How are reinforcement learning and deep learning algorithms used for big data based decision making in financial industries–A review and research agenda

Can Artificial Intelligence (AI) Manage Behavioural Biases Among Financial Planners?

“…This study proposes that deep reinforcement learning can overcome hindsight bias using AI. Matsuo et al (2022) and Singh et al (2022) state that deep learning and reinforcement learning are significant factors in the AI system. In addressing the hindsight bias, the system needs to separate individuals' existing perceptions or understanding of previous information, as they do not influence the new set of information.…”

Deep Learning in Business Analytics: a Clash of Expectations and Reality

“…Several studies have shown that deep learning architectures such as recurrent neural networks (RNN) and long short-term memory (LSTM) are strong candidates for time series data in finance and offer superior performance ( Fischer & Krauss, 2018 ). ( 5) Other areas for investigation would be reinforcement learning applications within business analytics ( Singh et al, 2022 ), bio-inspired computation/ML models ( Jain, Batra, Kar, Agrawal, & Tikkiwal, 2022 ; Kudithipudi et al, 2022 ), and also research that would further enhance the explainability of AI/ML, which would enable additional use cases in regulatory environments that require transparency ( Bücker, Szepannek, Gosiewska, & Biecek, 2022 ;Sharma, Kumar, & Chuah, 2021 ).…”

“…Artificial intelligence (AI) and machine learning (ML) have been widely accepted as general-purpose technology for decision-making ( Agrawal, Gans, & Goldfarb, 2019 ) across a variety of domains, industries, and functions including biotech, healthcare ( Sounderajah et al, 2022 ;Young & Steele, 2022 ), marketing ( Verma et al, 2021 ), human resource management ( Votto, Valecha, Najafirad, & Rao, 2021 ), financial services ( Schmitt, 2020 ;Singh, Chen, Singhania, Nanavati, & Gupta, 2022 ), insurance ( Rawat, Rawat, Kumar, & Sabitha, 2021 ), risk management ( Fujii, Sakaji, Masuyama, & Sasaki, 2022 ;Schmitt, 2022b ), cybersecurity ( Taddeo, McCutcheon, & Floridi, 2019 ;Thorat, Parekh, & Mangrulkar, 2021 ), and many others ( Kumar, Kar, & Ilavarasan, 2021 ). E-mail address: marcschmitt@hotmail.de Data analytics and information systems ( Kushwaha et al, 2021 ) build the foundation for actionable insights and are the primary enablers for AI-based decision-making across all domains.…”

When to choose ranked area integrals versus integrated gradient for explainable artificial intelligence – a comparison of algorithms

“…Figure 1 depicts a high-level structure of the current state of the art in XAI. On the one hand, emphasis has been placed here on the sequence of current research activities often performed by several scholars, their dependencies and order (Adadi and Berrada, 2018; Rudin, 2019; Singh et al. , 2022a).…”

“…Figure 1 depicts a high-level structure of the current state of the art in XAI. On the one hand, emphasis has been placed here on the sequence of current research activities often performed by several scholars, their dependencies and order (Adadi and Berrada, 2018;Rudin, 2019;Singh et al, 2022a). This sequence usually starts from input data that is then used for modeling purposes, employing connectionist data-driven learning or symbolic reasoning knowledge-based paradigms (Singh et al, 2022b) After a model has been formed, the XAI method is applied for its analysis and knowledge discovery, supporting its interpretability.…”