Matches in SemOpenAlex for { <https://semopenalex.org/work/W4221142947> ?p ?o ?g. }
Showing items 1 to 85 of
85
with 100 items per page.
- W4221142947 abstract "Modern computer vision applications rely on learning-based perception modules parameterized with neural networks for tasks like object detection. These modules frequently have low expected error overall but high error on atypical groups of data due to biases inherent in the training process. In building autonomous vehicles (AV), this problem is an especially important challenge because their perception modules are crucial to the overall system performance. After identifying failures in AV, a human team will comb through the associated data to group perception failures that share common causes. More data from these groups is then collected and annotated before retraining the model to fix the issue. In other words, error groups are found and addressed in hindsight. Our main contribution is a pseudo-automatic method to discover such groups in foresight by performing causal interventions on simulated scenes. To keep our interventions on the data manifold, we utilize masked language models. We verify that the prioritized groups found via intervention are challenging for the object detector and show that retraining with data collected from these groups helps inordinately compared to adding more IID data. We also plan to release software to run interventions in simulated scenes, which we hope will benefit the causality community." @default.
- W4221142947 created "2022-04-03" @default.
- W4221142947 creator A5006347979 @default.
- W4221142947 creator A5026850703 @default.
- W4221142947 creator A5032255237 @default.
- W4221142947 creator A5039510217 @default.
- W4221142947 creator A5070642269 @default.
- W4221142947 creator A5078831654 @default.
- W4221142947 creator A5086201201 @default.
- W4221142947 date "2022-02-08" @default.
- W4221142947 modified "2023-09-26" @default.
- W4221142947 title "Causal Scene BERT: Improving object detection by searching for challenging groups of data" @default.
- W4221142947 doi "https://doi.org/10.48550/arxiv.2202.03651" @default.
- W4221142947 hasPublicationYear "2022" @default.
- W4221142947 type Work @default.
- W4221142947 citedByCount "0" @default.
- W4221142947 crossrefType "posted-content" @default.
- W4221142947 hasAuthorship W4221142947A5006347979 @default.
- W4221142947 hasAuthorship W4221142947A5026850703 @default.
- W4221142947 hasAuthorship W4221142947A5032255237 @default.
- W4221142947 hasAuthorship W4221142947A5039510217 @default.
- W4221142947 hasAuthorship W4221142947A5070642269 @default.
- W4221142947 hasAuthorship W4221142947A5078831654 @default.
- W4221142947 hasAuthorship W4221142947A5086201201 @default.
- W4221142947 hasBestOaLocation W42211429471 @default.
- W4221142947 hasConcept C10347200 @default.
- W4221142947 hasConcept C107457646 @default.
- W4221142947 hasConcept C118552586 @default.
- W4221142947 hasConcept C119857082 @default.
- W4221142947 hasConcept C121332964 @default.
- W4221142947 hasConcept C144133560 @default.
- W4221142947 hasConcept C154945302 @default.
- W4221142947 hasConcept C155202549 @default.
- W4221142947 hasConcept C15744967 @default.
- W4221142947 hasConcept C169760540 @default.
- W4221142947 hasConcept C180747234 @default.
- W4221142947 hasConcept C199360897 @default.
- W4221142947 hasConcept C26760741 @default.
- W4221142947 hasConcept C27415008 @default.
- W4221142947 hasConcept C2776459999 @default.
- W4221142947 hasConcept C2778712577 @default.
- W4221142947 hasConcept C2781238097 @default.
- W4221142947 hasConcept C41008148 @default.
- W4221142947 hasConcept C62520636 @default.
- W4221142947 hasConcept C64357122 @default.
- W4221142947 hasConcept C76155785 @default.
- W4221142947 hasConcept C98045186 @default.
- W4221142947 hasConceptScore W4221142947C10347200 @default.
- W4221142947 hasConceptScore W4221142947C107457646 @default.
- W4221142947 hasConceptScore W4221142947C118552586 @default.
- W4221142947 hasConceptScore W4221142947C119857082 @default.
- W4221142947 hasConceptScore W4221142947C121332964 @default.
- W4221142947 hasConceptScore W4221142947C144133560 @default.
- W4221142947 hasConceptScore W4221142947C154945302 @default.
- W4221142947 hasConceptScore W4221142947C155202549 @default.
- W4221142947 hasConceptScore W4221142947C15744967 @default.
- W4221142947 hasConceptScore W4221142947C169760540 @default.
- W4221142947 hasConceptScore W4221142947C180747234 @default.
- W4221142947 hasConceptScore W4221142947C199360897 @default.
- W4221142947 hasConceptScore W4221142947C26760741 @default.
- W4221142947 hasConceptScore W4221142947C27415008 @default.
- W4221142947 hasConceptScore W4221142947C2776459999 @default.
- W4221142947 hasConceptScore W4221142947C2778712577 @default.
- W4221142947 hasConceptScore W4221142947C2781238097 @default.
- W4221142947 hasConceptScore W4221142947C41008148 @default.
- W4221142947 hasConceptScore W4221142947C62520636 @default.
- W4221142947 hasConceptScore W4221142947C64357122 @default.
- W4221142947 hasConceptScore W4221142947C76155785 @default.
- W4221142947 hasConceptScore W4221142947C98045186 @default.
- W4221142947 hasLocation W42211429471 @default.
- W4221142947 hasOpenAccess W4221142947 @default.
- W4221142947 hasPrimaryLocation W42211429471 @default.
- W4221142947 hasRelatedWork W2006651773 @default.
- W4221142947 hasRelatedWork W2091979774 @default.
- W4221142947 hasRelatedWork W2125659923 @default.
- W4221142947 hasRelatedWork W2889079158 @default.
- W4221142947 hasRelatedWork W2961085424 @default.
- W4221142947 hasRelatedWork W3034436240 @default.
- W4221142947 hasRelatedWork W3037919948 @default.
- W4221142947 hasRelatedWork W4286629047 @default.
- W4221142947 hasRelatedWork W4306674287 @default.
- W4221142947 hasRelatedWork W4224009465 @default.
- W4221142947 isParatext "false" @default.
- W4221142947 isRetracted "false" @default.
- W4221142947 workType "article" @default.