Towards Responsible AI for Financial Transactions

Maree, Charl; Modal, Jan Erik; Omlin, Christian W.

doi:10.1109/ssci47803.2020.9308456

Cited by 23 publications

(14 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The technique also enables AI developers to address typical challenges in automated systems, such as reducing social bias reinforcement, keeping people's jobs and talents, resolving responsibility to ensure confidence in an algorithm's results, and more [27]. Commercial AI systems in radiation clinics have just lately been developed, in contrast to the aerospace sector, with efforts concentrated on showing performance in academic or clinical settings, as well as product approval [28,29]. Until previously, commercial AI systems for radiation were only available as static goods, allowing cancer specialists to analyze their effectiveness.…”

Section: Literature Reviewmentioning

confidence: 99%

Enhancing Artificial Intelligence Control Mechanisms: Current Practices, Real Life Applications and Future Views

Usmani

Happonen

Watada

2022

Lecture Notes in Networks and Systems

View full text Add to dashboard Cite

The popularity of Artificial Intelligence has grown lately with the potential it promises for revolutionizing a wide range of different sectors. To achieve the change, whole community must overcome the Machine Learning (ML) related explainability barrier, an inherent obstacle of current sub symbolism-based approaches, e.g. in Deep Neural Networks, which was not existing during the last AI hype time including some expert and rule-based systems. Due to lack of transparency, privacy, biased systems, lack of governance and accountability, our society demands toolsets to create responsible AI solutions for enabling of unbiased AI systems. These solutions will help business owners to create AI applications which are trust enhancing, open and transparent and also explainable. Properly made systems will enhance trust among employees, business leaders, customers and other stakeholders. The process of overseeing artificial intelligence usage and its influence on related stakeholders belongs to the context of AI Governance. Our work gives a detailed overview of a governance model for Responsible AI, emphasizing fairness, model explainability, and responsibility in large-scale AI technology deployment in real-world organizations. Our goal is to provide the model developers in an organization to understand the Responsible AI with a comprehensive governance framework that outlines the details of the different roles and the key responsibilities. The results work as reference for future research is aimed to encourage area experts from other disciplines towards embracement of AI in their own business sectors, without interpretability shortcoming biases.

show abstract

Section: Literature Reviewmentioning

confidence: 99%

Enhancing Artificial Intelligence Control Mechanisms: Current Practices, Real Life Applications and Future Views

Usmani

Happonen

Watada

2022

Lecture Notes in Networks and Systems

View full text Add to dashboard Cite

show abstract

“…The lack of explainability and interpretability has thus far hindered the wider adoption of machine learning, mainly due to model opacity; model understanding is essential in financial services [12][13][14]. We distinguish between explainability and interpretability: explainability refers to a symbolic representation of the knowledge a model has learned, while interpretability is necessary for reasoning about a model's predictions.…”

Section: Introductionmentioning

confidence: 99%

Can Interpretable Reinforcement Learning Manage Prosperity Your Way?

Maree

Omlin

2022

Self Cite

View full text Add to dashboard Cite

Personalisation of products and services is fast becoming the driver of success in banking and commerce. Machine learning holds the promise of gaining a deeper understanding of and tailoring to customers’ needs and preferences. Whereas traditional solutions to financial decision problems frequently rely on model assumptions, reinforcement learning is able to exploit large amounts of data to improve customer modelling and decision-making in complex financial environments with fewer assumptions. Model explainability and interpretability present challenges from a regulatory perspective which demands transparency for acceptance; they also offer the opportunity for improved insight into and understanding of customers. Post-hoc approaches are typically used for explaining pretrained reinforcement learning models. Based on our previous modeling of customer spending behaviour, we adapt our recent reinforcement learning algorithm that intrinsically characterizes desirable behaviours and we transition to the problem of prosperity management. We train inherently interpretable reinforcement learning agents to give investment advice that is aligned with prototype financial personality traits which are combined to make a final recommendation. We observe that the trained agents’ advice adheres to their intended characteristics, they learn the value of compound growth, and, without any explicit reference, the notion of risk as well as improved policy convergence.

show abstract

“…These applications generally lack the personalisation needed to enhance customer relations and support service delivery for growing customer bases. The lack of explainability and interpretability has thus far hindered the wider adoption of machine learning, mainly due to model opacity; model understanding is essential in financial services [12,13,14]. We distinguish between explainabil-ity and interpretability: explainability refers to a symbolic representation of the knowledge a model has learned, while interpretability is necessary for reasoning about a model's predictions.…”

Section: Introductionmentioning

confidence: 99%

Can Interpretable Reinforcement Learning Manage Prosperity Your Way?

Maree,

Omlin

2022

Preprint

Self Cite

View full text Add to dashboard Cite

Personalisation of products and services is fast becoming the driver of success in banking and commerce. Machine learning holds the promise of gaining a deeper understanding of and tailoring to customers' needs and preferences. Whereas traditional solutions to financial decision problems frequently rely on model assumptions, reinforcement learning is able to exploit large amounts of data to improve customer modelling and decisionmaking in complex financial environments with fewer assumptions. Model explainability and interpretability present challenges from a regulatory perspective which demands transparency for acceptance; they also offer the opportunity for improved insight into and understanding of customers. Post-hoc approaches are typically used for explaining pretrained reinforcement learning models. Based on our previous modeling of customer spending behaviour, we adapt our recent reinforcement learning algorithm that intrinsically characterizes desirable behaviours and we transition to the problem of asset management. We train inherently interpretable reinforcement learning agents to give investment advice that is aligned with prototype financial personality traits which are combined to make a final recommendation. We observe that the trained agents' advice adheres to their intended characteristics, they learn the value of compound growth, and, without any explicit reference, the notion of risk as well as improved policy convergence.

show abstract

Towards Responsible AI for Financial Transactions

Cited by 23 publications

References 11 publications

Enhancing Artificial Intelligence Control Mechanisms: Current Practices, Real Life Applications and Future Views

Enhancing Artificial Intelligence Control Mechanisms: Current Practices, Real Life Applications and Future Views

Can Interpretable Reinforcement Learning Manage Prosperity Your Way?

Can Interpretable Reinforcement Learning Manage Prosperity Your Way?

Contact Info

Product

Resources

About