PulseExploreJournal ClubDebatesTrendingResearchersJournals
Instagram
HomeExploreJournal ClubTrending
Synapse
⌘+K
Synapse
March 19, 2026Mathematics1 citationsOpen Access

FedQuAD: Fast-Converging Curvature-Aware Federated Learning for Credit Default Prediction from Private Accounting Data

View Full Paper
DBDingwen BaiMWMuGa WaErQWQian Wu

Key Points

  • To enhance credit default prediction while managing sensitive firm-level accounting data using federated learning.
  • Proposed a novel FedQuAD algorithm combining curvature aggregation and lightweight updates.
  • Utilized a proximal variance-reduced local solver to address client drift under non-iid distributions.
  • Applied federated robust standardization of financial ratios using secure quantile statistics.
  • FedQuAD achieved target AUC and calibration error with fewer communication rounds than other methods.
  • Maintained privacy constraints consistent with secure aggregation and differential privacy.

Abstract

Credit default prediction from firm-level accounting statements is central to risk management, yet the underlying financial data are highly sensitive and often siloed across banks, auditors, and platforms. Federated learning (FL) offers a practical route to collaborative modeling without centralizing raw records, but standard FL optimization can converge slowly under severe client heterogeneity, heavy-tailed accounting features, and label imbalance typical of default events. This paper proposes FedQuAD, a novel fast-converging FL algorithm that couples (i) quasi-Newton curvature aggregation on the server with a lightweight limited-memory update to accelerate global progress, (ii) a proximal variance-reduced local solver that stabilizes client drift under non-IID accounting distributions, and (iii) federated robust standardization of tabular financial ratios via secure aggregated quantile statistics to mitigate scale instability and outliers. FedQuAD is communication-efficient by design: It transmits compact gradient and curvature sketches and adapts local computation to each client’s stochasticity and drift. We provide convergence guarantees for strongly convex default-risk objectives (logistic and calibrated GLM losses) under bounded heterogeneity, and extend the analysis to nonconvex deep tabular models via expected stationarity bounds. Experiments on public credit-risk benchmarks with simulated cross-silo (institutional) partitions demonstrate that FedQuAD reaches target AUC and calibration error with substantially fewer communication rounds than representative baselines while maintaining privacy constraints compatible with secure aggregation and optional client-level differential privacy accounting.

Ask AI
Helpful
Bookmark
Share
View Full Paper

Cite This Study

Bai et al. (2026) studied this question.

synapsesocial.com/papers/69bb92ae496e729e62980246https://doi.org/10.3390/math14061012
Ask AI
Helpful
Bookmark
Share
View Full Paper