John Hancock Life Insurance Company, U.S.A. is a Boston-based insurance company. Established April 21, 1862, it was named in honor of John Hancock, a prominent patriot.In 2004, John Hancock was acquired by the Canadian life insurance company Manulife Financial. The company and the majority of Manulife's U.S. assets continue to operate under the John Hancock name.
The paper proposes a hybrid AI framework that combines temporal and graph measures to measure financial risk in dynamic nonstationary markets. The structure has a temporal transformer encoder, a relation graph neural network (GNN) and multi-task probabilistic prediction heads to jointly score the probability of default (PD), value at risk (VaR), conditional value at risk (CVaR), and expected losses. The system uses many kinds of information. This involves market signals, economic information, firm information, randomly generated news-based features, and specific exposure networks. Our preprocessing pipeline aligns different time series at various resolutions. We use four concept-drift handling mechanisms namely online adaptation, divergence detection, ensembles and stress simulation for augmenting. This bolsters strength as market circumstances shift. The proposed model surpasses statistical baselines including CR, deep-learning baselines such as LSTM and GNN baselines like PGNN on three datasets (1,500 global firms over crisis regimes). The architecture improves the performance of traditional models by enhancing the PD AUC by 12.6 % as well as reducing the forecast errors of VaR and CVaR by 28-50 % and generating large expected-loss improvements at the portfolio level. There are various methods for explaining GNN outputs including SHAP feature attributions, GNN edge-level interpretability, and rule-based surrogate governance models that satisfy auditability requirements. According to the results, the novel temporalgraph multi-tasking system for systematic financial risk assessment is more flexible, interpretable and accurate approaches real-world volatility and systemic interdependence in comparison to existing methods.
As electronic banking rapidly expands, the demand for effective ATM card fraud prevention solutions to safeguard financial transactions increases correspondingly. As ATM usage increases, fraudsters perceive these systems as vulnerable targets, resulting in risks such as card cloning, PIN breaches, and unauthorised access. It is essential to ensure the security of electronic transactions to protect both users and financial institutions. The research presents an advanced approach for mitigating ATM card fraud by employing an AE-ProbRF model. Initially, data preprocessing techniques are employed to address missing values, verify data types, and examine unique characteristics inside the ATM transaction dataset. Utilising PSO, significant characteristics are selected to enhance model performance and simplify computations. The autoencoder generates low-dimensional representations from high-dimensional transaction data, facilitating feature extraction. The gathered features are further classified using a probabilistic random forest, an ensemble learning method that integrates probabilistic classification with bootstrap aggregating to distinguish between authentic and fraudulent transactions. The proposed AE-ProbRF model exhibits a detection accuracy of 96.34%, surpassing many alternative approaches. The findings indicate that the integration of feature optimisation, dimensionality reduction, and ensemble learning significantly enhances the reliability and utility of ATM card fraud prevention systems.
Blockchain-based systems have become widely used in secure and decentralized operation of transactions, but conventional consensus systems have high latency, low scalability, and are subject to complex adversarial actions. To overcome these drawbacks, this paper suggests a blockchain consensus mechanism that is enhanced with AI and incorporates intelligence, which is managed by machine learning, in the process of validating transactions. The framework has been proposed with use of dynamic node trust evaluation, intelligent selection of the validator, anomaly detection and tuning of the consensus parameter to enhance security and performance. It is experimentally proved that the suggested approach has a high transaction validation rate, 97.8%, in comparison with the conventional consensus methods, which have a 91.4% rate. The latency of the end-to-end consensus is minimized (820 ms) to 430 ms and the transaction throughput is increased (420 TPS) to 760 TPS. Moreover, the security reliability is achieved with a high level, the False Acceptance Rate is lowered to 1.5 instead of 4.6 and the False Rejection Rate is also lowered to 2.1 instead of 6.2. These findings validate the claim that blockchain consensus with artificial intelligence can be used to create secure, scalable, and adaptive validation of transactions to support next-generation decentralized applications.
