.