{"paper_id":"2402.06656","title":"DiffsFormer: A Diffusion Transformer on Stock Factor Augmentation","authors":["Yuan Gao","Haokun Chen","Xiang Wang","Zhicai Wang","Xue Wang","Jinyang Gao","Bolin Ding"],"paper_date":"2024-02-05","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2402.06656v1","page_url":"https://thequant.space/flowcharts/diffsformer-a-diffusion-transformer-on-stock-factor-augment/","scores":{"empirical_rigor":6,"math_complexity":8,"quadrant":"Holy Grail"},"proposes":{"status":"available","statement":"The paper proposes DiffsFormer, a conditional diffusion Transformer that generates stock factors to augment training data for stock forecasting.","quote":"we introduce the Diffusion Model to generate stock factors with Transformer architecture (DiffsFormer).","summary":"Machine learning models have demonstrated remarkable efficacy and efficiency in a wide range of stock forecasting tasks. However, the inherent challenges of data scarcity, including low signal-to-noise ratio (SNR) and data homogeneity, pose significant obstacles to accurate forecasting. To address…","strategy":"DiffsFormer"},"assumptions":{"status":"none_found","items":[]},"datasets":{"status":"available","items":[{"name":"CSI300","provider":"Shanghai Stock Exchange and the Shenzhen Stock Exchange","kind":"market-prices","access":"public","period":"2008-01-01 to 2022-09-30","universe":"largest 300 A‑share stocks in China","quote":"we evaluate the proposed framework on two real‑world stock datasets: CSI 300 and CSI 800."},{"name":"CSI800","provider":"Shanghai Stock Exchange and the Shenzhen Stock Exchange","kind":"market-prices","access":"public","period":"2008-01-01 to 2022-09-30","universe":"CSI 500 plus CSI 300, covering smaller‑size A‑share stocks","quote":"CSI 800 is a larger dataset consisting of CSI 500 and CSI 300, aiming to add some stocks with smaller size."}],"known":[],"sample":{"start":null,"end":null,"market":"CSI300 and CSI800 stocks","asset_class":"equities","frequency":"daily"},"data_resolution":"daily"},"transaction_costs":{"status":"available","modeled":"no","headline_net_of_costs":false,"detail":null,"quote":null,"source":"front_matter","rule":"R1","candidates":{"front_matter":"no","evidence_card":"unclear","evidence_card_quoted":false},"meaning":"a trading or execution result is reported without costs"},"results":{"status":"available","evaluation_type":"trading-pnl","findings":[{"finding":"DiffsFormer‑augmented training yields relative improvements of 7.2% and 27.8% in annualized return ratio on the CSI300 and CSI800 datasets respectively.","kind":"performance","quote":"The proposed method achieves relative improvements of 7.2% and 27.8% in annualized return ratio for the respective datasets."},{"finding":"Editing existing samples (DiffsFormer) achieves low Fréchet Inception Distance (FID) as low as 0.6872, indicating high fidelity of generated data.","kind":"performance","quote":"achieve FID as low as 0.6872, suggesting a high fidelity of the generated data."},{"finding":"Loss‑guided noise addition further increases the information ratio (IR) and reduces volatility compared to uniform noise addition.","kind":"performance","quote":"loss‑guided diffusion can further increase IR and decrease the volatility."}],"reported_performance":{"strategy":"DiffsFormer","sharpe":null,"sharpe_benchmark":null,"annual_return_pct":null,"total_return_pct":null,"max_drawdown_pct":null},"out_of_sample":{"method":"unclear","headline_out_of_sample":null}},"independent_evidence":{"status":"none_found","supports":[],"contradicts":[],"qualifies":[],"compared_with":23,"citations":{"cited_by_in_corpus":5,"by_independent_authors":2,"how_cited":{"used_as_baseline":0,"uses_method_or_data":0,"extends":0,"discusses_findings":0,"mentions":2,"unclassified":3},"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":"2026-11-14T18:15:13Z"},"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/"}