The Flow has not summarised this story yet — read it at arXiv Machine Learning.
The paper presents a deep learning credit risk early warning system that fuses heterogeneous data sources, such as transaction behaviors and social networks, using deep neural networks and attention mechanisms. By extracting multidimensional features, the system establishes an early identification mechanism for corporate and individual credit risks. Testing shows that this approach improves the accuracy and timeliness of risk warnings compared to traditional rule‑based engines.
The study investigates whether Large Language Models (LLMs) can translate technical explanations from credit risk models into stakeholder-friendly narratives. Using Freddie Mac loan data, the authors compare standard tabular models (XGBoost + SHAP) with alternative data pipelines (GNN + GNNExplainer and a bimodal mix) and generate explanations with three LLM configurations: a small fine‑tuned Gemma 3 4B, a large fine‑tuned DeepSeek R1 70B, and a zero‑shot Gemini 2.5. Findings show that the quality of explanations is more dependent on the evidence representation than on the LLM, that narratives reliably identify influential factors but are less consistent about the direction of influence, and that credit professionals demand higher evidentiary standards than non‑professionals.
arXiv:2607. 14416v1 Announce Type: new Abstract: The interconnected nature of global financial systems makes them vulnerable to systemic risks, where the failure of a few institutions can trigger catastrophic cascading defaults.
The paper introduces DTD‑VAE, a Variational Autoencoder that disentangles temporal dependencies to better predict credit risk. It uses an autoregressive feature inference module to capture temporal patterns among latent variables and an element‑wise gating mechanism in the generative module to assign independent weights to each latent dimension, especially those relevant to credit risk. Experiments on six real‑world datasets show the model outperforms existing methods, improving ROC‑AUC by 3.2%–4.86% and Accuracy Ratio by 6.41%–9.71%.
arXiv:2609.14234v1 Announce Type: cross Abstract: Financial fraud in corporate transaction networks has grown more coordinated and harder to detect with rule-based engines and with classical learning...
Credit risk detection, particularly mitigating individual fraud, is crucial for maintaining the stability of digital financial ecosystems. Accurately identifying credit fraud among billions of users is critical for minimizing financial losses and safeguarding the sustainability of inclusive financial services.