Jianping Yu

Chao Chen, Haoyu Geng, Nianzu Yang et al.

User interests are usually dynamic in the real world, which poses both theoretical and practical challenges for learning accurate preferences from rich behavior data. Among existing user behavior modeling solutions, attention networks are widely adopted for its effectiveness and relative simplicity. Despite being extensively studied, existing attentions still suffer from two limitations: i) conventional attentions mainly take into account the spatial correlation between user behaviors, regardless the distance between those behaviors in the continuous time space; and ii) these attentions mostly provide a dense and undistinguished distribution over all past behaviors then attentively encode them into the output latent representations. This is however not suitable in practical scenarios where a user's future actions are relevant to a small subset of her/his historical behaviors. In this paper, we propose a novel attention network, named self-modulating attention, that models the complex and non-linearly evolving dynamic user preferences. We empirically demonstrate the effectiveness of our method on top-N sequential recommendation tasks, and the results on three large-scale real-world datasets show that our model can achieve state-of-the-art performance.

QingChun ShenTu, JianPing Yu

The Bitcoin system is an anonymous, decentralized crypto-currency. There are some deanonymizating techniques to cluster Bitcoin addresses and to map them to users' identifications in the two research directions of Analysis of Transaction Chain (ATC) and Analysis of Bitcoin Protocol and Network (ABPN). Nowadays, there are also some anonymization methods such as coin-mixing and transaction remote release (TRR) to cover the relationship between Bitcoin address and the user. This paper studies anonymization and de-anonymization technologies and proposes some directions for further research.

QingChun ShenTu, JianPing Yu

To strengthen the anonymity of Bitcoin, several centralized coin-mixing providers (mixers) such as BitcoinFog.com, BitLaundry.com, and Blockchain.info assist users to mix Bitcoins through CoinJoin transactions with multiple inputs and multiple outputs to uncover the relationship between them. However, these mixers know the output address of each user, such that they cannot provide true anonymity. This paper proposes a centralized coin-mixing algorithm based on an elliptic curve blind signature scheme (denoted as Blind-Mixing) that obstructs mixers from linking an input address with an output address. Comparisons among three blind signature based algorithms, Blind-Mixing, BlindCoin, and RSA Coin-Mixing, are conducted. It is determined that BlindCoin may be deanonymized because of its use of a public log. In RSA Coin-Mixing, a user's Bitcoins may be falsely claimed by another. In addition, the blind signature scheme of Blind-Mixing executes 10.5 times faster than that of RSA Coin-Mixing.

QingChun ShenTu, JianPing Yu

The anonymity of the Bitcoin system has some shortcomings. Analysis of Transaction Chain (ATC) and Analysis of Bitcoin Protocol and Network (ABPN) are two important methods of deanonymizing bitcoin transactions. Nowadays, there are some anonymization methods to combat ATC but there has been little research into ways to counter ABPN. This paper proposes a new anonymization technology called Transaction Remote Release (TRR). Inspired by The Onion Router (TOR), TRR is able to render several typical attacking methods of ABPN ineffective. Furthermore, the performance of encryption and decryption of TRR is good and the growth rate of the cipher is very limited. Hence, TRR is suited for practical applications.