Matches in SemOpenAlex for { <https://semopenalex.org/work/W4384080231> ?p ?o ?g. }
Showing items 1 to 82 of
82
with 100 items per page.
- W4384080231 endingPage "18" @default.
- W4384080231 startingPage "1" @default.
- W4384080231 abstract "Optimizing deep neural network (DNN) models to meet quality of service (QoS) requirements in terms of accuracy and computation is of crucial importance for realizing efficient on-device inference in resource-constrained artificial intelligence of things (AIoT). However, most existing works can hardly satisfy the aforementioned QoS requirements since the intrinsic multi-scale characteristic of DNN structures has been seldom considered. In this paper, we formulate a QoS-ensured DNN model structure optimization problem as a novel multi-scale Markov decision process (MSMDP), which can collaboratively decide the DNN structures from different scales. To efficiently solve the above problem, we propose a multi-scale reinforcement learning (MSRL) algorithm, which jointly optimizes block and channel number by interactive multi-scale decision, while ensuring QoS by QoS-based decision evaluation and policy update. Extensive experiments are conducted in both the actual AIoT scenarios and public datasets for different tasks by using different AIoT devices. The results confirm that our proposed MSRL outperforms the baseline schemes in terms of QoS satisfaction, convergence performance, and complexity. Specifically, our algorithm respectively reduces 98.6% computation and 95.7% model size at most while ensuring the QoS compared with the state-of-the-art methods." @default.
- W4384080231 created "2023-07-13" @default.
- W4384080231 creator A5011900098 @default.
- W4384080231 creator A5022448322 @default.
- W4384080231 creator A5037036194 @default.
- W4384080231 creator A5038978459 @default.
- W4384080231 creator A5087954502 @default.
- W4384080231 date "2023-01-01" @default.
- W4384080231 modified "2023-09-23" @default.
- W4384080231 title "QoS-Ensured Model Optimization for AIoT: A Multi-Scale Reinforcement Learning Approach" @default.
- W4384080231 doi "https://doi.org/10.1109/tmc.2023.3294512" @default.
- W4384080231 hasPublicationYear "2023" @default.
- W4384080231 type Work @default.
- W4384080231 citedByCount "0" @default.
- W4384080231 crossrefType "journal-article" @default.
- W4384080231 hasAuthorship W4384080231A5011900098 @default.
- W4384080231 hasAuthorship W4384080231A5022448322 @default.
- W4384080231 hasAuthorship W4384080231A5037036194 @default.
- W4384080231 hasAuthorship W4384080231A5038978459 @default.
- W4384080231 hasAuthorship W4384080231A5087954502 @default.
- W4384080231 hasConcept C105795698 @default.
- W4384080231 hasConcept C106189395 @default.
- W4384080231 hasConcept C111919701 @default.
- W4384080231 hasConcept C119857082 @default.
- W4384080231 hasConcept C120314980 @default.
- W4384080231 hasConcept C121332964 @default.
- W4384080231 hasConcept C154945302 @default.
- W4384080231 hasConcept C159886148 @default.
- W4384080231 hasConcept C2524010 @default.
- W4384080231 hasConcept C2776214188 @default.
- W4384080231 hasConcept C2777210771 @default.
- W4384080231 hasConcept C2778755073 @default.
- W4384080231 hasConcept C31258907 @default.
- W4384080231 hasConcept C33923547 @default.
- W4384080231 hasConcept C41008148 @default.
- W4384080231 hasConcept C48044578 @default.
- W4384080231 hasConcept C50644808 @default.
- W4384080231 hasConcept C5119721 @default.
- W4384080231 hasConcept C62520636 @default.
- W4384080231 hasConcept C77088390 @default.
- W4384080231 hasConcept C97541855 @default.
- W4384080231 hasConcept C98045186 @default.
- W4384080231 hasConceptScore W4384080231C105795698 @default.
- W4384080231 hasConceptScore W4384080231C106189395 @default.
- W4384080231 hasConceptScore W4384080231C111919701 @default.
- W4384080231 hasConceptScore W4384080231C119857082 @default.
- W4384080231 hasConceptScore W4384080231C120314980 @default.
- W4384080231 hasConceptScore W4384080231C121332964 @default.
- W4384080231 hasConceptScore W4384080231C154945302 @default.
- W4384080231 hasConceptScore W4384080231C159886148 @default.
- W4384080231 hasConceptScore W4384080231C2524010 @default.
- W4384080231 hasConceptScore W4384080231C2776214188 @default.
- W4384080231 hasConceptScore W4384080231C2777210771 @default.
- W4384080231 hasConceptScore W4384080231C2778755073 @default.
- W4384080231 hasConceptScore W4384080231C31258907 @default.
- W4384080231 hasConceptScore W4384080231C33923547 @default.
- W4384080231 hasConceptScore W4384080231C41008148 @default.
- W4384080231 hasConceptScore W4384080231C48044578 @default.
- W4384080231 hasConceptScore W4384080231C50644808 @default.
- W4384080231 hasConceptScore W4384080231C5119721 @default.
- W4384080231 hasConceptScore W4384080231C62520636 @default.
- W4384080231 hasConceptScore W4384080231C77088390 @default.
- W4384080231 hasConceptScore W4384080231C97541855 @default.
- W4384080231 hasConceptScore W4384080231C98045186 @default.
- W4384080231 hasLocation W43840802311 @default.
- W4384080231 hasOpenAccess W4384080231 @default.
- W4384080231 hasPrimaryLocation W43840802311 @default.
- W4384080231 hasRelatedWork W1582107257 @default.
- W4384080231 hasRelatedWork W1596201972 @default.
- W4384080231 hasRelatedWork W1608069547 @default.
- W4384080231 hasRelatedWork W1967954938 @default.
- W4384080231 hasRelatedWork W1986253068 @default.
- W4384080231 hasRelatedWork W2364921833 @default.
- W4384080231 hasRelatedWork W2385146268 @default.
- W4384080231 hasRelatedWork W3186668962 @default.
- W4384080231 hasRelatedWork W3213838085 @default.
- W4384080231 hasRelatedWork W4313591620 @default.
- W4384080231 isParatext "false" @default.
- W4384080231 isRetracted "false" @default.
- W4384080231 workType "article" @default.