Blockchain technology is widely used in the field of digital currency, however, the typical blockchain systems suffer from high transaction confirmation latency and expensive fees, which make it difficult to meet the needs of daily payment scenarios. Schemes such as state channel and payment center shift the on-chain payment confirmation process off-chain, thereby reducing latency. Yet these solutions require locking in additional funds or introducing a trusted third party, and seldom consider the issue of lowering transaction fees. In this paper, we propose a cryptocurrency fast payment scheme that organises off-chain participants through smart contracts to enable fast payments. The transaction parties update their state off-chain to ensure fast confirmation. At settlement, off-chain transactions between customer and multiple merchants or merchant and multiple customers are aggregated into a single on-chain transaction, effectively reducing overall transaction fees. We protect transaction privacy by hiding the transaction amount and balance through range proof. We conduct experiments on Ethereum platform, and the results show that the system latency meets the needs of real-world scenarios.

Anonymous credential is widely used in online services, where issuers in authorities issue credentials to users and then users can selectively and privately prove their identities and attributes. However, users may misbehave under anonymous settings. Therefore, we need to trace the credential proof to obtain the user’s identity and link credential proofs to achieve supervision. Existing solutions either have the single point of failure problem or require multiple supervisors perform threshold computations on all users’ identities, it is inefficient in practice especially when the number of users increases. In this paper, we present a credential management system in multiple authorities with efficient and decentralized supervision. Specifically, we design a multi-authority credential management architecture, where each issuer in authorities issues credentials to users and supervisors trace and link credential proofs in multiple authorities. Then, we present efficient and decentralized credential proof tracing and linking protocols, where more than threshold supervisors can trace credential proofs to obtain users’ identities and generate users’ linking keys. Verifiers can link each malicious user’s credential proofs efficiently with those linking keys. We conduct experiments on our system in the WAN and LAN settings and compare it with another threshold attribute-based credential scheme. The experimental results demonstrate that our solution is efficient in practice.

Entity linking (EL) is a challenging task as it typically requires matching an ambiguous entity mention with its corresponding entity in a knowledge base (KB). The mainstream studies focus on learning and evaluating linking models on the same corpus and obtained significant performance achievement, however, they often overlook the generalization ability to out-of-domain corpus, which is more realistic yet much more challenging. To address this issue, we introduce a novel neural-symbolic model for entity linking, which is inspired by the symbol-manipulation mechanism in human brains. Specifically, we abstract diverse features into unified variables, then combine them using neural operators to capture diverse relevance requirements, and finally aggregate relevance scores through voting. We conduct experiments on eleven benchmark datasets with different types of text, and the results show that our method outperforms nearly all baselines. Notably, the best performance of our method on seven out-of-domain datasets highlights its generalization ability.

Table entity linking (TEL) aims to map entity mentions in the table to their corresponding entities in a knowledge base (KB).The core of this task is to leverage structured contexts, specifically row and column contexts, to enhance the semantics of mentions in entity disambiguation.Most entity linking (EL) methods primarily focus on understanding sequential text contexts, making it difficult to adapt to the row and column structure of tables.Additionally, existing methods for TEL indiscriminately mix row and column contexts together, overlooking their semantic differences.In this paper, we explicitly distinguish the modeling of row and column contexts, and propose a method called RoCEL to capture their distinct semantics.Specifically, for row contexts in tables, we take the attention mechanism to learn the implicit relational dependencies between each cell and the mention.For column contexts in tables, we employ a set-wise encoder to learn the categorical information about the group of mentions.At last, we merge both contexts to obtain the final mention embedding for link prediction.Experiments on four benchmarks show that our approach outperforms the stateof-the-art (SOTA) baseline by about 1.5% on the in-domain dataset, and by 3.7% on average across three out-of-domain datasets.

The use of deep learning to interpret ground-penetrating radar (GPR) data for automated target detection has become a rapidly developing field. However, the performance of deep learning networks is significantly affected by the quantity and diversity of the available GPR data for training. The finite-difference time-domain (FDTD) method is commonly used to supplement datasets. However, there are considerable differences between FDTD-simulated and real GPR data, which limit the effectiveness of using FDTD data for training. To address these challenges, we propose a generative adversarial network (GAN)-based domain adaptation (DA) network for detecting target reflections in GPR data. This network combines the you only look once (YOLO) network with a DA module, forming a GAN-based DA-YOLO that aligns features extracted from simulated and real GPR data to enhance the performance of the network on real GPR data. To further improve the network’s performance on real GPR data, a pseudo-labeling strategy is employed. This approach leverages the pseudo-labels generated by the network to iteratively refine and improve its accuracy. To validate the proposed method, we conduct comparative experiments to evaluate the impact of key factors, including network architecture, pseudo-labeling thresholds, and source-target-domain discrepancies, on feature alignment and the improvement of detection performance for target-domain data. The effectiveness of the proposed method is validated using real-world GPR data collected from a railway tunnel inspection. The results demonstrate that the proposed DA-YOLO network achieves the best detection performance on real data, with recall, precision,${F}1$score, and mAP values of 81.25%, 86.67%, 83.87%, and 69.30%, respectively, outperforming other networks that do not utilize DA methods. This study provides a solution for situations where real GPR data is scarce and only FDTD-simulated data can be used as training data.

Credential management is widely used in online services such as electronic identity cards, e-health, and e-voting, in which users prove their identity or attributes with credentials issued by authorities. Under some circumstances, a user needs to prove her/his identity or attributes in multiple credentials to a verifier. In existing credential management systems, a user either proves her/his credentials one by one or requests new credentials from authorities with the original ones, and they are inefficient in practice. Moreover, existing decentralized credential management systems either rely on multiple single parties or do not support attribute revocation. In this paper, we present a threshold credential management system with threshold issuance and revocation and user-autonomy aggregation. Specifically, we design a decentralized credential management architecture where multiple authorities form an alliance and manage credentials collaboratively. Then, we propose a threshold credential management scheme, where user issuance and revocation must be approved by multiple credential managers, and a user can aggregate her/his credentials and prove them to a verifier simultaneously. We conduct experiments on our system and the results demonstrate that it is suitable in practice.

Entity representation plays a central role in building effective entity retrieval models. Recent works propose to learn entity representations based on entity-centric contexts, which achieve SOTA performances on many tasks. However, these methods lead to poor representations for unseen entities since they rely on a multitude of occurrences for each entity to enable accurate representations. To address this issue, we propose to learn enhanced descriptional representations for unseen entities by distilling knowledge from distributional semantics into descriptional embeddings. Specifically, we infer enhanced embeddings for unseen entities based on descriptions by aligning the descriptional embedding space to the distributional embedding space with different granularities, i.e., element-level, batch-level and space-level alignment. Experimental results on four benchmark datasets show that our approach improves the performance over all baseline methods. In particular, our approach can achieve the effectiveness of the teacher model on almost all entities, and maintain such high performance on unseen entities.

In the last several years hundreds of thousands of SSDs have been deployed in the data centers of Baidu, China's largest Internet search company. Currently only 40\% or less of the raw bandwidth of the flash memory in the SSDs is delivered by the storage system to the applications. Moreover, because of space over-provisioning in the SSD to accommodate non-sequential or random writes, and additionally, parity coding across flash channels, typically only 50-70\% of the raw capacity of a commodity SSD can be used for user data. Given the large scale of Baidu's data center, making the most effective use of its SSDs is of great importance. Specifically, we seek to maximize both bandwidth and usable capacity.