Recommendations with Minimum Exposure Guarantees: A Post-Processing Framework. Lopes, R., Alves, R., Ledent, A., Santos, R., & Kloft, M. November 2022. ![Recommendations with Minimum Exposure Guarantees: A Post-Processing Framework [link]](https://bibbase.org/img/filetypes/link.svg "link")
Paper doi abstract bibtex 1 download Relevance-based ranking is a popular ingredient in recommenders, but it frequently struggles to meet fairness criteria because social and cultural norms may favor some item groups over others. For instance, some items might receive lower ratings due to some sort of bias (e.g. gender bias). A fair ranking should balance the exposure of items from advantaged and disadvantagedgroups. To this end, we propose a novel post-processing framework to produce fair, exposure-aware recommendations. Our approach is based on an integer linear programming model maximizing the expected utility while satisfying a minimum exposure constraint. The model has fewer variables than previous work and thus can be deployed to larger datasets and allows the organization to define a minimum level of exposure for groups of items. We conduct an extensive empirical evaluation indicating that our new framework can increase the exposure of items from disadvantaged groups at a small cost of recommendation accuracy.
@unpublished{lopes_recommendations_2022,
title = {Recommendations with {Minimum} {Exposure} {Guarantees}: {A} {Post}-{Processing} {Framework}},
url = {https://papers.ssrn.com/abstract=4274780},
abstract = {Relevance-based ranking is a popular ingredient in recommenders, but it
frequently struggles to meet fairness criteria because social and cultural
norms may favor some item groups over others. For instance, some items
might receive lower ratings due to some sort of bias (e.g. gender bias). A
fair ranking should balance the exposure of items from advantaged and
disadvantagedgroups. To this end, we propose a novel post-processing
framework to produce fair, exposure-aware recommendations. Our approach is
based on an integer linear programming model maximizing the expected
utility while satisfying a minimum exposure constraint. The model has
fewer variables than previous work and thus can be deployed to larger
datasets and allows the organization to define a minimum level of exposure
for groups of items. We conduct an extensive empirical evaluation
indicating that our new framework can increase the exposure of items from
disadvantaged groups at a small cost of recommendation accuracy.},
urldate = {2022-12-23},
author = {Lopes, Ramon and Alves, Rodrigo and Ledent, Antoine and Santos, Rodrygo and Kloft, Marius},
month = nov,
year = {2022},
doi = {10.2139/ssrn.4274780},
keywords = {recommender systems, fairness, exposure, integer linear programming},
}
Downloads: 1
{"_id":"GLqLfgnqKiXzeuAph","bibbaseid":"lopes-alves-ledent-santos-kloft-recommendationswithminimumexposureguaranteesapostprocessingframework-2022","author_short":["Lopes, R.","Alves, R.","Ledent, A.","Santos, R.","Kloft, M."],"bibdata":{"bibtype":"unpublished","type":"unpublished","title":"Recommendations with Minimum Exposure Guarantees: A Post-Processing Framework","url":"https://papers.ssrn.com/abstract=4274780","abstract":"Relevance-based ranking is a popular ingredient in recommenders, but it frequently struggles to meet fairness criteria because social and cultural norms may favor some item groups over others. For instance, some items might receive lower ratings due to some sort of bias (e.g. gender bias). A fair ranking should balance the exposure of items from advantaged and disadvantagedgroups. To this end, we propose a novel post-processing framework to produce fair, exposure-aware recommendations. Our approach is based on an integer linear programming model maximizing the expected utility while satisfying a minimum exposure constraint. The model has fewer variables than previous work and thus can be deployed to larger datasets and allows the organization to define a minimum level of exposure for groups of items. We conduct an extensive empirical evaluation indicating that our new framework can increase the exposure of items from disadvantaged groups at a small cost of recommendation accuracy.","urldate":"2022-12-23","author":[{"propositions":[],"lastnames":["Lopes"],"firstnames":["Ramon"],"suffixes":[]},{"propositions":[],"lastnames":["Alves"],"firstnames":["Rodrigo"],"suffixes":[]},{"propositions":[],"lastnames":["Ledent"],"firstnames":["Antoine"],"suffixes":[]},{"propositions":[],"lastnames":["Santos"],"firstnames":["Rodrygo"],"suffixes":[]},{"propositions":[],"lastnames":["Kloft"],"firstnames":["Marius"],"suffixes":[]}],"month":"November","year":"2022","doi":"10.2139/ssrn.4274780","keywords":"recommender systems, fairness, exposure, integer linear programming","bibtex":"@unpublished{lopes_recommendations_2022,\n\ttitle = {Recommendations with {Minimum} {Exposure} {Guarantees}: {A} {Post}-{Processing} {Framework}},\n\turl = {https://papers.ssrn.com/abstract=4274780},\n\tabstract = {Relevance-based ranking is a popular ingredient in recommenders, but it\nfrequently struggles to meet fairness criteria because social and cultural\nnorms may favor some item groups over others. For instance, some items\nmight receive lower ratings due to some sort of bias (e.g. gender bias). A\nfair ranking should balance the exposure of items from advantaged and\ndisadvantagedgroups. To this end, we propose a novel post-processing\nframework to produce fair, exposure-aware recommendations. Our approach is\nbased on an integer linear programming model maximizing the expected\nutility while satisfying a minimum exposure constraint. The model has\nfewer variables than previous work and thus can be deployed to larger\ndatasets and allows the organization to define a minimum level of exposure\nfor groups of items. We conduct an extensive empirical evaluation\nindicating that our new framework can increase the exposure of items from\ndisadvantaged groups at a small cost of recommendation accuracy.},\n\turldate = {2022-12-23},\n\tauthor = {Lopes, Ramon and Alves, Rodrigo and Ledent, Antoine and Santos, Rodrygo and Kloft, Marius},\n\tmonth = nov,\n\tyear = {2022},\n\tdoi = {10.2139/ssrn.4274780},\n\tkeywords = {recommender systems, fairness, exposure, integer linear programming},\n}\n\n","author_short":["Lopes, R.","Alves, R.","Ledent, A.","Santos, R.","Kloft, M."],"key":"lopes_recommendations_2022","id":"lopes_recommendations_2022","bibbaseid":"lopes-alves-ledent-santos-kloft-recommendationswithminimumexposureguaranteesapostprocessingframework-2022","role":"author","urls":{"Paper":"https://papers.ssrn.com/abstract=4274780"},"keyword":["recommender systems","fairness","exposure","integer linear programming"],"metadata":{"authorlinks":{}},"downloads":1},"bibtype":"unpublished","biburl":"https://api.zotero.org/users/6655/collections/3TB3KT36/items?key=VFvZhZXIoHNBbzoLZ1IM2zgf&format=bibtex&limit=100","dataSources":["HB6fr7bPytW2CAAzC","7KNAjxiv2tsagmbgY"],"keywords":["recommender systems","fairness","exposure","integer linear programming"],"search_terms":["recommendations","minimum","exposure","guarantees","post","processing","framework","lopes","alves","ledent","santos","kloft"],"title":"Recommendations with Minimum Exposure Guarantees: A Post-Processing Framework","year":2022,"downloads":1}