Conventional blockchains can provide data availability and integrity only. Tons of applications additionally need confidentiality with flexible access control such that data providers can decide how their data are shared through blockchains. This paper aims at enhancing Byzantine Fault Tolerance (BFT)-based permissioned blockchains with controllable access control. To this goal, we extend the concept of Proxy Re-Encryption (PRE) to a new variant called Controllable Threshold PRE (CT-PRE). The traditional PRE enables a proxy, using a re-encryption key, to convert a ciphertext meant for delegator A into another ciphertext meant for delegatee B, all without exposing the original message. CT-PRE extends PRE into the setting with multiple proxies (corresponding to blockchain servers and avoiding a single point of failure) and enables the delegator to fully take control of its ciphertext. We formally define CT-PRE and construct a provably secure CTPRE scheme. We further extend the CTPRE scheme to a verifiable one VCTPRE. We implement the Verifiable CT-PRE scheme with stronger security, integrate it in our BFT-based blockchain system, and deploy our system in a WAN on Amazon EC2 with 22 nodes across four continents. We show that our system is highly efficient, achieving a throughput of 5.15 ktx/sec (for access control operations) and 10.83 ktx/sec (for write operations, only slightly slower than our BFT write operations), respectively.

Dialogue coherence across multiple turns is still an open challenge. The entity grid model is arguably the most popular approach for coherence modeling. However, it heavily relies on the distribution of entities across adjacent sentences but ignores the emotional context embedded in non-entity text and fails to model long dependencies between speech intentions. These limitations become even more severe when applied to dialogue domain since sentences in dialogue are short, informal and colloquial, thereby, less entities could be extracted and less coherence information could be expressed in these grids. To address the limitations of entity gird methods and incorporate the structure knowledge of dialogue, we propose a new neural network architecture, Hierarchical Intention Enhance Network, to integrate semantic context and speech intention in both utterance and dialogue levels to hierarchically model the global coherence without any entity grids. Our proposed model outperforms the state-of-the-art entity-grid based coherence model on text discrimination task by 17.13% increase in accuracy, confirming the effectiveness of our hierarchical modeling in dialogue context and the crucial importance of intention information in dialogue coherence assessment.