Matches in SemOpenAlex for { <https://semopenalex.org/work/W2963979855> ?p ?o ?g. }
- W2963979855 abstract "We formulate tracking as an online decision-making process, where a tracking agent must follow an object despite ambiguous image frames and a limited computational bud- get. Crucially, the agent must decide where to look in the upcoming frames, when to reinitialize because it believes the target has been lost, and when to update its appearance model for the tracked object. Such decisions are typically made heuristically. Instead, we propose to learn an optimal decision-making policy by formulating tracking as a partially observable decision-making process (POMDP). We learn policies with deep reinforcement learning algorithms that need supervision (a reward signal) only when the track has gone awry. We demonstrate that sparse rewards al- low us to quickly train on massive datasets, several orders of magnitude more than past work. Interestingly, by treat- ing the data source of Internet videos as unlimited streams, we both learn and evaluate our trackers in a single, unified computational stream." @default.
- W2963979855 created "2019-07-30" @default.
- W2963979855 creator A5004353237 @default.
- W2963979855 creator A5086960085 @default.
- W2963979855 date "2017-10-01" @default.
- W2963979855 modified "2023-10-04" @default.
- W2963979855 title "Tracking as Online Decision-Making: Learning a Policy from Streaming Videos with Reinforcement Learning" @default.
- W2963979855 cites W1857884451 @default.
- W2963979855 cites W1915599933 @default.
- W2963979855 cites W1955514522 @default.
- W2963979855 cites W1982262386 @default.
- W2963979855 cites W1998738999 @default.
- W2963979855 cites W2015058673 @default.
- W2963979855 cites W2017293771 @default.
- W2963979855 cites W2024615826 @default.
- W2963979855 cites W203079324 @default.
- W2963979855 cites W2031342017 @default.
- W2963979855 cites W2045724293 @default.
- W2963979855 cites W2081580037 @default.
- W2963979855 cites W2089961441 @default.
- W2963979855 cites W2117539524 @default.
- W2963979855 cites W2118345389 @default.
- W2963979855 cites W2118897962 @default.
- W2963979855 cites W2122088301 @default.
- W2963979855 cites W2124211486 @default.
- W2963979855 cites W2124247605 @default.
- W2963979855 cites W2126302311 @default.
- W2963979855 cites W2138068405 @default.
- W2963979855 cites W2145339207 @default.
- W2963979855 cites W2148675068 @default.
- W2963979855 cites W2161986303 @default.
- W2963979855 cites W2163309385 @default.
- W2963979855 cites W2168359464 @default.
- W2963979855 cites W2194002514 @default.
- W2963979855 cites W2211629196 @default.
- W2963979855 cites W2214352687 @default.
- W2963979855 cites W2225887246 @default.
- W2963979855 cites W2237503480 @default.
- W2963979855 cites W2395596584 @default.
- W2963979855 cites W2469312016 @default.
- W2963979855 cites W2470456807 @default.
- W2963979855 cites W2562852216 @default.
- W2963979855 cites W2963249584 @default.
- W2963979855 cites W2964111344 @default.
- W2963979855 cites W3102624093 @default.
- W2963979855 doi "https://doi.org/10.1109/iccv.2017.43" @default.
- W2963979855 hasPublicationYear "2017" @default.
- W2963979855 type Work @default.
- W2963979855 sameAs 2963979855 @default.
- W2963979855 citedByCount "78" @default.
- W2963979855 countsByYear W29639798552017 @default.
- W2963979855 countsByYear W29639798552018 @default.
- W2963979855 countsByYear W29639798552019 @default.
- W2963979855 countsByYear W29639798552020 @default.
- W2963979855 countsByYear W29639798552021 @default.
- W2963979855 countsByYear W29639798552022 @default.
- W2963979855 countsByYear W29639798552023 @default.
- W2963979855 crossrefType "proceedings-article" @default.
- W2963979855 hasAuthorship W2963979855A5004353237 @default.
- W2963979855 hasAuthorship W2963979855A5086960085 @default.
- W2963979855 hasBestOaLocation W29639798552 @default.
- W2963979855 hasConcept C105795698 @default.
- W2963979855 hasConcept C106189395 @default.
- W2963979855 hasConcept C110875604 @default.
- W2963979855 hasConcept C111919701 @default.
- W2963979855 hasConcept C11413529 @default.
- W2963979855 hasConcept C115988155 @default.
- W2963979855 hasConcept C119857082 @default.
- W2963979855 hasConcept C136764020 @default.
- W2963979855 hasConcept C154945302 @default.
- W2963979855 hasConcept C15744967 @default.
- W2963979855 hasConcept C159886148 @default.
- W2963979855 hasConcept C163836022 @default.
- W2963979855 hasConcept C17098449 @default.
- W2963979855 hasConcept C19417346 @default.
- W2963979855 hasConcept C202474056 @default.
- W2963979855 hasConcept C2775936607 @default.
- W2963979855 hasConcept C2781238097 @default.
- W2963979855 hasConcept C2986087404 @default.
- W2963979855 hasConcept C33923547 @default.
- W2963979855 hasConcept C41008148 @default.
- W2963979855 hasConcept C49774154 @default.
- W2963979855 hasConcept C56461940 @default.
- W2963979855 hasConcept C57501372 @default.
- W2963979855 hasConcept C97541855 @default.
- W2963979855 hasConcept C98045186 @default.
- W2963979855 hasConcept C98763669 @default.
- W2963979855 hasConceptScore W2963979855C105795698 @default.
- W2963979855 hasConceptScore W2963979855C106189395 @default.
- W2963979855 hasConceptScore W2963979855C110875604 @default.
- W2963979855 hasConceptScore W2963979855C111919701 @default.
- W2963979855 hasConceptScore W2963979855C11413529 @default.
- W2963979855 hasConceptScore W2963979855C115988155 @default.
- W2963979855 hasConceptScore W2963979855C119857082 @default.
- W2963979855 hasConceptScore W2963979855C136764020 @default.
- W2963979855 hasConceptScore W2963979855C154945302 @default.
- W2963979855 hasConceptScore W2963979855C15744967 @default.
- W2963979855 hasConceptScore W2963979855C159886148 @default.
- W2963979855 hasConceptScore W2963979855C163836022 @default.
- W2963979855 hasConceptScore W2963979855C17098449 @default.