{"paper_id":"2511.07014","title":"Diffolio: A Diffusion Model for Multivariate Probabilistic Financial Time-Series Forecasting and Portfolio Construction","authors":["So-Yoon Cho","Jin-Young Kim","Kayoung Ban","Hyeng Keun Koo","Hyun-Gyoon Kim"],"paper_date":"2025-11-10","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2511.07014v1","page_url":"https://thequant.space/flowcharts/diffolio-a-diffusion-model-for-multivariate-probabilistic-f/","scores":{"empirical_rigor":6,"math_complexity":8.5,"quadrant":"Holy Grail"},"proposes":{"status":"available","statement":"The paper proposes Diffolio, a diffusion-based model with hierarchical attention for multivariate financial time-series forecasting and portfolio construction.","quote":"In this paper, we propose Diffolio, a diffusion model designed for multivariate financial time-series forecasting and portfolio construction.","summary":"Probabilistic forecasting is crucial in multivariate financial time-series for constructing efficient portfolios that account for complex cross-sectional dependencies. In this paper, we propose Diffolio, a diffusion model designed for multivariate financial time-series forecasting and portfolio…","strategy":"Diffolio"},"assumptions":{"status":"available","items":[{"assumption":"Portfolio optimization is performed without transaction cost constraints in the primary experiments.","type":"execution","stated":"implied","quote":"Since the portfolio optimization process is conducted without transaction cost constraints, the models focus exclusively on the predicted joint distribution to determine optimal weights at each time step."},{"assumption":"A stable target correlation matrix can be estimated via Ledoit‑Wolf shrinkage and used to guide attention.","type":"model","stated":"explicit","quote":"we introduce a correlation-guided regularizer that employs a stable target correlation matrix estimated via the Ledoit‑Wolf shrinkage method."},{"assumption":"Daily excess returns of industry portfolios are forecast one day ahead.","type":"data","stated":"explicit","quote":"we forecast one-day-ahead daily excess returns of 12 industry portfolios"},{"assumption":"Asset‑specific and systematic covariates are available and informative for forecasting.","type":"data","stated":"explicit","quote":"For the model’s covariates, we employ asset characteristics as the asset-specific variables and macroeconomic variables as the systematic variables"},{"assumption":"The market can be rebalanced weekly with a flat transaction cost of 10 basis points.","type":"execution","stated":"explicit","quote":"we explicitly integrate a transaction cost of 10 basis points (bps) as a penalty term directly into the portfolio optimization objective function and adjust the rebalancing frequency to a five-day interval (weekly)"},{"assumption":"The deterministic DDIM sampling procedure provides efficient generation of return samples.","type":"model","stated":"explicit","quote":"For inference, we employ the deterministic DDIM sampling procedure described in Section 3.2, which allows for efficient generation of return samples in a reduced number of steps."}]},"datasets":{"status":"available","items":[{"name":"12 industry portfolios","provider":"Kenneth R. French’s data library","kind":"market-prices","access":"public","period":"1958-2023","universe":"12 Fama‑French industry portfolios","quote":"We forecast one-day-ahead daily excess returns of 12 industry portfolios","known_id":"french-library"}],"known":[{"id":"french-library","name":"Kenneth French Data Library","provider":"Kenneth R. French","kind":"market-prices","access":"public","quote":"These portfolios—comprising Consumer Non-durables (NoDur), Consumer Durables (Durbl), Manufacturing (Manuf), Energy (En-3The data is sourced from Kenneth R. French’s data library: https://mba.tuck.dartmouth.edu/ pages/faculty/ken.french/data_library.html.","mentions":1,"used_in":1}],"sample":{"start":null,"end":null,"market":"U.S. industry portfolios","asset_class":"equities","frequency":"daily"},"data_resolution":"daily"},"transaction_costs":{"status":"available","modeled":"yes","headline_net_of_costs":true,"detail":"10 basis points per trade, weekly (five‑day) rebalancing","quote":"we explicitly integrate a transaction cost of 10 basis points (bps) as a penalty term directly into the portfolio optimization objective function and adjust the rebalancing frequency to a five-day interval (weekly)","source":"front_matter","rule":"R3","candidates":{"front_matter":"yes","evidence_card":"yes","evidence_card_quoted":true},"card_detail":"10 basis points per trade, weekly (five‑day) rebalancing","card_quote":"we explicitly integrate a transaction cost of 10 basis points (bps) as a penalty term directly into the portfolio optimization objective function and adjust the rebalancing frequency to a five-day interval (weekly)","meaning":"costs are part of the reported headline results"},"results":{"status":"available","evaluation_type":"trading-pnl","findings":[{"finding":"Diffolio outperforms baselines in multivariate forecasting accuracy and portfolio performance, achieving higher Sharpe ratios and certainty equivalents.","kind":"performance","quote":"Diffolio outperforms various probabilistic forecasting baselines in both statistical multivariate forecasting accuracy and portfolio performance"},{"finding":"Diffolio attains the best Energy Score (ES) and second‑best CRPS, with the lowest CRPS standard deviation across assets.","kind":"predictive-accuracy","quote":"the lowest standard deviation in CRPS across the 12 assets."},{"finding":"In a frictionless setting, most models exhibit high turnover, but Diffolio remains responsive to daily market signals.","kind":"other","quote":"... high turnover is incidental to the primary objectives of maximizing MVP‑SR or GOP‑CE; it serves as a measure of the model’s responsiveness."},{"finding":"When realistic transaction costs (10 bps) and weekly rebalancing are applied, Diffolio still delivers superior portfolio metrics compared to benchmarks.","kind":"performance","quote":"Following the portfolio results in the frictionless environment, we assess the practical viability"}],"reported_performance":{"strategy":"Diffolio","sharpe":0.7206,"sharpe_benchmark":0.5119,"annual_return_pct":16.8,"total_return_pct":null,"max_drawdown_pct":55.77},"out_of_sample":{"method":"unclear","headline_out_of_sample":null}},"independent_evidence":{"status":"none_found","supports":[],"contradicts":[],"qualifies":[],"compared_with":34,"citations":{"cited_by_in_corpus":1,"by_independent_authors":0,"how_cited":{"used_as_baseline":0,"uses_method_or_data":0,"extends":0,"discusses_findings":0,"mentions":0,"unclassified":1},"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":"2026-11-14T17:22:05Z"},"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/"}