{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,7]],"date-time":"2026-10-07T14:50:50Z","timestamp":1791384650715,"version":"4.3.3"},"posted":{"date-parts":[[2026]]},"group-title":"SSRN","reference-count":39,"publisher":"Elsevier BV","license":[{"start":{"date-parts":[[2026,1,1]],"date-time":"2026-01-01T00:00:00Z","timestamp":1767225600000},"content-version":"unspecified","delay-in-days":0,"URL":"https:\/\/www.uspto.gov\/ip-policy\/copyright-policy\/copyright-basics"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"abstract":"<jats:p>The joint modeling of market regimes, asset valuation, systemic risk, and portfolio optimization remains a major challenge in computational finance. Existing approaches typically address these components separately, limiting their adaptability to non-stationary markets and weakening the integration of risk assessment with institutional decision-making. Passive strategies such as the Equal Weight ( \u00a0) benchmark efficiently capture equity risk premia during stable market conditions but fail to account for compounding tail risks, while conventional optimization methods are highly sensitive to parameter estimation and often perform poorly out of sample. To overcome these limitations, this paper proposes the Multi-Agent Intelligent Financial System (MAIFS), a unified artificial intelligence governance framework that integrates six specialized agents: Market Understanding, Valuation, Risk Analysis, Decision, Learning, and Explainability. The Market Understanding Agent identifies latent market regimes using a Gaussian Hidden Markov Model enhanced by a Riemannian geometry-based instability indicator. The Valuation Agent combines multiple pricing models through a constrained ensemble with calibrated uncertainty estimates. The Risk Analysis Agent integrates dynamic covariance modeling, DebtRank-based contagion analysis, and tail risk estimation. The Decision Agent formulates portfolio allocation as a risk-constrained stochastic control problem solved through reinforcement learning. The Learning Agent supports adaptive online updates with sublinear regret, while the Explainability Agent provides SHAP-based explainable artificial intelligence (XAI) to enhance transparency and regulatory compliance. Experiments on global equity markets from 2005 to 2024 show that MAIFS achieves out-of-sample returns comparable to passive benchmarks while providing superior market regime detection, systemic risk mitigation, and regulatory compliance.<\/jats:p>","DOI":"10.2139\/ssrn.7576742","type":"posted-content","created":{"date-parts":[[2026,10,7]],"date-time":"2026-10-07T13:54:13Z","timestamp":1791381253000},"source":"Crossref","is-referenced-by-count":0,"title":["A Closed-Loop Multi-Agent Framework for Regime-Aware Portfolio Optimization, Systemic Risk Monitoring, and Regulatory Compliance in Global Capital Markets"],"prefix":"10.2139","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-8642-5688","authenticated-orcid":true,"given":"ABDUL  KADAR","family":"MUHAMMAD MASUM","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-3463-2738","authenticated-orcid":true,"given":"Md","family":"Azad","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0005-4704-1269","authenticated-orcid":true,"given":"CHANDA  RANI","family":"DEBI","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-3151-9855","authenticated-orcid":true,"given":"Paolo","family":"Saona","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"78","reference":[{"key":"ref1","first-page":"1","article-title":"Explainable algorithmic trading with embedded audit and tax trails: a technology-enabled behavioral oversight model","author":"Y Ibrahim","year":"2026","journal-title":"Int. J. Account. Inf. Manage"},{"key":"ref2","author":"S M Bartram","year":"2020","journal-title":"Artificial intelligence in asset management"},{"issue":"7","key":"ref3","doi-asserted-by":"crossref","first-page":"1947","DOI":"10.1086\/714090","article-title":"Asset pricing with omitted factors","volume":"129","author":"S Giglio","year":"2021","journal-title":"J. Polit. Econ"},{"issue":"5","key":"ref4","doi-asserted-by":"crossref","first-page":"2223","DOI":"10.1093\/rfs\/hhaa009","article-title":"Empirical asset pricing via machine learning","volume":"33","author":"S Gu","year":"2020","journal-title":"Rev. Financ. Stud"},{"issue":"4","key":"ref5","doi-asserted-by":"crossref","first-page":"1481","DOI":"10.1111\/j.1540-6261.2004.00670.x","article-title":"Risks for the long run: A potential resolution of asset pricing puzzles","volume":"59","author":"R Bansal","year":"2004","journal-title":"J. Finance"},{"issue":"2","key":"ref6","doi-asserted-by":"crossref","first-page":"557","DOI":"10.1111\/mafi.12388","article-title":"Reinforcement learning with dynamic convex risk measures","volume":"34","author":"A Coache","year":"2024","journal-title":"Math. Finance"},{"issue":"9","key":"ref7","doi-asserted-by":"crossref","DOI":"10.3390\/jrfm18090497","article-title":"Deep hedging under market frictions: A comparison of DRL models for options hedging with impact and transaction costs","volume":"18","author":"E Huang","year":"2025","journal-title":"J. Risk Financial Manag"},{"issue":"1","key":"ref8","doi-asserted-by":"crossref","first-page":"56","DOI":"10.1038\/s42256-019-0138-9","article-title":"From local explanations to global understanding with explainable AI for trees","volume":"2","author":"S M Lundberg","year":"2020","journal-title":"Nat. Mach. Intell"},{"issue":"4","key":"ref9","doi-asserted-by":"crossref","first-page":"494","DOI":"10.1561\/2200000101","article-title":"Conformal prediction: A gentle introduction","volume":"16","author":"A N Angelopoulos","year":"2023","journal-title":"Found. Trends Mach. Learn"},{"issue":"6275","key":"ref10","doi-asserted-by":"crossref","first-page":"818","DOI":"10.1126\/science.aad0299","article-title":"Complexity theory and financial regulation","volume":"351","author":"S Battiston","year":"2016","journal-title":"Science"},{"key":"ref11","author":"J Aarnio","year":"2025","journal-title":"AI-Driven Portfolio Management: A Comparative Research of Deep Reinforcement Learning Techniques Against The 1\/N Portfolio Strategy"},{"key":"ref12","doi-asserted-by":"crossref","DOI":"10.1016\/j.eswa.2023.121711","article-title":"A deep Q-learning based algorithmic trading system for commodity futures markets","volume":"237","author":"M Massahi","year":"2024","journal-title":"Expert Syst. Appl"},{"issue":"2","key":"ref13","doi-asserted-by":"crossref","first-page":"714","DOI":"10.1287\/mnsc.2023.4695","article-title":"Deep learning in asset pricing","volume":"70","author":"L Chen","year":"2024","journal-title":"Manag. Sci"},{"issue":"4","key":"ref14","doi-asserted-by":"crossref","first-page":"969","DOI":"10.3390\/forecast4040053","article-title":"The Lasso and the Factor Zoo-Predicting Expected Returns in the CrossSection","volume":"4","author":"M Messmer","year":"2022","journal-title":"Forecasting"},{"key":"ref15","article-title":"Reinforcement learning in financial markets-a survey","author":"T G Fischer","year":"2018","journal-title":"FAU Discussion Papers in Economics"},{"key":"ref16","doi-asserted-by":"crossref","first-page":"1835","DOI":"10.52202\/068431-0134","article-title":"FinRL-Meta: Market environments and benchmarks for data-driven financial reinforcement learning","volume":"35","author":"X.-Y Liu","year":"2022","journal-title":"Advances in Neural Information Processing Systems"},{"key":"ref17","first-page":"14129","article-title":"MOPO: Modelbased offline policy optimization","volume":"33","author":"T Yu","year":"2020","journal-title":"Advances in Neural Information Processing Systems, H. Larochelle, M. Ranzato"},{"key":"ref18","article-title":"Replicating anomalies","author":"K Hou","year":"2017","journal-title":"National Bureau of Economic Research"},{"key":"ref19","first-page":"1","article-title":"Time-series and cross-sectional stock return forecasting: New machine learning methods","author":"D E Rapach","year":"2020","journal-title":"Machine Learning for Asset Management: New Developments and Financial Applications"},{"issue":"2","key":"ref20","doi-asserted-by":"crossref","first-page":"62","DOI":"10.3905\/jpm.2018.44.2.062","article-title":"Dynamic allocation or diversification: A regime-based approach to multiple assets","volume":"44","author":"P Nystrup","year":"2018","journal-title":"J. Portfolio Manag"},{"issue":"4","key":"ref21","doi-asserted-by":"crossref","first-page":"2153","DOI":"10.1109\/TIT.2017.2653803","article-title":"Riemannian Gaussian distributions on the space of symmetric positive definite matrices","volume":"63","author":"S Said","year":"2017","journal-title":"IEEE Trans. Inf. Theory"},{"key":"ref22","author":"N","year":"2023","journal-title":"An Introduction to Optimization on Smooth Manifolds"},{"issue":"1","key":"ref23","doi-asserted-by":"crossref","DOI":"10.1002\/sta4.291","article-title":"Relating and comparing methods for detecting changes in mean","volume":"9","author":"P Fearnhead","year":"2020","journal-title":"Stat"},{"key":"ref24","author":"N Chopin","year":"2020","journal-title":"An Introduction to Sequential Monte Carlo"},{"key":"ref25","first-page":"1861","article-title":"Soft actor-critic: Off-policy maximum entropy deep reinforcement learning with a stochastic actor","author":"T Haarnoja","year":"2018","journal-title":"Proc. Int. Conf. Mach. Learn. (PMLR)"},{"key":"ref26","author":"O Gottesman","year":"2018","journal-title":"Evaluating reinforcement learning algorithms in observational health settings"},{"key":"ref27","first-page":"1126","article-title":"Model-agnostic meta-learning for fast adaptation of deep networks","author":"C Finn","year":"2017","journal-title":"Proc. Int. Conf. Mach. Learn. (PMLR)"},{"issue":"1","key":"ref28","doi-asserted-by":"crossref","DOI":"10.1038\/srep00541","article-title":"DebtRank: Too central to fail? Financial networks, the FED and systemic risk","volume":"2","author":"S Battiston","year":"2012","journal-title":"Sci. Rep"},{"key":"ref29","article-title":"On the Impact of Network Topology on Systemic Risk: Network Reconstruction, Stress Testing and Control","author":"D Grigat","year":"2018","journal-title":"Dept. Comput. Sci"},{"issue":"10","key":"ref30","doi-asserted-by":"crossref","first-page":"4839","DOI":"10.1093\/rfs\/hhz147","article-title":"Shock transmission through cross-border bank lending: Credit and real effects","volume":"33","author":"G Hale","year":"2020","journal-title":"Rev. Financ. Stud"},{"issue":"3","key":"ref31","doi-asserted-by":"crossref","first-page":"339","DOI":"10.1198\/073500102288618487","article-title":"Dynamic conditional correlation: A simple class of multivariate generalized autoregressive conditional heteroskedasticity models","volume":"20","author":"R Engle","year":"2002","journal-title":"J. Bus. Econ. Stat"},{"issue":"3","key":"ref32","doi-asserted-by":"crossref","first-page":"282","DOI":"10.1080\/07350015.2013.771027","article-title":"Dynamic conditional correlation: On properties and estimation","volume":"31","author":"G P Aielli","year":"2013","journal-title":"J. Bus. Econ. Stat"},{"issue":"3","key":"ref33","doi-asserted-by":"crossref","first-page":"652","DOI":"10.1080\/07350015.2020.1713795","article-title":"Fitting vast dimensional time-varying covariance models","volume":"39","author":"C Pakel","year":"2021","journal-title":"J. Bus. Econ. Stat"},{"issue":"3","key":"ref34","doi-asserted-by":"crossref","first-page":"1263","DOI":"10.1111\/0022-1082.00247","article-title":"Conditional skewness in asset pricing tests","volume":"55","author":"C R Harvey","year":"2000","journal-title":"J. Finance"},{"key":"ref35","first-page":"261","article-title":"Explaining models by propagating Shapley values of local components","author":"H Chen","year":"2020","journal-title":"Explainable AI in Healthcare and Medicine: Building a Culture of Transparency and Accountability"},{"issue":"1","key":"ref36","doi-asserted-by":"crossref","first-page":"203","DOI":"10.1007\/s10614-020-10042-0","article-title":"Explainable machine learning in credit risk management","volume":"57","author":"N Bussmann","year":"2021","journal-title":"Comput. Econ"},{"issue":"2","key":"ref37","doi-asserted-by":"crossref","first-page":"816","DOI":"10.1214\/23-AOS2276","article-title":"Conformal prediction beyond exchangeability","volume":"51","author":"R F Barber","year":"2023","journal-title":"Ann. Stat"},{"key":"ref38","author":"V Vovk","year":"2005","journal-title":"Algorithmic Learning in a Random World"},{"issue":"8","key":"ref39","doi-asserted-by":"crossref","first-page":"1271","DOI":"10.1080\/14697688.2019.1571683","article-title":"Deep hedging","volume":"19","author":"H Buehler","year":"2019","journal-title":"Quant. Finance"}],"container-title":[],"original-title":[],"deposited":{"date-parts":[[2026,10,7]],"date-time":"2026-10-07T13:55:40Z","timestamp":1791381340000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.ssrn.com\/abstract=7576742"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026]]},"references-count":39,"URL":"https:\/\/doi.org\/10.2139\/ssrn.7576742","relation":{},"subject":[],"published":{"date-parts":[[2026]]},"subtype":"preprint"}}