{"paper_id":"2307.13546","title":"Transfer Learning for Portfolio Optimization","authors":["Haoyang Cao","Haotian Gu","Xin Guo","Mathieu Rosenbaum"],"paper_date":"2023-07-25","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2307.13546v1","page_url":"https://thequant.space/flowcharts/transfer-learning-for-portfolio-optimization/","scores":{"empirical_rigor":6,"math_complexity":8,"quadrant":"Holy Grail"},"proposes":{"status":"available","statement":"The paper proposes a novel concept called transfer risk for portfolio optimization via transfer learning.","quote":"concept called “transfer risk”, within the optimization framework for transfer learning","summary":"In this work, we explore the possibility of utilizing transfer learning techniques to address the financial portfolio optimization problem. We introduce a novel concept called “transfer risk”, within the optimization framework of transfer learning. A series of numerical experiments are conducted…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"Source and target tasks share the same input and output spaces (Rd and R).","type":"model","stated":"explicit","quote":"assume that the source and target tasks S and T share the same input and output spaces denoted by XS = XT = Rd and YS = YT = R"},{"assumption":"Both source and target models are linear mappings from returns to portfolio return.","type":"model","stated":"explicit","quote":"the admissible sets of source and target models are restricted to: AS = AT = {f : Rd →R|f(r) = ϕ⊤r for some ϕ ∈X}"},{"assumption":"The regularization hyper‑parameter λ is positive and controls the strength of the L2 penalty.","type":"model","stated":"explicit","quote":"λ > 0 is a hyper-parameter controls the power of the regularization"},{"assumption":"Over‑night holding is not allowed for trading frequencies ranging from 1‑minute to 130‑minute.","type":"execution","stated":"explicit","quote":"we assume that over-night holding is not allowed for trading frequencies ranging from 1-minute to 130-minute"},{"assumption":"Historical daily returns are used to estimate mean and covariance matrices for both source and target tasks.","type":"data","stated":"explicit","quote":"Empirically, µP and ΣP are estimated from the historical return"}]},"datasets":{"status":"available","items":[{"name":"Source training data (cross‑continent transfer)","provider":null,"kind":"market-prices","access":"unknown","period":"February 2000 to February 2020","universe":"ten S&P500 component stocks","quote":"Source training data: it consists of the daily returns of ten source assets, from February 2000 to February 2020."},{"name":"Target training data (cross‑continent transfer)","provider":null,"kind":"market-prices","access":"unknown","period":"February 2015 to February 2020","universe":"ten target assets (largest market caps in a given market)","quote":"Target training data: it consists of the daily returns of ten target assets, from February 2015 to February 2020."},{"name":"Target testing data (cross‑continent transfer)","provider":null,"kind":"market-prices","access":"unknown","period":"February 2020 to September 2021","universe":"same ten target assets","quote":"Target testing data: it consists of the daily returns of ten target assets, from February 2020 to September 2021."},{"name":"Sector transfer data","provider":null,"kind":"market-prices","access":"unknown","period":"February 2015 to February 2020 (training","universe":"ten stocks from a given target sector and ten stocks from a source sector","quote":"Target training data: it consists of the daily returns of ten stocks from a given target sector, from February 2015 to February 2020."},{"name":"Frequency transfer data","provider":null,"kind":"market-prices","access":"unknown","period":"September 2019 to February 2020 (trainin","universe":"ten stocks sampled under a given target frequency","quote":"it consists of the returns of ten stocks, sampled under a given target frequency, from September 2019 to February 2020."}],"known":[],"sample":{"start":null,"end":null,"market":null,"asset_class":null,"frequency":null},"data_resolution":"intraday-bars"},"transaction_costs":{"status":"available","modeled":"no","headline_net_of_costs":null,"detail":null,"quote":null,"source":"front_matter","rule":"R1","candidates":{"front_matter":"no","evidence_card":"no","evidence_card_quoted":false},"meaning":"a trading or execution result is reported without costs"},"results":{"status":"available","evaluation_type":"trading-pnl","findings":[{"finding":"A strong correlation is observed between transfer risk and transfer learning performance across markets.","kind":"statistical","quote":"we observe a strong correlation between transfer risk and the transfer learning performance across all the different markets."},{"finding":"Negative correlation between transfer risk scores and Sharpe ratios; lower transfer risk leads to higher Sharpe.","kind":"performance","quote":"In general, the negative correlation between Sharpe ratios and scores is observed"},{"finding":"For the Health Care target sector, source sectors Information Technology, Health Care, and Consumer Discretionary yield higher Sharpe ratios, while Energy is unsuitable.","kind":"performance","quote":"Information Technology, Health Care, and Consumer Discretionary are desirable source sectors when the target sector is Health Care, while Energy is not a suitable choice."},{"finding":"Source frequencies of 65‑minute and 130‑minute provide better transfer learning performance than higher frequencies.","kind":"performance","quote":"Figure 7 reveal that 65-minute and 130-minute are more appropriate candidates for the source frequency, since they lead to much better transfer learning performance"},{"finding":"Transfer risk can be computed efficiently and used as a prior estimate of final learning outcome before full transfer learning.","kind":"theoretical","quote":"transfer risk is an easy-to-compute and viable quantity as a prior estimate of the final learning outcome"}],"reported_performance":null,"out_of_sample":{"method":"holdout","headline_out_of_sample":null}},"independent_evidence":{"status":"available","supports":[],"contradicts":[],"qualifies":[],"compared_with":17,"citations":{"cited_by_in_corpus":2,"by_independent_authors":2,"how_cited":{"used_as_baseline":1,"uses_method_or_data":0,"extends":0,"discusses_findings":0,"mentions":0,"unclassified":1},"independent_papers":[{"paper":{"paper_id":"2511.21221","title":"Portfolio Optimization via Transfer Learning","paper_date":"2025-11-26","authors":["Kexin Wang","Xiaomeng Zhang","Xinyu Zhang"],"empirical_rigor":6,"page_url":"https://thequant.space/flowcharts/portfolio-optimization-via-transfer-learning/"},"how":"used_as_baseline","quote":null}],"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":"2026-11-14T18:24:21Z"},"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/"}