FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2399218780> ?p ?o ?g. }

• W2399218780 endingPage "79" @default.
• W2399218780 startingPage "75" @default.
• W2399218780 abstract "This paper describes an approach for creating detailed fullcoverage labellings of human activity. Our goal is to create global maps of physical positions labelled with a distribution over the most likely place name and most likely activity. We ground our ontology of labels as: the term that a person would want to display to someone before they initiate a communication. Rather than compiling a canonical list of possible labels, we piggyback the label data collection in a situated communicative exchange. Using ideas inspired by image segmentation and extended to support our goals we propose machine learning techniques for smoothing distributions across gaps in existing data. Introduction and Related Work The consumerization of sensor-laden platforms such as mobile phones, laptops, and vehicles, that also provide access to the collected data, is enabling data sharing and activity reasoning to scale to new levels. A difficult challenge which remains is to understand how the sensors from different users, ostensibly deployed for specific different reasons, can be principally combined and aggregated for new types of uses. In this paper we introduce a data collection approach and inference mechanism by which it may be possible to create dense geographic maps that are identified with place and activity labels. We obtain information about place and activity labelling in a supervised manner by piggybacking onto the communication practices of instant messaging and cellphone users. Related work on plan recognition (Kautz and Allen 1986; Wilensky 1983; Colbry, Peintner, and Pollack 2002), visionbased activity recognition (Jebara and Pentland 1999; Boger et al. 2006; Shi et al. 2004), and object-interaction-based activity recognition (Perkowitz et al. 2004; Philipose et al. 2004) formulates activities as collections of stereotypical sequential actions. When these approaches are applied to real-world data, they tend to be successful in modelling low-level behaviors in controlled environments or in very domain-specific applications (e.g., hand-washing). We believe that when applied to global scale environments these Copyright c © 2009, Association for the Advancement of Artificial Intelligence (www.aaai.org). All rights reserved. approaches will be of limited success. The wide variety of valid approaches to achieving a goal (executing an activity) and correspondingly, the wide variety of reasons why such an activity might want to be recognized, make a global statebased recognition system a daunting pursuit. For example, successfully recognizing that someone has “made tea” has important modelling implications if the application automatically calls social partners to the table as opposed to ordering new tea bags or monitoring caffeine intake. An alternative approach in the literature has been to treat activity recognition as a classification problem with less strict interpretation, if any at all, of the steps involved. This includes unimodal evaluations of activity and social context from audio (Choudhury and Pentland 2003; Stager et al. 2003; Stager, Lukowicz, and Troster 2004), video (Fitzpatrick and Kemp 2003), accelerometers (K. Van Laerhoven and Gellersen 2004) and RFID (Patterson et al. 2005). Although in this work we also approach the activity recognition task as one of classification, we emphasize multi-modal sensor evaluations like (Kern et al. 2004) and (Choudhury, Lester, and Borriello 2005) in order to avoid biasing our effectiveness on activities that are easily discriminated by one sensor (e.g., choosing activities such as “hammering” or “grinding coffee” because sensing is done with a microphone). In this paper we describe a system for collecting place and activity data from users. We characterize the existing data and propose two models for smoothing the estimations of place and activity names using loopy belief propagation. Additionally we suggest some relevant extensions that are necessary in order to account for application specific effects. Our goal is to use our data to create a dense map of priors over place and activity labels for every location on earth." @default.
• W2399218780 created "2016-06-24" @default.
• W2399218780 creator A5006649277 @default.
• W2399218780 date "2009-01-01" @default.
• W2399218780 modified "2023-09-27" @default.
• W2399218780 title "Global Priors of Place and Activity Tags." @default.
• W2399218780 cites W123994226 @default.
• W2399218780 cites W141859514 @default.
• W2399218780 cites W146707186 @default.
• W2399218780 cites W1491797991 @default.
• W2399218780 cites W1569107316 @default.
• W2399218780 cites W2003637680 @default.
• W2399218780 cites W2081695901 @default.
• W2399218780 cites W2117976903 @default.
• W2399218780 cites W2119380192 @default.
• W2399218780 cites W2132603770 @default.
• W2399218780 cites W2135787934 @default.
• W2399218780 cites W2136585082 @default.
• W2399218780 cites W2146370044 @default.
• W2399218780 cites W2148108980 @default.
• W2399218780 cites W2150338977 @default.
• W2399218780 cites W2152887363 @default.
• W2399218780 cites W2156136571 @default.
• W2399218780 cites W2157181132 @default.
• W2399218780 cites W2546817834 @default.
• W2399218780 cites W94200858 @default.
• W2399218780 hasPublicationYear "2009" @default.
• W2399218780 type Work @default.
• W2399218780 sameAs 2399218780 @default.
• W2399218780 citedByCount "1" @default.
• W2399218780 countsByYear W23992187802014 @default.
• W2399218780 crossrefType "proceedings-article" @default.
• W2399218780 hasAuthorship W2399218780A5006649277 @default.
• W2399218780 hasConcept C107457646 @default.
• W2399218780 hasConcept C111472728 @default.
• W2399218780 hasConcept C119857082 @default.
• W2399218780 hasConcept C121687571 @default.
• W2399218780 hasConcept C132829578 @default.
• W2399218780 hasConcept C136764020 @default.
• W2399218780 hasConcept C138885662 @default.
• W2399218780 hasConcept C154945302 @default.
• W2399218780 hasConcept C23123220 @default.
• W2399218780 hasConcept C2522767166 @default.
• W2399218780 hasConcept C25810664 @default.
• W2399218780 hasConcept C2776214188 @default.
• W2399218780 hasConcept C31972630 @default.
• W2399218780 hasConcept C3770464 @default.
• W2399218780 hasConcept C41008148 @default.
• W2399218780 hasConceptScore W2399218780C107457646 @default.
• W2399218780 hasConceptScore W2399218780C111472728 @default.
• W2399218780 hasConceptScore W2399218780C119857082 @default.
• W2399218780 hasConceptScore W2399218780C121687571 @default.
• W2399218780 hasConceptScore W2399218780C132829578 @default.
• W2399218780 hasConceptScore W2399218780C136764020 @default.
• W2399218780 hasConceptScore W2399218780C138885662 @default.
• W2399218780 hasConceptScore W2399218780C154945302 @default.
• W2399218780 hasConceptScore W2399218780C23123220 @default.
• W2399218780 hasConceptScore W2399218780C2522767166 @default.
• W2399218780 hasConceptScore W2399218780C25810664 @default.
• W2399218780 hasConceptScore W2399218780C2776214188 @default.
• W2399218780 hasConceptScore W2399218780C31972630 @default.
• W2399218780 hasConceptScore W2399218780C3770464 @default.
• W2399218780 hasConceptScore W2399218780C41008148 @default.
• W2399218780 hasLocation W23992187801 @default.
• W2399218780 hasOpenAccess W2399218780 @default.
• W2399218780 hasPrimaryLocation W23992187801 @default.
• W2399218780 hasRelatedWork W1572192454 @default.
• W2399218780 hasRelatedWork W2018671208 @default.
• W2399218780 hasRelatedWork W2037532757 @default.
• W2399218780 hasRelatedWork W2108335620 @default.
• W2399218780 hasRelatedWork W2246318539 @default.
• W2399218780 hasRelatedWork W2292669296 @default.
• W2399218780 hasRelatedWork W2516402518 @default.
• W2399218780 hasRelatedWork W2574048019 @default.
• W2399218780 hasRelatedWork W2799320946 @default.
• W2399218780 hasRelatedWork W2894793156 @default.
• W2399218780 hasRelatedWork W2937608033 @default.
• W2399218780 hasRelatedWork W2945952120 @default.
• W2399218780 hasRelatedWork W2967300251 @default.
• W2399218780 hasRelatedWork W2969451795 @default.
• W2399218780 hasRelatedWork W3090764769 @default.
• W2399218780 hasRelatedWork W3127090497 @default.
• W2399218780 hasRelatedWork W3139423735 @default.
• W2399218780 hasRelatedWork W3185436414 @default.
• W2399218780 hasRelatedWork W55152634 @default.
• W2399218780 hasRelatedWork W1961590190 @default.
• W2399218780 isParatext "false" @default.
• W2399218780 isRetracted "false" @default.
• W2399218780 magId "2399218780" @default.
• W2399218780 workType "article" @default.