pith. sign in

arxiv: 2606.10333 · v1 · pith:IUD7ODY5new · submitted 2026-06-09 · 💻 cs.LG · cs.CR

Privacy-Preserving Credit Risk Prediction with Alternative Data

Hongzhe Zhang , Jiarong Xu , Jing He , Xiao Fang This is my paper

Pith reviewed 2026-06-27 13:37 UTC · model grok-4.3

classification 💻 cs.LG cs.CR
keywords privacy-preserving machine learningcredit risk predictionalternative dataconsumer privacymodel confidentialitylossless learning
0
0 comments X

The pith

PrivacyCredit lets lenders combine traditional and alternative data for credit risk models while keeping consumer privacy and the model itself confidential and without any accuracy loss.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper defines a new problem where credit risk prediction must meet three constraints at once: protect consumer privacy from alternative data sources, keep the trained model confidential at the financial institution, and preserve full predictive performance. It introduces PrivacyCredit as a machine learning method that meets all three constraints simultaneously. The authors prove these properties theoretically and confirm through experiments on a linked real-world dataset that the method matches the accuracy of an insecure model trained on the combined plaintext data. A reader would care because it removes the privacy barrier that currently prevents use of richer alternative data like mobile records in lending decisions.

Core claim

PrivacyCredit is a privacy-preserving machine learning method that incorporates alternative data held by external parties such that the privacy-preserving constraint, the model-confidentiality constraint, and the lossless constraint all hold at the same time, allowing the learned model to achieve exactly the same predictive performance as one trained on the insecure plaintext combination of traditional and alternative data.

What carries the argument

PrivacyCredit, a privacy-preserving machine learning method that enables secure incorporation of alternative data while satisfying the three constraints.

If this is right

  • Lenders gain access to richer borrower profiles from mobile communication data without direct data sharing.
  • The final model remains stored and used only at the financial institution with no exposure of its internals.
  • Predictive performance equals that of models trained directly on the combined plaintext datasets.
  • Theoretical guarantees ensure the three constraints hold without trade-offs in the demonstrated setting.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Similar techniques could apply to other split-data prediction tasks such as insurance underwriting or fraud detection where privacy rules block data pooling.
  • The lossless property suggests the method might serve as a template for privacy-preserving tabular learning beyond credit scoring.
  • If the constraints can be met for one alternative data type, the same structure may extend to multiple external data holders without added performance cost.

Load-bearing premise

Alternative data can be incorporated through PrivacyCredit such that privacy protection, model confidentiality, and full predictive performance are achieved simultaneously with no degradation or leakage.

What would settle it

An experiment on the real-world linked credit dataset in which PrivacyCredit's accuracy or AUC falls measurably below the plaintext combined model's or in which private information can be recovered from the method's outputs.

read the original abstract

Credit risk prediction is a critical problem in the consumer credit industry. Traditionally, financial institutions construct credit risk prediction models using borrowers' demographic, financial, and credit history data, collectively referred to as traditional data. Recent studies have demonstrated that alternative data, such as borrowers' mobile phone communication data, enable lenders to acquire fuller and more accurate profiles of borrowers' creditworthiness, thereby improving credit risk prediction performance. Nevertheless, alternative data are held by external entities independent of financial institutions. Directly sharing alternative data with financial institutions infringe on consumer privacy, yet existing credit risk prediction studies largely overlook this issue. To address this gap, we define a new problem, namely privacy-preserving credit risk prediction with alternative data, which simultaneously considers three practical constraints: the privacy-preserving constraint that protects consumer privacy, the model-confidentiality constraint that learns and stores the model centrally at the financial institution, and the lossless constraint that maintains the performance of the learned model. To solve this problem, we develop PrivacyCredit, a novel privacy-preserving machine learning method. We then theoretically demonstrate the privacy-preserving, model-confidential, and lossless properties of PrivacyCredit. Through extensive experiments using a real-world credit dataset linked with alternative data, we demonstrate the predictive value of securely incorporating alternative data into credit risk prediction and show that PrivacyCredit achieves the same predictive performance as the model learned from the insecure plaintext combination of traditional and alternative data. We further evaluate its model-confidentiality property and computational efficiency.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The manuscript defines a new problem of privacy-preserving credit risk prediction that incorporates alternative data (e.g., mobile phone records) held by external parties while satisfying three simultaneous constraints: privacy preservation for consumers, model confidentiality (model learned and stored only at the financial institution), and lossless performance relative to the insecure plaintext union of traditional and alternative data. It proposes the PrivacyCredit method, provides theoretical arguments for the three properties, and reports experiments on a real-world linked credit dataset showing equivalent predictive performance to the plaintext baseline together with evaluations of confidentiality and efficiency.

Significance. If the theoretical arguments establish the three properties without hidden assumptions and the experiments confirm statistical equivalence on representative data, the result would be significant for regulated financial applications: it offers a concrete route to improve credit models with alternative data without direct data sharing or performance trade-offs. The explicit model-confidentiality requirement distinguishes the work from standard federated or secure-MPC approaches.

major comments (2)
  1. [§4] §4 (theoretical analysis): the lossless claim requires an explicit argument that the model parameters or decision function obtained under PrivacyCredit are identical to those obtained from the plaintext concatenation; if the proof relies on any cryptographic or statistical assumption, these must be stated and shown to hold for the credit-risk loss functions used.
  2. [§5] §5 (experiments): the statement that PrivacyCredit 'achieves the same predictive performance' must be supported by the exact metrics (AUC, F1, etc.), dataset sizes, train/test splits, and statistical tests (e.g., paired t-test or bootstrap confidence intervals) comparing PrivacyCredit to the plaintext baseline; without these, the equivalence claim cannot be verified.
minor comments (2)
  1. [Abstract] The abstract is information-dense; a short enumerated list of the three constraints would improve readability.
  2. [§2] Notation for the alternative-data holder and the financial institution should be introduced once and used consistently throughout.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the positive evaluation and recommendation of minor revision. The comments on the theoretical and experimental sections are helpful, and we address each one below with plans for clarification in the revised manuscript.

read point-by-point responses

  1. Referee: [§4] §4 (theoretical analysis): the lossless claim requires an explicit argument that the model parameters or decision function obtained under PrivacyCredit are identical to those obtained from the plaintext concatenation; if the proof relies on any cryptographic or statistical assumption, these must be stated and shown to hold for the credit-risk loss functions used.

    Authors: We agree that an explicit argument is required to establish that the model parameters are identical. In the revised manuscript we will expand the proof in §4 to directly show equivalence of the learned parameters (and thus the decision function) between PrivacyCredit and plaintext concatenation. The argument relies only on the algebraic equivalence of the secure computation steps to centralized gradient descent and holds for the logistic loss used in credit-risk modeling; no additional statistical or cryptographic assumptions beyond the standard secure-computation model are needed. We will state this explicitly. revision: yes

  2. Referee: [§5] §5 (experiments): the statement that PrivacyCredit 'achieves the same predictive performance' must be supported by the exact metrics (AUC, F1, etc.), dataset sizes, train/test splits, and statistical tests (e.g., paired t-test or bootstrap confidence intervals) comparing PrivacyCredit to the plaintext baseline; without these, the equivalence claim cannot be verified.

    Authors: We accept that the equivalence claim requires more granular reporting and statistical support. The revised §5 will include the precise AUC and F1 values, the exact dataset sizes, the train/test split ratios, and the results of paired t-tests (or bootstrap confidence intervals) confirming that performance differences are statistically insignificant. These details are already computed in our experimental pipeline and will be added to the text and tables. revision: yes

Circularity Check

0 steps flagged

No significant circularity identified

full rationale

The paper defines PrivacyCredit to satisfy three explicit constraints (privacy-preserving, model-confidential, lossless) by construction and then asserts that theoretical proofs establish these properties while experiments confirm equivalence to the plaintext baseline. No equations, self-citations, or fitted-parameter renamings are visible in the provided abstract that would reduce the central claims to tautological inputs. The theoretical demonstrations are presented as independent of the experimental results, and the lossless property is framed as a maintained performance guarantee rather than a post-hoc fit. This is the most common honest finding for a method paper whose core contribution is a new protocol whose correctness is asserted via separate proofs.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only review yields no explicit free parameters, axioms, or invented entities; the method is presented as satisfying three constraints without detailing underlying assumptions or new entities.

pith-pipeline@v0.9.1-grok · 5793 in / 1120 out tokens · 13429 ms · 2026-06-27T13:37:53.668426+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

300 extracted references · 5 linked inside Pith

  1. [1]

    IEEE Transactions on Dependable and Secure Computing , year=

    Differentially Private Publication of Vertically Partitioned Data , author=. IEEE Transactions on Dependable and Secure Computing , year=

  2. [2]

    IEEE transactions on dependable and secure computing , volume=

    Secure two-party differentially private data release for vertically partitioned data , author=. IEEE transactions on dependable and secure computing , volume=. 2013 , publisher=

    2013

  3. [3]

    The VLDB Journal , volume=

    Anonymity meets game theory: secure data integration with malicious participants , author=. The VLDB Journal , volume=. 2011 , publisher=

    2011

  4. [4]

    The VLDB journal , volume=

    A secure distributed framework for achieving k-anonymity , author=. The VLDB journal , volume=. 2006 , publisher=

    2006

  5. [5]

    Information Systems Research , volume=

    Anonymizing and sharing medical text records , author=. Information Systems Research , volume=. 2017 , publisher=

    2017

  6. [6]

    INFORMS Journal on Computing , volume=

    T-closeness slicing: A new privacy-preserving approach for transactional data publishing , author=. INFORMS Journal on Computing , volume=. 2018 , publisher=

    2018

  7. [7]

    IEEE Transactions on Knowledge and Data Engineering , volume=

    Differentially private mixture of generative neural networks , author=. IEEE Transactions on Knowledge and Data Engineering , volume=. 2018 , publisher=

    2018

  8. [8]

    ACM Transactions on Database Systems (TODS) , volume=

    Privbayes: Private data release via bayesian networks , author=. ACM Transactions on Database Systems (TODS) , volume=. 2017 , publisher=

    2017

  9. [9]

    Proceedings of the VLDB Endowment , volume=

    Plausible Deniability for Privacy-Preserving Data Synthesis , author=. Proceedings of the VLDB Endowment , volume=

  10. [10]

    Proceedings of the 17th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining , pages=

    Differentially private data release for data mining , author=. Proceedings of the 17th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining , pages=

  11. [11]

    Encyclopedia of Cryptography, Security and Privacy , pages=

    Differential privacy , author=. Encyclopedia of Cryptography, Security and Privacy , pages=. 2025 , publisher=

    2025

  12. [12]

    Proceedings of the 32nd international conference on Very large data bases , pages=

    Anatomy: Simple and effective privacy preservation , author=. Proceedings of the 32nd international conference on Very large data bases , pages=

  13. [13]

    Information Sciences , volume=

    A hybrid approach to prevent composition attacks for independent data releases , author=. Information Sciences , volume=. 2016 , publisher=

    2016

  14. [14]

    MIS Quarterly , volume=

    Digression and value concatenation to enable privacy-preserving regression , author=. MIS Quarterly , volume=

  15. [15]

    Operations Research , volume=

    Against classification attacks: A decision tree pruning approach to privacy protection in data mining , author=. Operations Research , volume=. 2009 , publisher=

    2009

  16. [16]

    IEEE Transactions on Knowledge and Data Engineering , volume=

    Slicing: A New Approach for Privacy Preserving Data Publishing , author=. IEEE Transactions on Knowledge and Data Engineering , volume=. 2012 , publisher=

    2012

  17. [17]

    Information Systems Research , volume=

    Protecting privacy against record linkage disclosure: A bounded swapping approach for numeric data , author=. Information Systems Research , volume=. 2011 , publisher=

    2011

  18. [18]

    Information Systems Research , volume=

    Privacy protection in data mining: A perturbation approach for categorical data , author=. Information Systems Research , volume=. 2006 , publisher=

    2006

  19. [19]

    2007 IEEE 23rd International Conference on Data Engineering , pages=

    t-closeness: Privacy beyond k-anonymity and l-diversity , author=. 2007 IEEE 23rd International Conference on Data Engineering , pages=. 2007 , organization=

    2007

  20. [20]

    International Journal of Uncertainty, Fuzziness and Knowledge-Based Systems , volume=

    Achieving k-anonymity privacy protection using generalization and suppression , author=. International Journal of Uncertainty, Fuzziness and Knowledge-Based Systems , volume=. 2002 , publisher=

    2002

  21. [21]

    ACM Transactions on Knowledge Discovery from Data , volume=

    l -diversity: Privacy beyond k -anonymity , author=. ACM Transactions on Knowledge Discovery from Data , volume=. 2007 , publisher=

    2007

  22. [22]

    PODS , volume=

    Generalizing data to provide anonymity when disclosing information , author=. PODS , volume=

  23. [23]

    1998 , publisher=

    Protecting privacy when disclosing information: k-anonymity and its enforcement through generalization and suppression , author=. 1998 , publisher=

    1998

  24. [24]

    arXiv preprint cs/0610105 , year=

    How to break anonymity of the netflix prize dataset , author=. arXiv preprint cs/0610105 , year=

    Pith/arXiv arXiv

  25. [25]

    ACM Computing Surveys , volume=

    Privacy-preserving data publishing: A survey of recent developments , author=. ACM Computing Surveys , volume=. 2010 , publisher=

    2010

  26. [26]

    25th \ USENIX \ Security Symposium ( \ USENIX \ Security 16) , pages=

    Oblivious multi-party machine learning on trusted processors , author=. 25th \ USENIX \ Security Symposium ( \ USENIX \ Security 16) , pages=

  27. [27]

    Artificial Intelligence and Statistics , pages=

    Communication-efficient learning of deep networks from decentralized data , author=. Artificial Intelligence and Statistics , pages=

  28. [28]

    IEEE Transactions on Knowledge and Data Engineering , volume=

    Privacy-preserving distributed mining of association rules on horizontally partitioned data , author=. IEEE Transactions on Knowledge and Data Engineering , volume=. 2004 , publisher=

    2004

  29. [29]

    Advances in Neural Information Processing Systems , volume=

    A Scalable Approach for Privacy-Preserving Collaborative Machine Learning , author=. Advances in Neural Information Processing Systems , volume=

  30. [30]

    Privacy-Preserving Data Mining , pages=

    A general survey of privacy-preserving data mining models and algorithms , author=. Privacy-Preserving Data Mining , pages=. 2008 , publisher=

    2008

  31. [31]

    Proceedings of the 2016 international conference on management of data , pages=

    Publishing attributed social graphs with formal privacy guarantees , author=. Proceedings of the 2016 international conference on management of data , pages=

    2016

  32. [32]

    The VLDB Journal , volume=

    Correlated network data publication via differential privacy , author=. The VLDB Journal , volume=. 2014 , publisher=

    2014

  33. [33]

    Proceedings of the 20th ACM SIGKDD international conference on Knowledge discovery and data mining , pages=

    Differentially private network data release via structural inference , author=. Proceedings of the 20th ACM SIGKDD international conference on Knowledge discovery and data mining , pages=

  34. [34]

    Proceedings of the 20th ACM SIGKDD international conference on Knowledge discovery and data mining , pages=

    Exponential random graph estimation under differential privacy , author=. Proceedings of the 20th ACM SIGKDD international conference on Knowledge discovery and data mining , pages=

  35. [35]

    Proceedings of the 2011 ACM SIGCOMM conference on Internet measurement conference , pages=

    Sharing graphs using differentially private graph models , author=. Proceedings of the 2011 ACM SIGCOMM conference on Internet measurement conference , pages=

    2011

  36. [36]

    2009 IEEE International Conference on Data Mining Workshops , pages=

    A differentially private graph estimator , author=. 2009 IEEE International Conference on Data Mining Workshops , pages=. 2009 , organization=

    2009

  37. [37]

    Theory of cryptography conference , pages=

    Calibrating noise to sensitivity in private data analysis , author=. Theory of cryptography conference , pages=. 2006 , organization=

    2006

  38. [38]

    2010 ieee symposium on security and privacy , pages=

    A practical attack to de-anonymize social network users , author=. 2010 ieee symposium on security and privacy , pages=. 2010 , organization=

    2010

  39. [39]

    2009 30th IEEE symposium on security and privacy , pages=

    De-anonymizing social networks , author=. 2009 30th IEEE symposium on security and privacy , pages=. 2009 , organization=

    2009

  40. [40]

    24th \ USENIX \ Security Symposium ( \ USENIX \ Security 15) , pages=

    Secgraph: A uniform and open-source evaluation system for graph data anonymization and de-anonymization , author=. 24th \ USENIX \ Security Symposium ( \ USENIX \ Security 15) , pages=

  41. [41]

    Proceedings of the 20th Annual Network and Distributed System Security Symposium , pages=

    Preserving link privacy in social network based systems , author=. Proceedings of the 20th Annual Network and Distributed System Security Symposium , pages=

  42. [42]

    Proceedings of the VLDB Endowment , volume=

    Personalized privacy protection in social networks , author=. Proceedings of the VLDB Endowment , volume=. 2010 , publisher=

    2010

  43. [43]

    Proceedings of the 2010 ACM SIGMOD International Conference on Management of data , pages=

    K-isomorphism: privacy preserving network publication against structural attacks , author=. Proceedings of the 2010 ACM SIGMOD International Conference on Management of data , pages=

    2010

  44. [44]

    Proceedings of the VLDB Endowment , volume=

    K-automorphism: A general framework for privacy preserving network publication , author=. Proceedings of the VLDB Endowment , volume=. 2009 , publisher=

    2009

  45. [45]

    IEEE Transactions on Network Science and Engineering , year=

    Publishing social network graph Eigen-Spectrum with privacy guarantees , author=. IEEE Transactions on Network Science and Engineering , year=

  46. [46]

    Advances in neural information processing systems , volume=

    Near-optimal differentially private principal components , author=. Advances in neural information processing systems , volume=

  47. [47]

    Proceedings of the forty-fifth annual ACM symposium on Theory of computing , pages=

    Beyond worst-case analysis in private singular vector computation , author=. Proceedings of the forty-fifth annual ACM symposium on Theory of computing , pages=

  48. [48]

    Theory of Cryptography Conference , pages=

    Analyzing graphs with node differential privacy , author=. Theory of Cryptography Conference , pages=. 2013 , organization=

    2013

  49. [49]

    Proceedings of the 2015 ACM SIGMOD international conference on management of data , pages=

    Private release of graph statistics using ladder functions , author=. Proceedings of the 2015 ACM SIGMOD international conference on management of data , pages=

    2015

  50. [50]

    Proceedings of the VLDB Endowment , volume=

    Private analysis of graph structure , author=. Proceedings of the VLDB Endowment , volume=. 2011 , publisher=

    2011

  51. [51]

    Advances in Neural Information Processing Systems , pages=

    Efficiently estimating erdos-renyi graphs with node differential privacy , author=. Advances in Neural Information Processing Systems , pages=

  52. [52]

    Proceedings of the 2016 International Conference on Management of Data , pages=

    Publishing graph degree distribution with node differential privacy , author=. Proceedings of the 2016 International Conference on Management of Data , pages=

    2016

  53. [53]

    Proceedings of the VLDB Endowment , volume=

    Boosting the Accuracy of Differentially Private Histograms Through Consistency , author=. Proceedings of the VLDB Endowment , volume=

  54. [54]

    2009 Ninth IEEE International Conference on Data Mining , pages=

    Accurate estimation of the degree distribution of private networks , author=. 2009 Ninth IEEE International Conference on Data Mining , pages=. 2009 , organization=

    2009

  55. [55]

    Annual International Conference on the Theory and Applications of Cryptographic Techniques , pages=

    Efficient binary conversion for Paillier encrypted values , author=. Annual International Conference on the Theory and Applications of Cryptographic Techniques , pages=. 2006 , organization=

    2006

  56. [56]

    Multimedia Signal Processing Group, Delft University of Technology, The Netherlands, and TNO Information and Communication Technology, Delft, The Netherlands, Tech

    Comparing encrypted data , author=. Multimedia Signal Processing Group, Delft University of Technology, The Netherlands, and TNO Information and Communication Technology, Delft, The Netherlands, Tech. Rep , year=

  57. [57]

    Proceedings of the 2004 SIAM International Conference on Data Mining , pages=

    Privacy-preserving multivariate statistical analysis: Linear regression and classification , author=. Proceedings of the 2004 SIAM International Conference on Data Mining , pages=

    2004

  58. [58]

    International Journal of Information Security , volume=

    A generalization of Paillier?s public-key system with applications to electronic voting , author=. International Journal of Information Security , volume=. 2010 , publisher=

    2010

  59. [59]

    Proceedings of the Sixth ACM Conference on Data and Application Security and Privacy , pages=

    Scalable and secure logistic regression via homomorphic encryption , author=. Proceedings of the Sixth ACM Conference on Data and Application Security and Privacy , pages=

  60. [60]

    Ryan, Peter YA , journal=. Pr. 2008 , publisher=

    2008

  61. [61]

    International Conference on the Theory and Applications of Cryptographic Techniques , pages=

    The bit security of Paillier?s encryption scheme and its applications , author=. International Conference on the Theory and Applications of Cryptographic Techniques , pages=. 2001 , organization=

    2001

  62. [62]

    Proceedings of the 2007 ACM SIGMOD international conference on Management of data , pages=

    Privacy preserving schema and data matching , author=. Proceedings of the 2007 ACM SIGMOD international conference on Management of data , pages=

    2007

  63. [63]

    International Conference on Intelligence and Security Informatics , pages=

    Privacy-preserving inter-database operations , author=. International Conference on Intelligence and Security Informatics , pages=. 2004 , organization=

    2004

  64. [64]

    2009 , publisher=

    Foundations of cryptography: volume 2, basic applications , author=. 2009 , publisher=

    2009

  65. [65]

    Journal of computer and system sciences , volume=

    Probabilistic encryption , author=. Journal of computer and system sciences , volume=. 1984 , publisher=

    1984

  66. [66]

    International Conference on the Theory and Applications of Cryptographic Techniques , pages=

    Public-key cryptosystems based on composite degree residuosity classes , author=. International Conference on the Theory and Applications of Cryptographic Techniques , pages=. 1999 , organization=

    1999

  67. [67]

    , author=

    Practical Federated Gradient Boosting Decision Trees. , author=. AAAI , pages=

  68. [68]

    Advances in Neural Information Processing Systems , pages=

    Federated multi-task learning , author=. Advances in Neural Information Processing Systems , pages=

  69. [69]

    Proceedings of the 2017 ACM SIGSAC Conference on Computer and Communications Security , pages=

    Practical secure aggregation for privacy-preserving machine learning , author=. Proceedings of the 2017 ACM SIGSAC Conference on Computer and Communications Security , pages=

    2017

  70. [70]

    arXiv preprint arXiv:1610.05492 , year=

    Federated learning: Strategies for improving communication efficiency , author=. arXiv preprint arXiv:1610.05492 , year=

    Pith/arXiv arXiv

  71. [71]

    Synthesis Lectures on Artificial Intelligence and Machine Learning , volume=

    Federated learning , author=. Synthesis Lectures on Artificial Intelligence and Machine Learning , volume=. 2019 , publisher=

    2019

  72. [72]

    ACM Transactions on Intelligent Systems and Technology (TIST) , volume=

    Federated machine learning: Concept and applications , author=. ACM Transactions on Intelligent Systems and Technology (TIST) , volume=. 2019 , publisher=

    2019

  73. [73]

    arXiv preprint arXiv:2007.06081 , year=

    VAFL: a Method of Vertical Asynchronous Federated Learning , author=. arXiv preprint arXiv:2007.06081 , year=

    arXiv 2007

  74. [74]

    arXiv preprint arXiv:2008.10838 , year=

    FedMVT: Semi-supervised Vertical Federated Learning with MultiView Training , author=. arXiv preprint arXiv:2008.10838 , year=

    arXiv 2008

  75. [75]

    IEEE Intelligent Systems , year=

    A Secure Federated Transfer Learning Framework , author=. IEEE Intelligent Systems , year=

  76. [76]

    arXiv preprint arXiv:2001.11154 , year=

    Multi-Participant Multi-Class Vertical Federated Learning , author=. arXiv preprint arXiv:2001.11154 , year=

    arXiv 2001

  77. [77]

    Proceedings of the VLDB Endowment , volume=

    Privacy Preserving Vertical Federated Learning for Tree-Based Models , author=. Proceedings of the VLDB Endowment , volume=. 2020 , publisher=

    2020

  78. [78]

    IEEE Intelligent Systems , volume=

    Secureboost: A lossless federated learning framework , author=. IEEE Intelligent Systems , volume=. 2021 , publisher=

    2021

  79. [79]

    arXiv preprint arXiv:1911.09824 , year=

    Parallel distributed logistic regression for vertical federated learning without third-party coordinator , author=. arXiv preprint arXiv:1911.09824 , year=

    arXiv 1911

  80. [80]

    arXiv preprint arXiv:1912.00513 , year=

    A quasi-newton method based vertical federated learning framework for logistic regression , author=. arXiv preprint arXiv:1912.00513 , year=

    arXiv 1912

Showing first 80 references.