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

Rate-Optimal Rank Aggregation with Private Pairwise Rankings (2402.16792v2)

Published 26 Feb 2024 in stat.ML, cs.CR, and cs.LG

Abstract: In various real-world scenarios, such as recommender systems and political surveys, pairwise rankings are commonly collected and utilized for rank aggregation to obtain an overall ranking of items. However, preference rankings can reveal individuals' personal preferences, underscoring the need to protect them from being released for downstream analysis. In this paper, we address the challenge of preserving privacy while ensuring the utility of rank aggregation based on pairwise rankings generated from a general comparison model. Using the randomized response mechanism to perturb raw pairwise rankings is a common privacy protection strategy used in practice. However, a critical challenge arises because the privatized rankings no longer adhere to the original model, resulting in significant bias in downstream rank aggregation tasks. Motivated by this, we propose to adaptively debiasing the rankings from the randomized response mechanism, ensuring consistent estimation of true preferences and enhancing the utility of downstream rank aggregation. Theoretically, we offer insights into the relationship between overall privacy guarantees and estimation errors from private ranking data, and establish minimax rates for estimation errors. This enables the determination of optimal privacy guarantees that balance consistency in rank aggregation with privacy protection. We also investigate convergence rates of expected ranking errors for partial and full ranking recovery, quantifying how privacy protection influences the specification of top-$K$ item sets and complete rankings. Our findings are validated through extensive simulations and a real application.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (57)
  1. Elections with partially ordered preferences. Public Choice, 157:145–168.
  2. Private rank aggregation in central and local models. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 36, pages 5984–5991.
  3. Local differential privacy for deep learning. IEEE Internet of Things Journal, 7(7):5827–5842.
  4. Distribution-invariant differential privacy. Journal of econometrics, 235(2):444–453.
  5. Rank analysis of incomplete block designs: I. the method of paired comparisons. Biometrika, 39(3/4):324–345.
  6. The cost of privacy: Optimal rates of convergence for parameter estimation with differential privacy. The Annals of Statistics, 49(5):2825–2850.
  7. A pairwise comparison framework for fast, flexible, and reliable human coding of political texts. American Political Science Review, 111(4):835–843.
  8. Partial recovery for top-k𝑘kitalic_k ranking: optimality of MLE, and suboptimality of the spectral method. The Annals of Statistics, 50(3):1618–1652.
  9. Spectral method and regularized mle are both optimal for top-k ranking. Annals of statistics, 47(4):2204.
  10. Spectral mle: Top-k rank aggregation from pairwise comparisons. In International Conference on Machine Learning, pages 371–380. PMLR.
  11. Robust estimation of discrete distributions under local differential privacy. In International Conference on Algorithmic Learning Theory, pages 411–446. PMLR.
  12. Privacy-aware web service composition and ranking. In 2013 IEEE 20th International Conference on Web Services, pages 131–138. IEEE.
  13. Gaussian differential privacy. Journal of the Royal Statistical Society Series B, 84(1):3–37.
  14. Minimax optimal procedures for locally private estimation. Journal of the American Statistical Association, 113(521):182–201.
  15. Dwork, C. (2006). Differential privacy. In International Colloquium on Automata, Languages, and Programming, pages 1–12. Springer.
  16. Rank aggregation methods for the web. In Proceedings of the 10th International Conference on World Wide Web, pages 613–622.
  17. The algorithmic foundations of differential privacy. Foundations and Trends® in Theoretical Computer Science, 9(3–4):211–407.
  18. Rappor: Randomized aggregatable privacy-preserving ordinal response. In Proceedings of the 2014 ACM SIGSAC Conference on Computer and Communications Security, pages 1054–1067.
  19. Dplcf: differentially private local collaborative filtering. In Proceedings of the 43rd International ACM SIGIR Conference on Research and Development in Information Retrieval, pages 961–970.
  20. Uncertainty quantification in the bradley–terry–luce model. Information and Inference: A Journal of the IMA, 12(2):1073–1140.
  21. Minimax-optimal inference from partial rankings. Advances in Neural Information Processing Systems, 27.
  22. The movielens datasets: History and context. Acm transactions on interactive intelligent systems (tiis), 5(4):1–19.
  23. Differentially private rank aggregation. In Proceedings of the 2017 SIAM International Conference on Data Mining, pages 669–677. SIAM.
  24. Ichihashi, S. (2020). Online privacy and information disclosure by consumers. American Economic Review, 110(2):569–595.
  25. Ranking recovery under privacy considerations. Transactions on Machine Learning Research.
  26. Eliciting pairwise preferences in recommender systems. In Proceedings of the 12th ACM Conference on Recommender Systems, pages 329–337.
  27. Learning to rank for recommender systems. In Proceedings of the 7th ACM Conference on Recommender Systems, pages 493–494.
  28. Dynamic ranking with the btl model: a nearest neighbor based rank centrality method. Journal of Machine Learning Research, 24(269):1–57.
  29. Concentration around the mean for maxima of empirical processes. The Annals of Probability, 33(3):1060–1077.
  30. Koltchinskii, V. (2011). Oracle inequalities in empirical risk minimization and sparse recovery problems: École D’Été de Probabilités de Saint-Flour XXXVIII-2008, volume 2033. Springer Science & Business Media.
  31. Lee, D. T. (2015). Efficient, private, and eps-strategyproof elicitation of tournament voting rules. In Twenty-Fourth International Joint Conference on Artificial Intelligence.
  32. Differentially private condorcet voting. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 37, pages 5755–5763.
  33. Lagrangian inference for ranking problems. Operations Research, 71(1):202–223.
  34. Supervised rank aggregation. In Proceedings of the 16th International Conference on World Wide Web, pages 481–490.
  35. Luce, R. D. (2012). Individual choice behavior: A theoretical analysis. Courier Corporation.
  36. Ranked choice voting as a generational issue in modern american politics. Politics & Policy, 49(1):33–60.
  37. Differentially private recommender systems: Building privacy into the netflix prize contenders. In Proceedings of the 15th ACM SIGKDD international conference on Knowledge discovery and data mining, pages 627–636.
  38. Rank centrality: Ranking from pairwise com-parisons. Operations Research, 65(1):266–287.
  39. Nielson, L. (2017). Ranked choice voting and attitudes toward democracy in the united states: Results from a survey experiment. Politics & Policy, 45(4):535–570.
  40. Consumer privacy concerns and preference for degree of regulatory control. Journal of Advertising, 38(4):63–77.
  41. Is rank aggregation effective in recommender systems? an experimental analysis. ACM Transactions on Intelligent Systems and Technology (TIST), 11(2):1–26.
  42. Simple, robust and optimal ranking from pairwise comparisons. Journal of Machine Learning Research, 18(199):1–38.
  43. The application of differential privacy for rank aggregation: Privacy and accuracy. In 17th International Conference on Information Fusion (FUSION), pages 1–7. IEEE.
  44. Boosting data analytics with synthetic volume expansion. arXiv preprint arXiv:2310.17848.
  45. Adversarial top-k𝑘kitalic_k ranking. IEEE Transactions on Information Theory, 63(4):2201–2225.
  46. Online rank elicitation for plackett-luce: A dueling bandits approach. Advances in Neural Information Processing Systems, 28.
  47. Privacy loss in apple’s implementation of differential privacy on macos 10.12. arXiv preprint arXiv:1709.02753.
  48. Rényi divergence and kullback-leibler divergence. IEEE Transactions on Information Theory, 60(7):3797–3820.
  49. On sparse linear regression in the local differential privacy model. In International Conference on Machine Learning, pages 6628–6637. PMLR.
  50. Locally differentially private protocols for frequency estimation. In 26th USENIX Security Symposium (USENIX Security 17), pages 729–745.
  51. Warner, S. L. (1965). Randomized response: A survey technique for eliminating evasive answer bias. Journal of the American Statistical Association, 60(309):63–69.
  52. A statistical framework for differential privacy. Journal of the American Statistical Association, 105(489):375–389.
  53. A ranking model motivated by nonnegative matrix factorization with applications to tennis tournaments. In Machine Learning and Knowledge Discovery in Databases: European Conference, ECML PKDD 2019, Würzburg, Germany, September 16–20, 2019, Proceedings, Part III, pages 187–203. Springer.
  54. Binary classification under local label differential privacy using randomized response mechanisms. Transactions on Machine Learning Research.
  55. Private rank aggregation under local differential privacy. International Journal of Intelligent Systems, 35(10):1492–1519.
  56. Local differential privacy and its applications: A comprehensive survey. arXiv preprint arXiv:2008.03686.
  57. Principled reinforcement learning with human feedback from pairwise or k𝑘kitalic_k-wise comparisons. arXiv preprint arXiv:2301.11270.
User Edit Pencil Streamline Icon: https://streamlinehq.com
Authors (3)
  1. Shirong Xu (8 papers)
  2. Will Wei Sun (32 papers)
  3. Guang Cheng (136 papers)

Summary

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

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

Tweets