Matches in SemOpenAlex for { <https://semopenalex.org/work/W4312321611> ?p ?o ?g. }
Showing items 1 to 96 of
96
with 100 items per page.
- W4312321611 endingPage "50" @default.
- W4312321611 startingPage "34" @default.
- W4312321611 abstract "Deep learning has recently achieved significant progress in trajectory forecasting. However, the scarcity of trajectory data inhibits the data-hungry deep-learning models from learning good representations. While pre-training methods for representation learning exist in computer vision and natural language processing, they still require large-scale data. It is hard to replicate their success in trajectory forecasting due to the inadequate trajectory data (e.g., 34K samples in the nuScenes dataset). To work around the scarcity of trajectory data, we resort to another data modality closely related to trajectories—HD-maps, which is abundantly provided in existing datasets. In this paper, we propose PreTraM, a self-supervised Pre-training scheme via connecting Trajectories and Maps for trajectory forecasting. PreTraM consists of two parts: 1) Trajectory-Map Contrastive Learning, where we project trajectories and maps to a shared embedding space with cross-modal contrastive learning, 2) Map Contrastive Learning, where we enhance map representation with contrastive learning on large quantities of HD-maps. On top of popular baselines such as AgentFormer and Trajectron++, PreTraM reduces their errors by 5.5% and 6.9% relatively on the nuScenes dataset. We show that PreTraM improves data efficiency and scales well with model size. Our code and pre-trained models will be released at https://github.com/chenfengxu714/PreTraM ." @default.
- W4312321611 created "2023-01-04" @default.
- W4312321611 creator A5018019822 @default.
- W4312321611 creator A5047285420 @default.
- W4312321611 creator A5054449834 @default.
- W4312321611 creator A5063473743 @default.
- W4312321611 creator A5064077634 @default.
- W4312321611 creator A5065597482 @default.
- W4312321611 creator A5068465714 @default.
- W4312321611 creator A5083914796 @default.
- W4312321611 date "2022-01-01" @default.
- W4312321611 modified "2023-10-16" @default.
- W4312321611 title "PreTraM: Self-supervised Pre-training via Connecting Trajectory and Map" @default.
- W4312321611 cites W2007339694 @default.
- W4312321611 cites W2963341956 @default.
- W4312321611 cites W2967177252 @default.
- W4312321611 cites W3034722190 @default.
- W4312321611 cites W3035524453 @default.
- W4312321611 cites W3035574168 @default.
- W4312321611 cites W3035671534 @default.
- W4312321611 cites W3108262631 @default.
- W4312321611 cites W3108486966 @default.
- W4312321611 cites W3116651890 @default.
- W4312321611 cites W3139491754 @default.
- W4312321611 cites W3156636935 @default.
- W4312321611 cites W3173252859 @default.
- W4312321611 cites W3199505925 @default.
- W4312321611 cites W3204875639 @default.
- W4312321611 cites W3209373652 @default.
- W4312321611 cites W3209837334 @default.
- W4312321611 cites W3214950490 @default.
- W4312321611 doi "https://doi.org/10.1007/978-3-031-19842-7_3" @default.
- W4312321611 hasPublicationYear "2022" @default.
- W4312321611 type Work @default.
- W4312321611 citedByCount "2" @default.
- W4312321611 countsByYear W43123216112022 @default.
- W4312321611 countsByYear W43123216112023 @default.
- W4312321611 crossrefType "book-chapter" @default.
- W4312321611 hasAuthorship W4312321611A5018019822 @default.
- W4312321611 hasAuthorship W4312321611A5047285420 @default.
- W4312321611 hasAuthorship W4312321611A5054449834 @default.
- W4312321611 hasAuthorship W4312321611A5063473743 @default.
- W4312321611 hasAuthorship W4312321611A5064077634 @default.
- W4312321611 hasAuthorship W4312321611A5065597482 @default.
- W4312321611 hasAuthorship W4312321611A5068465714 @default.
- W4312321611 hasAuthorship W4312321611A5083914796 @default.
- W4312321611 hasBestOaLocation W43123216112 @default.
- W4312321611 hasConcept C108583219 @default.
- W4312321611 hasConcept C119857082 @default.
- W4312321611 hasConcept C121332964 @default.
- W4312321611 hasConcept C1276947 @default.
- W4312321611 hasConcept C13662910 @default.
- W4312321611 hasConcept C154945302 @default.
- W4312321611 hasConcept C177264268 @default.
- W4312321611 hasConcept C17744445 @default.
- W4312321611 hasConcept C199360897 @default.
- W4312321611 hasConcept C199539241 @default.
- W4312321611 hasConcept C2776359362 @default.
- W4312321611 hasConcept C2776760102 @default.
- W4312321611 hasConcept C41008148 @default.
- W4312321611 hasConcept C41608201 @default.
- W4312321611 hasConcept C94625758 @default.
- W4312321611 hasConceptScore W4312321611C108583219 @default.
- W4312321611 hasConceptScore W4312321611C119857082 @default.
- W4312321611 hasConceptScore W4312321611C121332964 @default.
- W4312321611 hasConceptScore W4312321611C1276947 @default.
- W4312321611 hasConceptScore W4312321611C13662910 @default.
- W4312321611 hasConceptScore W4312321611C154945302 @default.
- W4312321611 hasConceptScore W4312321611C177264268 @default.
- W4312321611 hasConceptScore W4312321611C17744445 @default.
- W4312321611 hasConceptScore W4312321611C199360897 @default.
- W4312321611 hasConceptScore W4312321611C199539241 @default.
- W4312321611 hasConceptScore W4312321611C2776359362 @default.
- W4312321611 hasConceptScore W4312321611C2776760102 @default.
- W4312321611 hasConceptScore W4312321611C41008148 @default.
- W4312321611 hasConceptScore W4312321611C41608201 @default.
- W4312321611 hasConceptScore W4312321611C94625758 @default.
- W4312321611 hasLocation W43123216111 @default.
- W4312321611 hasLocation W43123216112 @default.
- W4312321611 hasOpenAccess W4312321611 @default.
- W4312321611 hasPrimaryLocation W43123216111 @default.
- W4312321611 hasRelatedWork W3014300295 @default.
- W4312321611 hasRelatedWork W3164822677 @default.
- W4312321611 hasRelatedWork W4223943233 @default.
- W4312321611 hasRelatedWork W4225161397 @default.
- W4312321611 hasRelatedWork W4250304930 @default.
- W4312321611 hasRelatedWork W4312200629 @default.
- W4312321611 hasRelatedWork W4360585206 @default.
- W4312321611 hasRelatedWork W4364306694 @default.
- W4312321611 hasRelatedWork W4380075502 @default.
- W4312321611 hasRelatedWork W4380086463 @default.
- W4312321611 isParatext "false" @default.
- W4312321611 isRetracted "false" @default.
- W4312321611 workType "book-chapter" @default.