{"paper_id":"2405.05449","title":"Markowitz Meets Bellman: Knowledge-distilled Reinforcement Learning for Portfolio Management","authors":["Gang Hu","Ming Gu"],"paper_date":"2024-05-08","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2405.05449v1","page_url":"https://thequant.space/flowcharts/markowitz-meets-bellman-knowledge-distilled-reinforcement-l/","scores":{"empirical_rigor":6,"math_complexity":8.5,"quadrant":"Holy Grail"},"proposes":{"status":"available","statement":"The paper proposes a hybrid portfolio management model called Knowledge Distillation DDPG (KDD) that combines Markowitz theory with reinforcement learning in two training stages.","quote":"In particular, our proposed method, called KDD (Knowledge Distillation DDPG), consist of two training stages: supervised and reinforcement learning stages.","summary":"Investment portfolios, central to finance, balance potential returns and risks. This paper introduces a hybrid approach combining Markowitz’s portfolio theory with reinforcement learning, utilizing knowledge distillation for training agents. In particular, our proposed method, called KDD (Knowledge…","strategy":"KDD (Knowledge Distillation DDPG)"},"assumptions":{"status":"available","items":[{"assumption":"Investors are risk‑averse, seeking to minimize uncertainty and potential losses while maximizing expected returns for a given level of risk.","type":"market","stated":"explicit","quote":"This theory is predicated on the assumption that investors are risk‑averse, seeking to minimize uncertainty and potential losses while maximizing expected returns for a given level of risk."},{"assumption":"The Markowitz model provides optimal portfolio allocations that can be used as target actions for supervised pre‑training of the DDPG actor network.","type":"model","stated":"implied","quote":"Initially, we generate a dataset based on the Markowitz model, simulating optimal investment portfolios under various market conditions."}]},"datasets":{"status":"available","items":[{"name":"Markowitz‑generated synthetic portfolio dataset","provider":null,"kind":"synthetic","access":"unknown","period":null,"universe":null,"quote":"Initially, we generate a dataset based on the Markowitz model, simulating optimal investment portfolios under various market conditions."}],"known":[],"sample":{"start":null,"end":null,"market":null,"asset_class":null,"frequency":null},"data_resolution":"daily"},"transaction_costs":{"status":"available","modeled":"no","headline_net_of_costs":false,"detail":null,"quote":"these models frequently overlook critical factors like transaction costs, thus limiting their practicality in real‑world trading scenarios.","source":"front_matter","rule":"R3","candidates":{"front_matter":"no","evidence_card":"no","evidence_card_quoted":true},"card_detail":null,"card_quote":"these models frequently overlook critical factors like transaction costs, thus limiting their practicality in real‑world trading scenarios.","meaning":"a trading or execution result is reported without costs"},"results":{"status":"available","evaluation_type":"trading-pnl","findings":[{"finding":"The KDD model achieves a total return of 138.38%, far surpassing benchmarks such as DJI (51.41%) and the Markowitz portfolio (69.43%).","kind":"performance","quote":"The KDD model demonstrates exceptional performance, with a Total Return of 138.38%, outshining all other strategies, including the DJI and Markowitz portfolio, which yielded returns of 51.41% and 69.43%, respectively."},{"finding":"The model attains an annualized return of 38.74%, the highest among all compared strategies.","kind":"performance","quote":"Moreover, the model achieves an Annualized Return of 38.74%, surpassing all competitors."},{"finding":"The KDD model records the highest Sharpe Ratio of 2.03, indicating superior risk‑adjusted performance.","kind":"performance","quote":"In terms of risk‑adjusted returns, the model attains the highest Sharpe Ratio of 2.03, indicating not only higher returns but also a superior risk‑return balance."}],"reported_performance":{"strategy":"KDD (Knowledge Distillation DDPG)","sharpe":2.03,"sharpe_benchmark":null,"annual_return_pct":null,"total_return_pct":null,"max_drawdown_pct":null},"out_of_sample":{"method":"holdout","headline_out_of_sample":true}},"independent_evidence":{"status":"none_found","supports":[],"contradicts":[],"qualifies":[],"compared_with":30,"citations":{"cited_by_in_corpus":0,"by_independent_authors":0,"how_cited":{"used_as_baseline":0,"uses_method_or_data":0,"extends":0,"discusses_findings":0,"mentions":0,"unclassified":0},"independent_papers":[],"scope":"Citations between papers in this corpus only; citations from outside it are not counted."}},"implementation":{"status":"available","public":false,"code_url":null,"note":"No repository presented by the authors as their own was found in the paper. Absence here is not proof that no code exists."},"replication":{"status":"none_found","card":null,"note":"No replication of this paper has been run by thequant.space yet."},"extraction":{"text_source":"fulltext","model":"openai/gpt-oss-120b","version":1,"extracted_at":"2026-11-14T18:09:47Z"},"provenance":{"claims":"Everything here is what the authors state in the paper. Nothing was re-run or reproduced by the platform.","quotes":"Assumptions, datasets, cost detail and findings were extracted by a language model. Every quote is verbatim paper text. Where an item has `support`, \"full\" means every figure in it appears in the quoted sentences and \"partial\" means a figure is stated elsewhere in the paper; items with a figure found nowhere in the paper were discarded.","datasets_known":"datasets.known lists well-known vendors and public datasets by a stable id, found by a dictionary over the full text where the sentence says the data is used; it is independent of the model and covers only names in the dictionary.","flags":"transaction_costs.modeled, results.evaluation_type and results.out_of_sample.method are platform classifications (structured-decision models); see https://thequant.space/score-guide/#penalty-flags","comparisons":"supports / contradicts / qualifies are a language model's reading of two papers' quoted findings, restricted to closely related papers that share no author. They are pointers to read, not verified replications, and coverage is partial.","implementation":"code_url is a repository the paper itself presents as its own (rule-scored, ambiguous cases judged by a model), checked for liveness when it was recorded."},"docs":"https://thequant.space/agents/"}