Matches in SemOpenAlex for { <https://semopenalex.org/work/W4312278571> ?p ?o ?g. }
Showing items 1 to 95 of
95
with 100 items per page.
- W4312278571 endingPage "12426" @default.
- W4312278571 startingPage "12419" @default.
- W4312278571 abstract "Source-free model adaptation (SFMA) plays an important role in robust robot automation, which aims to mitigate distributional inconsistency between source and target data while avoiding accessing source data. SFMA methods generally benefit from self-training, which can be prone to overfitting noisy pseudo labels. Previous methods used various techniques to filter out noisy predictions and improve self-training performance. Nonetheless, the filtering is heavily reliant on pre-defined thresholds and suffers from a class imbalance problem, resulting in a bias towards the majority classes and poor performance of the other classes. Aiming at this pitfall, this study proposes to combat the noisy pseudo labels via multi-level perturbation, which involves all the pseudo predictions in self-training. The proposed method introduces different perturbations at three different levels: input, feature, and model. These perturbations boost the diversity of training data and increase the difficulty of fitting pseudo labels, which avoids overfitting the noise without ignoring the minority classes. Besides, the proposed method helps mitigate the intra-domain gap by maximizing the consistency between the original predictions and the perturbed predictions. Experimental results on two domain adaptive segmentation benchmarks confirm the effectiveness of the proposed method, which outperforms state-of-the-art SFMA methods." @default.
- W4312278571 created "2023-01-04" @default.
- W4312278571 creator A5025409170 @default.
- W4312278571 creator A5032447166 @default.
- W4312278571 creator A5034435564 @default.
- W4312278571 creator A5060664791 @default.
- W4312278571 creator A5076427704 @default.
- W4312278571 date "2022-10-01" @default.
- W4312278571 modified "2023-10-16" @default.
- W4312278571 title "Multi-Level Consistency Learning for Source-Free Model Adaptation" @default.
- W4312278571 cites W2194775991 @default.
- W4312278571 cites W2340897893 @default.
- W4312278571 cites W2412782625 @default.
- W4312278571 cites W2431874326 @default.
- W4312278571 cites W2487365028 @default.
- W4312278571 cites W2807907893 @default.
- W4312278571 cites W2895281799 @default.
- W4312278571 cites W2932414082 @default.
- W4312278571 cites W2962687275 @default.
- W4312278571 cites W2963073217 @default.
- W4312278571 cites W2969893028 @default.
- W4312278571 cites W2972285644 @default.
- W4312278571 cites W2985406498 @default.
- W4312278571 cites W2992308087 @default.
- W4312278571 cites W3034272105 @default.
- W4312278571 cites W3035256099 @default.
- W4312278571 cites W3035294798 @default.
- W4312278571 cites W3099319035 @default.
- W4312278571 cites W3102977943 @default.
- W4312278571 cites W3173206925 @default.
- W4312278571 cites W3175042385 @default.
- W4312278571 cites W3177316921 @default.
- W4312278571 cites W3181127262 @default.
- W4312278571 cites W3205594649 @default.
- W4312278571 cites W3206487064 @default.
- W4312278571 cites W4214611271 @default.
- W4312278571 cites W4214715922 @default.
- W4312278571 doi "https://doi.org/10.1109/lra.2022.3216997" @default.
- W4312278571 hasPublicationYear "2022" @default.
- W4312278571 type Work @default.
- W4312278571 citedByCount "0" @default.
- W4312278571 crossrefType "journal-article" @default.
- W4312278571 hasAuthorship W4312278571A5025409170 @default.
- W4312278571 hasAuthorship W4312278571A5032447166 @default.
- W4312278571 hasAuthorship W4312278571A5034435564 @default.
- W4312278571 hasAuthorship W4312278571A5060664791 @default.
- W4312278571 hasAuthorship W4312278571A5076427704 @default.
- W4312278571 hasConcept C115961682 @default.
- W4312278571 hasConcept C119857082 @default.
- W4312278571 hasConcept C124101348 @default.
- W4312278571 hasConcept C153180895 @default.
- W4312278571 hasConcept C154945302 @default.
- W4312278571 hasConcept C22019652 @default.
- W4312278571 hasConcept C2776434776 @default.
- W4312278571 hasConcept C2776436953 @default.
- W4312278571 hasConcept C41008148 @default.
- W4312278571 hasConcept C50644808 @default.
- W4312278571 hasConcept C89600930 @default.
- W4312278571 hasConcept C95623464 @default.
- W4312278571 hasConcept C99498987 @default.
- W4312278571 hasConceptScore W4312278571C115961682 @default.
- W4312278571 hasConceptScore W4312278571C119857082 @default.
- W4312278571 hasConceptScore W4312278571C124101348 @default.
- W4312278571 hasConceptScore W4312278571C153180895 @default.
- W4312278571 hasConceptScore W4312278571C154945302 @default.
- W4312278571 hasConceptScore W4312278571C22019652 @default.
- W4312278571 hasConceptScore W4312278571C2776434776 @default.
- W4312278571 hasConceptScore W4312278571C2776436953 @default.
- W4312278571 hasConceptScore W4312278571C41008148 @default.
- W4312278571 hasConceptScore W4312278571C50644808 @default.
- W4312278571 hasConceptScore W4312278571C89600930 @default.
- W4312278571 hasConceptScore W4312278571C95623464 @default.
- W4312278571 hasConceptScore W4312278571C99498987 @default.
- W4312278571 hasFunder F4320322866 @default.
- W4312278571 hasFunder F4320324150 @default.
- W4312278571 hasIssue "4" @default.
- W4312278571 hasLocation W43122785711 @default.
- W4312278571 hasOpenAccess W4312278571 @default.
- W4312278571 hasPrimaryLocation W43122785711 @default.
- W4312278571 hasRelatedWork W1996541855 @default.
- W4312278571 hasRelatedWork W2944843851 @default.
- W4312278571 hasRelatedWork W2985459377 @default.
- W4312278571 hasRelatedWork W2989932438 @default.
- W4312278571 hasRelatedWork W3011996705 @default.
- W4312278571 hasRelatedWork W3099765033 @default.
- W4312278571 hasRelatedWork W3175189414 @default.
- W4312278571 hasRelatedWork W3206592002 @default.
- W4312278571 hasRelatedWork W4210794429 @default.
- W4312278571 hasRelatedWork W4225691219 @default.
- W4312278571 hasVolume "7" @default.
- W4312278571 isParatext "false" @default.
- W4312278571 isRetracted "false" @default.
- W4312278571 workType "article" @default.