Matches in SemOpenAlex for { <https://semopenalex.org/work/W3173192176> ?p ?o ?g. }
Showing items 1 to 82 of
82
with 100 items per page.
- W3173192176 endingPage "129" @default.
- W3173192176 startingPage "117" @default.
- W3173192176 abstract "Filter bubbles can have undesirable consequences on society by limiting user recommendations to those that already conform to their own beliefs rather than a diversity of opinions. In this work, we consider the problem of enabling users to choose which filter bubbles they want to be in and which ones they want to eliminate. We propose a method of eliminating filter bubbles in user-selected topics by unbiasing them with respect to a certain attribute, e.g., political inclination while exempting biases in the remaining topics/categories. We note that an attempt to make all the recommendations free of biases without any provision for exemptions can be highly restrictive. E.g., a user might be invested in a particular topic, e.g., climate change, and want to be in a filter bubble that motivates them. For other topics, however, the user might want to avoid filter bubbles and learn from all sides of the debate. We first propose a set of desirable properties to quantify bias in recommendations with topic exemptions, and then arrive at a simple information-theoretic measure – Value-based Conditional Mutual Information (VCMI) – that satisfies our proposed desirable properties. Next, we also demonstrate how one can use this measure in practice to train models that eliminate biases in recommendations while exempting certain biases based on preferred topics/categories. We consider two datasets: (i) a completely synthetic dataset; and (ii) a dataset that we create based on the publicly available Twitter News Sharing User Behaviour dataset." @default.
- W3173192176 created "2021-07-05" @default.
- W3173192176 creator A5035222388 @default.
- W3173192176 creator A5054746425 @default.
- W3173192176 creator A5071770749 @default.
- W3173192176 creator A5079787466 @default.
- W3173192176 date "2021-01-01" @default.
- W3173192176 modified "2023-10-01" @default.
- W3173192176 title "An Information-Theoretic Measure for Enabling Category Exemptions with an Application to Filter Bubbles" @default.
- W3173192176 cites W1686065478 @default.
- W3173192176 cites W2037451280 @default.
- W3173192176 cites W2083685419 @default.
- W3173192176 cites W2090883204 @default.
- W3173192176 cites W2127109099 @default.
- W3173192176 cites W2146950091 @default.
- W3173192176 cites W2155653793 @default.
- W3173192176 cites W2214169095 @default.
- W3173192176 cites W2476960546 @default.
- W3173192176 cites W2584805976 @default.
- W3173192176 cites W2774443469 @default.
- W3173192176 cites W2911936302 @default.
- W3173192176 cites W2932004331 @default.
- W3173192176 cites W2948130259 @default.
- W3173192176 cites W2964261743 @default.
- W3173192176 cites W2997892292 @default.
- W3173192176 cites W3034161109 @default.
- W3173192176 cites W3035446616 @default.
- W3173192176 cites W4299687421 @default.
- W3173192176 cites W4313003290 @default.
- W3173192176 doi "https://doi.org/10.1007/978-3-030-78818-6_11" @default.
- W3173192176 hasPublicationYear "2021" @default.
- W3173192176 type Work @default.
- W3173192176 sameAs 3173192176 @default.
- W3173192176 citedByCount "0" @default.
- W3173192176 crossrefType "book-chapter" @default.
- W3173192176 hasAuthorship W3173192176A5035222388 @default.
- W3173192176 hasAuthorship W3173192176A5054746425 @default.
- W3173192176 hasAuthorship W3173192176A5071770749 @default.
- W3173192176 hasAuthorship W3173192176A5079787466 @default.
- W3173192176 hasConcept C106131492 @default.
- W3173192176 hasConcept C124101348 @default.
- W3173192176 hasConcept C127413603 @default.
- W3173192176 hasConcept C177264268 @default.
- W3173192176 hasConcept C188198153 @default.
- W3173192176 hasConcept C199360897 @default.
- W3173192176 hasConcept C23123220 @default.
- W3173192176 hasConcept C2522767166 @default.
- W3173192176 hasConcept C2780009758 @default.
- W3173192176 hasConcept C31972630 @default.
- W3173192176 hasConcept C41008148 @default.
- W3173192176 hasConcept C78519656 @default.
- W3173192176 hasConceptScore W3173192176C106131492 @default.
- W3173192176 hasConceptScore W3173192176C124101348 @default.
- W3173192176 hasConceptScore W3173192176C127413603 @default.
- W3173192176 hasConceptScore W3173192176C177264268 @default.
- W3173192176 hasConceptScore W3173192176C188198153 @default.
- W3173192176 hasConceptScore W3173192176C199360897 @default.
- W3173192176 hasConceptScore W3173192176C23123220 @default.
- W3173192176 hasConceptScore W3173192176C2522767166 @default.
- W3173192176 hasConceptScore W3173192176C2780009758 @default.
- W3173192176 hasConceptScore W3173192176C31972630 @default.
- W3173192176 hasConceptScore W3173192176C41008148 @default.
- W3173192176 hasConceptScore W3173192176C78519656 @default.
- W3173192176 hasLocation W31731921761 @default.
- W3173192176 hasOpenAccess W3173192176 @default.
- W3173192176 hasPrimaryLocation W31731921761 @default.
- W3173192176 hasRelatedWork W1509467138 @default.
- W3173192176 hasRelatedWork W2115485936 @default.
- W3173192176 hasRelatedWork W2128719260 @default.
- W3173192176 hasRelatedWork W2144190808 @default.
- W3173192176 hasRelatedWork W2357241418 @default.
- W3173192176 hasRelatedWork W2366644548 @default.
- W3173192176 hasRelatedWork W2376314740 @default.
- W3173192176 hasRelatedWork W2384888906 @default.
- W3173192176 hasRelatedWork W2899084033 @default.
- W3173192176 hasRelatedWork W341479092 @default.
- W3173192176 isParatext "false" @default.
- W3173192176 isRetracted "false" @default.
- W3173192176 magId "3173192176" @default.
- W3173192176 workType "book-chapter" @default.