Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
119 tokens/sec
GPT-4o
56 tokens/sec
Gemini 2.5 Pro Pro
43 tokens/sec
o3 Pro
6 tokens/sec
GPT-4.1 Pro
47 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Break Out of a Pigeonhole: A Unified Framework for Examining Miscalibration, Bias, and Stereotype in Recommender Systems (2312.17443v1)

Published 29 Dec 2023 in cs.IR, cs.AI, and cs.LG

Abstract: Despite the benefits of personalizing items and information tailored to users' needs, it has been found that recommender systems tend to introduce biases that favor popular items or certain categories of items, and dominant user groups. In this study, we aim to characterize the systematic errors of a recommendation system and how they manifest in various accountability issues, such as stereotypes, biases, and miscalibration. We propose a unified framework that distinguishes the sources of prediction errors into a set of key measures that quantify the various types of system-induced effects, both at the individual and collective levels. Based on our measuring framework, we examine the most widely adopted algorithms in the context of movie recommendation. Our research reveals three important findings: (1) Differences between algorithms: recommendations generated by simpler algorithms tend to be more stereotypical but less biased than those generated by more complex algorithms. (2) Disparate impact on groups and individuals: system-induced biases and stereotypes have a disproportionate effect on atypical users and minority groups (e.g., women and older users). (3) Mitigation opportunity: using structural equation modeling, we identify the interactions between user characteristics (typicality and diversity), system-induced effects, and miscalibration. We further investigate the possibility of mitigating system-induced effects by oversampling underrepresented groups and individuals, which was found to be effective in reducing stereotypes and improving recommendation quality. Our research is the first systematic examination of not only system-induced effects and miscalibration but also the stereotyping issue in recommender systems.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (48)
  1. Controlling Popularity Bias in Learning-to-Rank Recommendation. In Proceedings of the Eleventh ACM Conference on Recommender Systems. ACM, 42–46.
  2. The Impact of Popularity Bias on Fairness and Calibration in Recommendation. http://arxiv.org/abs/1910.05755 arXiv:1910.05755 [cs].
  3. The Connection Between Popularity Bias, Calibration, and Fairness in Recommendation. In Fourteenth ACM Conference on Recommender Systems. ACM, Virtual Event Brazil, 726–731. https://doi.org/10.1145/3383313.3418487
  4. Nourah A ALRossais and Daniel Kudenko. 2018. Evaluating Stereotype and Non-Stereotype Recommender Systems.. In KaRS@ RecSys. 23–28.
  5. Exposure to ideologically diverse news and opinion on Facebook. Science 348, 6239 (2015), 1130–1132.
  6. The Minority Matters: A Diversity-Promoting Collaborative Metric Learning Algorithm. http://arxiv.org/abs/2209.15292 arXiv:2209.15292 [cs].
  7. Stereotype and most-popular recommendations in the digital library sowiport. Humboldt-Universität zu Berlin.
  8. Lawrence Blum. 2004. Stereotypes And Stereotyping: A Moral Analysis. Philosophical Papers 33, 3 (Nov. 2004), 251–289. https://doi.org/10.1080/05568640409485143
  9. A Bias-Variance Decomposition for Bayesian Deep Learning. ([n. d.]).
  10. A bias-variance decomposition for Bayesian deep learning. In NeurIPS 2019 Workshop on Bayesian Deep Learning.
  11. Òscar Celma and Pedro Cano. 2008. From hits to niches?: or how popular artists can bias music recommendation and discovery. In Proceedings of the 2nd KDD Workshop on Large-Scale Recommender Systems and the Netflix Prize Competition. ACM, Las Vegas Nevada, 1–8. https://doi.org/10.1145/1722149.1722154
  12. How algorithmic confounding in recommendation systems increases homogeneity and decreases utility. In Proceedings of the 12th ACM Conference on Recommender Systems. ACM, Vancouver British Columbia Canada, 224–232. https://doi.org/10.1145/3240323.3240370
  13. Learning to Recommend Accurate and Diverse Items. In Proceedings of the 26th International Conference on World Wide Web. International World Wide Web Conferences Steering Committee, Perth Australia, 183–192. https://doi.org/10.1145/3038912.3052585
  14. All The Cool Kids, How Do They Fit In?: Popularity and Demographic Biases in Recommender Evaluation and Effectiveness. In Proceedings of the 1st Conference on Fairness, Accountability and Transparency. PMLR, 172–186. https://proceedings.mlr.press/v81/ekstrand18b.html ISSN: 2640-3498.
  15. Exploring author gender in book rating and recommendation. In Proceedings of the 12th ACM Conference on Recommender Systems. ACM, Vancouver British Columbia Canada, 242–250. https://doi.org/10.1145/3240323.3240373
  16. Gender stereotype reinforcement: Measuring the gender bias conveyed by ranking algorithms. Information Processing & Management 57, 6 (Nov. 2020), 102377. https://doi.org/10.1016/j.ipm.2020.102377
  17. Understanding Echo Chambers in E-commerce Recommender Systems. In Proceedings of the 43rd International ACM SIGIR Conference on Research and Development in Information Retrieval. ACM, Virtual Event China, 2261–2270. https://doi.org/10.1145/3397271.3401431
  18. Understanding echo chambers in e-commerce recommender systems. In Proceedings of the 43rd international ACM SIGIR conference on research and development in information retrieval. 2261–2270.
  19. The Stereotyping Problem in Collaboratively Filtered Recommender Systems. In Equity and Access in Algorithms, Mechanisms, and Optimization. ACM, – NY USA, 1–10. https://doi.org/10.1145/3465416.3483298
  20. Neural collaborative filtering. In Proceedings of the 26th international conference on world wide web. 173–182.
  21. Tom Heskes. 1998. Bias/Variance Decompositions for Likelihood-Based Estimators. Neural Computation 10, 6 (Aug. 1998), 1425–1433. https://doi.org/10.1162/089976698300017232
  22. Collaborative filtering for implicit feedback datasets. In 2008 Eighth IEEE international conference on data mining. Ieee, 263–272.
  23. Gareth M James. 2003. Variance and bias for general loss functions. Machine learning 51 (2003), 115–135.
  24. Kalervo Järvelin and Jaana Kekäläinen. 2002. Cumulated gain-based evaluation of IR techniques. ACM Transactions on Information Systems (TOIS) 20, 4 (2002), 422–446.
  25. Unequal Representation and Gender Stereotypes in Image Search Results for Occupations. In Proceedings of the 33rd Annual ACM Conference on Human Factors in Computing Systems. ACM, Seoul Republic of Korea, 3819–3828. https://doi.org/10.1145/2702123.2702520
  26. Bias plus variance decomposition for zero-one loss functions. In ICML, Vol. 96. Citeseer, 275–283.
  27. Crank up the volume: preference bias amplification in collaborative recommendation. http://arxiv.org/abs/1909.06362 arXiv:1909.06362 [cs, stat].
  28. Calibration in Collaborative Filtering Recommender Systems: a User-Centered Analysis. In Proceedings of the 31st ACM Conference on Hypertext and Social Media. ACM, Virtual Event USA, 197–206. https://doi.org/10.1145/3372923.3404793
  29. Amazon. com recommendations: Item-to-item collaborative filtering. IEEE Internet computing 7, 1 (2003), 76–80.
  30. Investigating Potential Factors Associated with Gender Discrimination in Collaborative Recommender Systems. http://arxiv.org/abs/2002.07786 arXiv:2002.07786 [cs].
  31. Masoud Mansoury and Robin Burke. 2019. Algorithm Selection with Librec-auto.. In AMIR@ ECIR. 11–17.
  32. Automating recommender systems experimentation with librec-auto. In Proceedings of the 12th ACM Conference on Recommender Systems. 500–501.
  33. Bias Disparity in Collaborative Recommendation: Algorithmic Evaluation and Comparison. http://arxiv.org/abs/1908.00831 arXiv:1908.00831 [cs].
  34. Umap: Uniform manifold approximation and projection for dimension reduction. arXiv preprint arXiv:1802.03426 (2018).
  35. Improving new user recommendations with rule-based induction on cold user data. In Proceedings of the 2007 ACM conference on Recommender systems. 121–128.
  36. Exploring the filter bubble: the effect of using recommender systems on content diversity. In Proceedings of the 23rd international conference on World wide web. ACM, Seoul Korea, 677–686. https://doi.org/10.1145/2566486.2568012
  37. BPR: Bayesian personalized ranking from implicit feedback. arXiv preprint arXiv:1205.2618 (2012).
  38. Grouplens: An open architecture for collaborative filtering of netnews. In Proceedings of the 1994 ACM conference on Computer supported cooperative work. 175–186.
  39. Harald Steck. 2018. Calibrated recommendations. In Proceedings of the 12th ACM Conference on Recommender Systems. ACM, Vancouver British Columbia Canada, 154–162. https://doi.org/10.1145/3240323.3240372
  40. Cass R Sunstein. 2009. Going to extremes: How like minds unite and divide. Oxford University Press.
  41. Bias Disparity in Recommendation Systems. arXiv:1811.01461 [cs] (Nov. 2018). http://arxiv.org/abs/1811.01461 arXiv: 1811.01461.
  42. Jodie B Ullman and Peter M Bentler. 2012. Structural equation modeling. Handbook of Psychology, Second Edition 2 (2012).
  43. Deconfounded Recommendation for Alleviating Bias Amplification. Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining (Aug. 2021), 1717–1725. https://doi.org/10.1145/3447548.3467249 arXiv: 2105.10648.
  44. User-controllable Recommendation Against Filter Bubbles. In Proceedings of the 45th International ACM SIGIR Conference on Research and Development in Information Retrieval. ACM, Madrid Spain, 1251–1261. https://doi.org/10.1145/3477495.3532075
  45. Tears or Fears? Comparing Gender Stereotypes about Movie Preferences to Actual Preferences. Frontiers in Psychology 8 (March 2017). https://doi.org/10.3389/fpsyg.2017.00428
  46. Challenging the Long Tail Recommendation. http://arxiv.org/abs/1205.6700 arXiv:1205.6700 [cs].
  47. Socially-aware self-supervised tri-training for recommendation. In Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining. 2084–2092.
  48. Improving recommendation lists through topic diversification. In Proceedings of the 14th international conference on World Wide Web - WWW ’05. ACM Press, Chiba, Japan, 22. https://doi.org/10.1145/1060745.1060754
User Edit Pencil Streamline Icon: https://streamlinehq.com
Authors (2)
  1. Yongsu Ahn (10 papers)
  2. Yu-Ru Lin (30 papers)
Citations (2)
View on Semantic Scholar

Summary

We haven't generated a summary for this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Now