Privacy-Preserving ML for Analyzing Usage Telemetry to Improve Software Reliability without Exposing User Data

  • Authors

    • Sanjay Verma Operations Manager, HCL Technologies, India. Author
    • Naveen Kumar Product Manager, Zoho Corporation, India. Author

    DOI:

    https://doi.org/10.67228/3142788X/IJMLPA-2021PI6D3K

    Published 03-02-2021

  • Privacy-Preserving Machine Learning, Usage Telemetry, Software Reliability, Federated Learning, Differential Privacy, Anomaly Detection, Privacy-Aware Software Engineering

    Issue

    Vol. 4 No. 1 (2021)

    Section

    Articles

    How to Cite

    [1]
    S. Verma and N. Kumar, “Privacy-Preserving ML for Analyzing Usage Telemetry to Improve Software Reliability without Exposing User Data”, IJMLPA, vol. 4, no. 1, pp. 01–10, Mar. 2021, doi: 10.67228/3142788X/IJMLPA-2021PI6D3K.

  • Abstract

    Ensuring software reliability is critical in modern applications, yet analyzing usage telemetry often involves sensitive user data, raising significant privacy concerns. This paper proposes a privacy-preserving machine learning framework for analyzing usage telemetry to improve software reliability without exposing individual user information. We explore techniques such as federated learning and differential privacy to enable collaborative model training while keeping raw data localized. Our approach leverages telemetry data to predict software failures, detect anomalies, and identify reliability hotspots. Experimental results demonstrate that our framework achieves competitive predictive performance while maintaining strong privacy guarantees, offering a practical solution for privacy-aware software reliability analysis.

  • References

    [1] B. Ding, J. Kulkarni, and S. Yekhanin, Collecting Telemetry Data Privately, arXiv:1712.01524 (2017). arXiv

    [2] R. Xu, N. Baracaldo, and J. Joshi, Privacy-Preserving Machine Learning: Methods, Challenges and Directions, arXiv:2108.04417 (2021). arXiv

    [3] T. Wang, A Comprehensive Survey on Local Differential Privacy, Sensors, 20(24):7030 (2020). MDPI

    [4] M. Naseri, J. Hayes & E. De Cristofaro, Local and Central Differential Privacy for Robustness and Privacy in Federated Learning, arXiv:2009.03561 (2020). arXiv

    [5] P. Kairouz, Z. Liu & T. Steinke, The Composition Theorem for Differential Privacy, Advances in Neural Information Processing Systems (2013). Wikipedia

    [6] Bolin Ding, Janardhan Kulkarni, and Sergey Yekhanin, "Collecting Telemetry Data Privately," Advances in Neural Information Processing Systems (NeurIPS), vol. 30, pp. 3574–3583, 2017.

    [7] Benjamin Baron and Mirco Musolesi, "Interpretable Machine Learning for Privacy-Preserving Pervasive Systems," IEEE Pervasive Computing, vol. 19, no. 1, pp. 73–82, Jan.–Mar. 2020.

    [8] Dayeol Lee, Dmitrii Kuvaiskii, Anjo Vahldiek-Oberwagner, and Mona Vij, "Privacy-Preserving Machine Learning in Untrusted Clouds Made Simple," arXiv preprint arXiv:2009.04390, 2020.

    [9] Sina Shaham, Ming Ding, Bo Liu, Shuping Dang, Zihuai Lin, and Jun Li, "Privacy Preserving Location Data Publishing: A Machine Learning Approach," IEEE Transactions on Knowledge and Data Engineering, vol. 33, no. 4, pp. 1290–1305, 2021 (online first, 2020).

    [10] Yanqing Yang, Qinghua Zheng, and co-authors, "Differentially Private Model Publishing in Cyber Physical Systems," Future Generation Computer Systems, vol. 108, pp. 1297–1306, 2020.

    [11] Arunima Jaiswal and Ruchika Malhotra, "Software Reliability Prediction Using Machine Learning Techniques," International Journal of System Assurance Engineering and Management, vol. 8, no. 1, pp. 230–244, 2017.

    [12] Cynthia Dwork and Aaron Roth, "The Algorithmic Foundations of Differential Privacy," Foundations and Trends in Theoretical Computer Science, vol. 9, nos. 3–4, pp. 211–407, 2014.

  • Downloads