Matches in SemOpenAlex for { <https://semopenalex.org/work/W4292551330> ?p ?o ?g. }
Showing items 1 to 67 of
67
with 100 items per page.
- W4292551330 endingPage "109601" @default.
- W4292551330 startingPage "109601" @default.
- W4292551330 abstract "Incremental event detection is a challenging subfield of event detection, which could be applied in various complex situations, especially for the real-time streaming data scenario where new event classes are continuously emerging. To handle the catastrophic forgetting problem, knowledge distillation has been proved to be a promising mechanism since it requires no extra tools or data storage. However, conventional knowledge distillation-based methods are not effective due to two challenging issues: (a) there is an inherent conflict between reserving history information from old tasks and adjusting to new tasks with shared model output features. (b) the historical category information is not fully utilized, making the model prone to forgetting knowledge on the observed tasks. In this work, we present a History-Enhanced Feature Transfer (HEFT) framework to solve the two challenges mentioned above. For the first challenge, we employ a feature transfer module to separate reserving and adjusting sub-functions by reconstructing the old task features. In this manner, the reconstructed features are responsible for preserving historical knowledge, and the feature extractor with followed classifier only requires to focus on new task classification. For the second issue, a history-enhanced question answering model is introduced to reinforce the ability to memorize historical information, which leverages the trained categories as clues to guide the reasoning process. The experimental results show HEFT outperforms the state-of-the-art model by 6.2% and 9.1% of the whole F1 score on ACE 2005 and TAC KBP 2017 benchmarks, respectively. The ablation study and case study also demonstrate that HEFT could overcome catastrophic forgetting of old class event triggers without heavily relying on historically preserved samples." @default.
- W4292551330 created "2022-08-22" @default.
- W4292551330 creator A5035960468 @default.
- W4292551330 creator A5050604078 @default.
- W4292551330 creator A5060210998 @default.
- W4292551330 creator A5066586426 @default.
- W4292551330 creator A5075955258 @default.
- W4292551330 creator A5086257285 @default.
- W4292551330 creator A5090121996 @default.
- W4292551330 date "2022-10-01" @default.
- W4292551330 modified "2023-09-27" @default.
- W4292551330 title "HEFT: A History-Enhanced Feature Transfer framework for incremental event detection" @default.
- W4292551330 doi "https://doi.org/10.1016/j.knosys.2022.109601" @default.
- W4292551330 hasPublicationYear "2022" @default.
- W4292551330 type Work @default.
- W4292551330 citedByCount "0" @default.
- W4292551330 crossrefType "journal-article" @default.
- W4292551330 hasAuthorship W4292551330A5035960468 @default.
- W4292551330 hasAuthorship W4292551330A5050604078 @default.
- W4292551330 hasAuthorship W4292551330A5060210998 @default.
- W4292551330 hasAuthorship W4292551330A5066586426 @default.
- W4292551330 hasAuthorship W4292551330A5075955258 @default.
- W4292551330 hasAuthorship W4292551330A5086257285 @default.
- W4292551330 hasAuthorship W4292551330A5090121996 @default.
- W4292551330 hasConcept C111919701 @default.
- W4292551330 hasConcept C121332964 @default.
- W4292551330 hasConcept C124101348 @default.
- W4292551330 hasConcept C138885662 @default.
- W4292551330 hasConcept C154945302 @default.
- W4292551330 hasConcept C2776175482 @default.
- W4292551330 hasConcept C2776401178 @default.
- W4292551330 hasConcept C2779662365 @default.
- W4292551330 hasConcept C41008148 @default.
- W4292551330 hasConcept C41895202 @default.
- W4292551330 hasConcept C62520636 @default.
- W4292551330 hasConcept C79403827 @default.
- W4292551330 hasConceptScore W4292551330C111919701 @default.
- W4292551330 hasConceptScore W4292551330C121332964 @default.
- W4292551330 hasConceptScore W4292551330C124101348 @default.
- W4292551330 hasConceptScore W4292551330C138885662 @default.
- W4292551330 hasConceptScore W4292551330C154945302 @default.
- W4292551330 hasConceptScore W4292551330C2776175482 @default.
- W4292551330 hasConceptScore W4292551330C2776401178 @default.
- W4292551330 hasConceptScore W4292551330C2779662365 @default.
- W4292551330 hasConceptScore W4292551330C41008148 @default.
- W4292551330 hasConceptScore W4292551330C41895202 @default.
- W4292551330 hasConceptScore W4292551330C62520636 @default.
- W4292551330 hasConceptScore W4292551330C79403827 @default.
- W4292551330 hasLocation W42925513301 @default.
- W4292551330 hasOpenAccess W4292551330 @default.
- W4292551330 hasPrimaryLocation W42925513301 @default.
- W4292551330 hasRelatedWork W2045393060 @default.
- W4292551330 hasRelatedWork W2112973065 @default.
- W4292551330 hasRelatedWork W2166015963 @default.
- W4292551330 hasRelatedWork W2296762741 @default.
- W4292551330 hasRelatedWork W2348097614 @default.
- W4292551330 hasRelatedWork W2348531541 @default.
- W4292551330 hasRelatedWork W2349359195 @default.
- W4292551330 hasRelatedWork W2375382787 @default.
- W4292551330 hasRelatedWork W2772218528 @default.
- W4292551330 hasRelatedWork W3107474891 @default.
- W4292551330 hasVolume "254" @default.
- W4292551330 isParatext "false" @default.
- W4292551330 isRetracted "false" @default.
- W4292551330 workType "article" @default.