Edge–cloud collaboration has become essential for managing the limited computational capacity of edge servers while meeting the dynamic and latency-critical demands of wireless devices in 5G and beyond networks. However, efficient offloading of interdependent tasks remains challenging due to heterogeneous computing resources, time-varying wireless channels, and complex dependency structures among subtasks. This paper proposes a Quantum Graph Attention Network-based Deep Reinforcement Learning (QGAT-DRL) framework for dependency-aware task offloading in collaborative edge–cloud systems. Computation tasks generated by user equipments (UEs) are represented as directed acyclic graphs (DAGs) to capture subtask dependencies, while a quantum-enhanced graph attention mechanism exploits superposition and entanglement to encode high-order correlations across wireless and computing layers with improved representational efficiency. The offloading policy is optimized through a Proximal Policy Optimization (PPO) algorithm to jointly minimize task completion latency and energy consumption. Simulation results demonstrate that QGAT-DRL achieves faster convergence, lower latency, and superior energy efficiency compared with state-of-the-art baselines, demonstrating the promise of hybrid quantum–classical learning for scalable resource orchestration in next-generation edge–cloud networks.
Yuhui Wang, Xingqi Wu, Junaid Farooq et al.· International Conference on...· 0 citations
Retrieval-augmented generation relies mostly on flat, fixed-granularity indexes: documents are cut into uniform chunks and retrieved by similarity, discarding the hierarchical structure of the source. We introduce Semantic Compression Trees (SCT), a hierarchical index in which each node stores only its semantic residual -- the information it adds beyond its parent -- and retrieval proceeds by progressive descent from the root, so that per-query cost is governed by tree depth rather than collection size. We evaluate on QASPER (50 papers, 173 questions) under two protocols differing only in whether the benchmark supplies the relevant document, with bootstrap confidence intervals and paired significance tests throughout. The results are mixed and we report them as such. When the document is given, SCT with a zero-LLM extractive compressor matches dense retrieval on answer quality (0.274 vs. 0.277 F1, $p = 0.37$) using 30% fewer context tokens and no LLM calls to build the index, and residual storage beats storing full summaries at each node (0.274 vs. 0.205, $p<0.001$). Increasing the collection fifty-fold multiplies flat retrieval's per-query scoring work by 48.9x and SCT's by 6.4x. Progressive descent itself is not supported. Retrieving the same residuals without the tree performs identically when the document is given ($p = 0.27$), and descent is substantially worse when the system must select the document (0.122 vs. 0.165, $p<0.001$). Routing accuracy localises the cause: descent selects the correct paper 20.2% of the time against 39.3% for flat retrieval, because that choice is made from the root residual, the most compressed node in the tree. We conclude that the residual representation is worth keeping and top-down routing is not.