A5084519698- Cryptography and Data Security
- Privacy-Preserving Technologies in Data
- Cloud Computing and Resource Management
- Stochastic Gradient Optimization Techniques
- Caching and Content Delivery
- Video Coding and Compression Technologies
- Ferroelectric and Negative Capacitance Devices
- Advanced Data Compression Techniques
- Image and Video Quality Assessment
- Complexity and Algorithms in Graphs
- Software-Defined Networks and 5G
- Advanced Data Storage Technologies
Beijing Institute of Technology
2023
University of Calgary
2019-2023
Multi-domain network resource reservation systems are being deployed, driven by the demand and substantial benefits of providing predictable resources. However, a major lack existing is their coarse granularity, due to participating networks' concern revealing sensitive information, which can result in inefficiencies. This paper presents Mercator, novel multi-domain discovery system provide fine-grained, global for collaborative sciences. The foundation Mercator abstraction through...
Heterogeneous Federated Learning (HtFL) enables collaborative learning on multiple clients with different model architectures while preserving privacy. Despite recent research progress, knowledge sharing in HtFL is still difficult due to data and heterogeneity. To tackle this issue, we leverage the stored pre-trained generators propose a new upload-efficient transfer scheme called Knowledge-Transfer Loop (FedKTL). Our FedKTL can produce client-task-related prototypical image-vector pairs via...
Persistent key-value (KV) stores are widely used by cloud services at ByteDance as local storage engines, and RocksDB to be the de facto implementation since it can tailored a variety of workloads requirements. In this paper, we provide key insights into engine usage ByteDance, explain why combination highly write-intensive stringent requirements on cost efficiency point lookup tail latency may pose challenges general-purpose such RocksDB, present design LavaStore , high-performance...
Traditional Adaptive BitRate (ABR) streaming faces a challenge of providing smooth experience under highly variable network conditions, especially when low latency is required. Effective adaptation techniques exist for deep-buffer scenarios, such as long-form Video-on-Demand content, but remain elusive short-form or low-latency cases, even short segment may be delivered too late and cause stall. Recently proposed temporal aims to mitigate this problem by being robust losing part the video...
Secure storage and computing of ciphertext data is an important research area in cloud computing. Homomorphic encryption (HE) a popular technique, but most homomorphic requires using the same public key. This restricts application scenarios compromises security system. To address this issue, we propose verifiable multi-key secure system for off-line non-real-time data, which solves issue with double-trapdoor HE scheme. Additionally, lightweight commitment algorithm(LC) that generates...
Ciphertext policy attribute-based encryption (CP-ABE) has been widely used as a fine-grained access control mechanism for data sharing. Existing research on CP-ABE schemes in cloud computing environment predominantly focuses single attribute authority, which may lead to high computation overhead or point of failure. Besides, the encrypted stored restricts ciphertext retrieval. Therefore, we propose multi-authority searchable scheme storage called MASEP sharing securely. In addition, supports...