Decentralized Physical Infrastructure Networks (DePINs) utilize token incentives to construct permissionless physical infrastructure, but face challenges in ensuring fair compensation for resource contributors. Focusing on bandwidth provision in decentralized data delivery, existing decentralized incentive mechanisms incur prohibitive on-chain costs or employ oversimplified network topologies. We proposeFairRelay, a protocol enablingfair,cost-efficientpayments incomplex multi-hop data delivery. We design two cryptographic primitives: 1)Accountable Multi-hop Data Delivery (AMDD)guaranteeing either correct data receipt or verifiable proof of misbehavior, reducing fair compensation to fee-for-secret exchange; and 2)Enforceable Accumulative HTLC (Enforceable A-HTLC)enabling atomic settlement across multiple off-chain payments via Payment Channel Networks (PCNs). FairRelay's fairness is formally proven within the Universal Composability (UC) framework. Evaluations demonstrate that FairRelay achieveszeroon-chain costs in optimistic execution. Pessimistic scenarios incur constant-cost disputes (O(1) complexity), achieving 13.5% lower overhead than FDE (CCS'24), the state-of-the-art simplified two-party exchange solution (no relays). FairRelay achieves over$95\%$encoding efficiency in 10-hop transmissions.

Options are fundamental to blockchain-based financial services, offering essential tools for risk management and price speculation, which enhance liquidity, flexibility, and market efficiency in decentralized finance (DeFi). Despite the growing interest in options for blockchain-resident assets, such as cryptocurrencies, current option mechanisms face significant challenges, including a high reliance on trusted third parties, limited asset support, high trading delays, and the requirement for option holders to provide upfront collateral. In this paper, we present a protocol that addresses the aforementioned issues. Our protocol is the first to eliminate the need for holders to post collateral when establishing options in trustless service environments (i.e., without a cross-chain bridge), which is achieved by introducing a guarantee from the option writer. Its universality allows for cross-chain options involving nearly any assets on any two different blockchains, provided the chains' programming languages can enforce and execute the necessary contract logic. Another key innovation is reducing option position transfer latency, which uses Double-Authentication-Preventing Signatures (DAPS). Our evaluation demonstrates that the proposed scheme reduces option transfer latency to less than half of that in existing methods. Rigorous security analysis proves that our protocol achieves secure option trading, even when facing adversarial behaviors.

Current benchmarks for evaluating large language models (LLMs) in social media moderation completely overlook a serious threat: covert advertisements, which disguise themselves as regular posts to deceive and mislead consumers into making purchases, leading to significant ethical and legal concerns. In this paper, we present the CHASM, a first-of-its-kind dataset designed to evaluate the capability of Multimodal Large Language Models (MLLMs) in detecting covert advertisements on social media. CHASM is a high-quality, anonymized, manually curated dataset consisting of 4,992 instances, based on real-world scenarios from the Chinese social media platform Rednote. The dataset was collected and annotated under strict privacy protection and quality control protocols. It includes many product experience sharing posts that closely resemble covert advertisements, making the dataset particularly challenging.The results show that under both zero-shot and in-context learning settings, none of the current MLLMs are sufficiently reliable for detecting covert advertisements.Our further experiments revealed that fine-tuning open-source MLLMs on our dataset yielded noticeable performance gains. However, significant challenges persist, such as detecting subtle cues in comments and differences in visual and textual structures.We provide in-depth error analysis and outline future research directions. We hope our study can serve as a call for the research community and platform moderators to develop more precise defenses against this emerging threat.

The COVID-19 pandemic has triggered a surge in misinformation and disinformation online, particularly on social media platforms. The World Health Organization has highlighted the urgent need to combat this infodemic, as false information can lead to the spread of conspiracy theories, false remedies, and xenophobia. Efforts to combat the COVID-19 infodemic have faced challenges, with existing research often oversimplifying the issue by focusing solely on verifying the reliability of information. To effectively address this complex problem, a multifaceted approach is necessary. This study aims to broaden the perspective by analyzing texts from diverse angles, considering societal implications and the necessity of government intervention. We employ a prompt-based contrastive learning framework to meet the challenges. A prompt is a passage of text or query fed into a pretrained language model (PLM) to elicit a response. Prompts provide clear guidance and precise context for language models, which has been shown to make them more effective with limited training samples. This approach can help mitigate class imbalance and data sparsity issues in combating COVID-19 infodemics. We also incorporate prompts into a contrastive learning framework to better understand complex utterances within short, informal, unstructured social media texts. Contrastive learning is effective at distinguishing between useful and irrelevant input samples and focusing on the most discriminative features. Our research demonstrates prompt-based contrastive learning can reinforce each other and provide more accurate assessments of input text reliability than current baseline techniques. The framework not only addresses the challenges of data scarcity and class imbalance but also shows potential for application in other text classification tasks, particularly in low-resource languages with extreme class imbalance issues. In conclusion, the prompt-based contrastive learning approach presented in this study offers a promising strategy to combat the infectious diseases infodemic on social media platforms. By considering the multifaceted nature of misinformation and incorporating prompts and contrastive learning, this method provides a more accurate assessment of text reliability, contributing to the broader efforts to mitigate the harmful effects of misinformation during the pandemic.