Matches in SemOpenAlex for { <https://semopenalex.org/work/W3216334431> ?p ?o ?g. }
Showing items 1 to 99 of
99
with 100 items per page.
- W3216334431 abstract "Vision-and-Language Navigation (VLN) is a challenging task in the field of artificial intelligence. Although massive progress has been made in this task over the past few years attributed to breakthroughs in deep vision and language models, it remains tough to build VLN models that can generalize as well as humans. In this paper, we provide a new perspective to improve VLN models. Based on our discovery that snapshots of the same VLN model behave significantly differently even when their success rates are relatively the same, we propose a snapshot-based ensemble solution that leverages predictions among multiple snapshots. Constructed on the snapshots of the existing state-of-the-art (SOTA) model $circlearrowright$BERT and our past-action-aware modification, our proposed ensemble achieves the new SOTA performance in the R2R dataset challenge in Navigation Error (NE) and Success weighted by Path Length (SPL)." @default.
- W3216334431 created "2021-12-06" @default.
- W3216334431 creator A5027174211 @default.
- W3216334431 creator A5084637778 @default.
- W3216334431 creator A5085882524 @default.
- W3216334431 date "2021-11-28" @default.
- W3216334431 modified "2023-09-24" @default.
- W3216334431 title "Explore the Potential Performance of Vision-and-Language Navigation Model: a Snapshot Ensemble Method" @default.
- W3216334431 cites W1525482321 @default.
- W3216334431 cites W1797070008 @default.
- W3216334431 cites W1902237438 @default.
- W3216334431 cites W1930624869 @default.
- W3216334431 cites W1988790447 @default.
- W3216334431 cites W2135293965 @default.
- W3216334431 cites W2402268235 @default.
- W3216334431 cites W2518108298 @default.
- W3216334431 cites W2557465155 @default.
- W3216334431 cites W2613763509 @default.
- W3216334431 cites W2767722847 @default.
- W3216334431 cites W2774661387 @default.
- W3216334431 cites W2776202271 @default.
- W3216334431 cites W2783375473 @default.
- W3216334431 cites W2784820185 @default.
- W3216334431 cites W2912934387 @default.
- W3216334431 cites W2918407751 @default.
- W3216334431 cites W2926977875 @default.
- W3216334431 cites W2947067399 @default.
- W3216334431 cites W2947178363 @default.
- W3216334431 cites W2951696358 @default.
- W3216334431 cites W2951970475 @default.
- W3216334431 cites W2963403868 @default.
- W3216334431 cites W2963575117 @default.
- W3216334431 cites W2963726321 @default.
- W3216334431 cites W2963800628 @default.
- W3216334431 cites W2964043796 @default.
- W3216334431 cites W2964339842 @default.
- W3216334431 cites W2969862959 @default.
- W3216334431 cites W2970608575 @default.
- W3216334431 cites W2971820544 @default.
- W3216334431 cites W3015180014 @default.
- W3216334431 cites W3023306062 @default.
- W3216334431 cites W3034500398 @default.
- W3216334431 cites W3035232877 @default.
- W3216334431 cites W3091588028 @default.
- W3216334431 cites W3100923070 @default.
- W3216334431 cites W3109085430 @default.
- W3216334431 cites W3165915253 @default.
- W3216334431 cites W3172675210 @default.
- W3216334431 cites W3174887918 @default.
- W3216334431 cites W3176974620 @default.
- W3216334431 cites W3092825347 @default.
- W3216334431 doi "https://doi.org/10.48550/arxiv.2111.14267" @default.
- W3216334431 hasPublicationYear "2021" @default.
- W3216334431 type Work @default.
- W3216334431 sameAs 3216334431 @default.
- W3216334431 citedByCount "0" @default.
- W3216334431 crossrefType "posted-content" @default.
- W3216334431 hasAuthorship W3216334431A5027174211 @default.
- W3216334431 hasAuthorship W3216334431A5084637778 @default.
- W3216334431 hasAuthorship W3216334431A5085882524 @default.
- W3216334431 hasBestOaLocation W32163344311 @default.
- W3216334431 hasConcept C111919701 @default.
- W3216334431 hasConcept C119857082 @default.
- W3216334431 hasConcept C119898033 @default.
- W3216334431 hasConcept C12713177 @default.
- W3216334431 hasConcept C137293760 @default.
- W3216334431 hasConcept C154945302 @default.
- W3216334431 hasConcept C2983448237 @default.
- W3216334431 hasConcept C41008148 @default.
- W3216334431 hasConcept C45942800 @default.
- W3216334431 hasConcept C55282118 @default.
- W3216334431 hasConceptScore W3216334431C111919701 @default.
- W3216334431 hasConceptScore W3216334431C119857082 @default.
- W3216334431 hasConceptScore W3216334431C119898033 @default.
- W3216334431 hasConceptScore W3216334431C12713177 @default.
- W3216334431 hasConceptScore W3216334431C137293760 @default.
- W3216334431 hasConceptScore W3216334431C154945302 @default.
- W3216334431 hasConceptScore W3216334431C2983448237 @default.
- W3216334431 hasConceptScore W3216334431C41008148 @default.
- W3216334431 hasConceptScore W3216334431C45942800 @default.
- W3216334431 hasConceptScore W3216334431C55282118 @default.
- W3216334431 hasLocation W32163344311 @default.
- W3216334431 hasLocation W32163344312 @default.
- W3216334431 hasOpenAccess W3216334431 @default.
- W3216334431 hasPrimaryLocation W32163344311 @default.
- W3216334431 hasRelatedWork W2810053714 @default.
- W3216334431 hasRelatedWork W3124390867 @default.
- W3216334431 hasRelatedWork W3136979370 @default.
- W3216334431 hasRelatedWork W4281560664 @default.
- W3216334431 hasRelatedWork W4281757034 @default.
- W3216334431 hasRelatedWork W4285046548 @default.
- W3216334431 hasRelatedWork W4285741730 @default.
- W3216334431 hasRelatedWork W4311847748 @default.
- W3216334431 hasRelatedWork W4313488044 @default.
- W3216334431 hasRelatedWork W3214927170 @default.
- W3216334431 isParatext "false" @default.
- W3216334431 isRetracted "false" @default.
- W3216334431 magId "3216334431" @default.
- W3216334431 workType "article" @default.