Learning Transactions Representations for Information Management in Banks: Mastering Local, Global, and External Knowledge

2024-04-02

AI Summary

To address challenges in banking transaction data—including difficulty in multi-task coordination (e.g., customer satisfaction enhancement and repayment capacity prediction), high model redundancy, and weak generalization of local modeling—this paper proposes a universal transaction representation model. The model overcomes the limitation of isolated single-customer modeling via a representation enhancement mechanism that integrates external knowledge across customers. It unifies optimization of both local tasks (e.g., transaction prediction) and global tasks (e.g., repayment prediction) through a synergistic framework combining contrastive self-supervised learning, generative modeling, and unsupervised cross-customer knowledge injection. Evaluated on 11 downstream tasks, the model consistently outperforms classical baselines, achieving up to a 20% absolute accuracy improvement. Results demonstrate its strong generalizability, task-agnostic applicability, and deployment efficiency—enabling scalable, lightweight integration into production banking systems.

Technology Category

• Machine Learning: Representation Learning
• Search and Optimization: Learning to Search
• Natural Language Processing: Generation

Application Category

• User Modeling, Personalization and Recommendation: Accountability, Transparency, and Ethics for personalization
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

In today's world, banks use artificial intelligence to optimize diverse business processes, aiming to improve customer experience. Most of the customer-related tasks can be categorized into two groups: 1) local ones, which focus on a client's current state, such as transaction forecasting, and 2) global ones, which consider the general customer behaviour, e.g., predicting successful loan repayment. Unfortunately, maintaining separate models for each task is costly. Therefore, to better facilitate information management, we compared eight state-of-the-art unsupervised methods on 11 tasks in search for a one-size-fits-all solution. Contrastive self-supervised learning methods were demonstrated to excel at global problems, while generative techniques were superior at local tasks. We also introduced a novel approach, which enriches the client's representation by incorporating external information gathered from other clients. Our method outperforms classical models, boosting accuracy by up to 20%.

Problem

• Banking Transaction Optimization
• Customer Satisfaction
• Predictive Analytics

Innovation

• Integrated Customer Information
• Contrastive Learning
• Generative Techniques