To cope with the explosive traffic demands and limited capacity provided by the current cellular networks, Delay Tolerant Networking (DTN) is used to migrate traffic from the cellular networks to the free and high capacity device-todevice networks. The current DTN-based mobile data offloading models do not address the heterogeneity of mobile traffic and are based on simple network assumptions. In this paper, we establish a mathematical framework to study the problem of multiple mobile data offloading under realistic network assumptions, where 1) mobile data is heterogeneous in terms of size and lifetime, 2) mobile users have different data subscribing interests, and 3) the storage of offloading helpers is limited. We formulate the maximum mobile data offloading as a Submodular Function Maximization problem with multiple linear constraints of limited storage and propose greedy, approximated and optimal algorithms for different offloading scenarios. We show that our algorithms can effectively offload data to DTNs by extensive simulations which employ real traces of both humans and vehicles.
Continual learning aims to enable a single model to learn a sequence of tasks without catastrophic forgetting. Top-performing methods usually require a rehearsal buffer to store past pristine examples for experience replay, which, however, limits their practical value due to privacy and memory constraints. In this work, we present a simple yet effective framework, DualPrompt, which learns a tiny set of parameters, called prompts, to properly instruct a pre-trained model to learn tasks arriving sequentially without buffering past examples. DualPrompt presents a novel approach to attach complementary prompts to the pre-trained backbone, and then formulates the objective as learning task-invariant and task-specific "instructions". With extensive experimental validation, DualPrompt consistently sets state-of-the-art performance under the challenging class-incremental setting. In particular, DualPrompt outperforms recent advanced continual learning methods with relatively large buffer sizes. We also introduce a more challenging benchmark, Split ImageNet-R, to help generalize rehearsal-free continual learning research. Source code is available at https://github.com/google-research/l2p.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.