Fighting Fire with Fire: Using Antidote Data to Improve โฆ
WEBDec 2, 2018 — We formalize the antidote data problem, and develop optimization-based solutions. We take as our model system the matrix factorization approach to recommendation, and we propose a set of measures to capture the polarization or โฆ
- We start by presenting the system setup, notation, and problem definition. Assume X โ Rn×d is a partially observed rating matrix ofn users andditems such that elementxijdenotes the rating given by user to item j. Let ฮฉ be the set of indices of known ratings in i X. Also ฮฉi denotes the indices of known item ratings for user i, and ฮฉ denotes the indi...
- Author: Bashir Rastegarpanah, Krishna P. Gummadi, Mark Crovella
- Publish Year: 2019
Fighting Fire with Fire: Using Antidote Data to Improve โฆ
WEBJan 30, 2019 — As an alternative, in this paper we introduce the idea of improving the social desirability of recommender system outputs by adding more data to the input, an โฆ
Fighting Fire with Fire: Using Antidote Data to Improve โฆ
WEBDec 2, 2018 — This paper formalizes the antidote data problem, and develops optimization-based solutions, and takes as its model system the matrix factorization approach to โฆ
Fighting Fire with Fire: Using Antidote Data to Improve โฆ
WEBDec 2, 2018 — Reducing polarization and increasing diverse navigability in graphs by inserting edges and swapping edge weights. ... Recommender systems serve a โฆ
Fighting Fire with Fire: Using Antidote Data to Improve โฆ
WEBOur results show that the antidote data generation framework can successfully either minimize or maximize polarization. Antidote data generated by our method are โฆ
Fighting Fire with Fire: Using Antidote Data to Improve โฆ
WEBFighting Fire with Fire: Using Antidote Data to Improve Polarization and Fairness of Recommender Systems
Fighting Fire with Fire: Using Antidote Data to Improve โฆ
WEBRastegarpanah, B., Gummadi, K., & Crovella, M. (2019). Fighting Fire with Fire: Using Antidote Data to Improve Polarization and Fairness of Recommender Systems. In โฆ
WEBIn this section, we first discuss how our measures of fairness and polarization for recommender systems relate to those discussed in prior works. Later, we describe โฆ
Fighting Fire with Fire: Using Antidote Data to Improve โฆ
WEBThe increasing role of recommender systems in many aspects of society makes it essential to consider how such systems may impact social good. Various modifications โฆ
- Some results have been removed
Skip to content