{"paper_id":"2407.18645","title":"Contrastive Learning of Asset Embeddings from Financial Time Series","authors":["Rian Dolphin","Barry Smyth","Ruihai Dong"],"paper_date":"2024-07-26","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2407.18645v1","page_url":"https://thequant.space/flowcharts/contrastive-learning-of-asset-embeddings-from-financial-time/","scores":{"empirical_rigor":7,"math_complexity":5.5,"quadrant":"Holy Grail"},"proposes":{"status":"available","statement":"The paper proposes a novel contrastive learning framework to generate asset embeddings from financial time series data, using a statistical sampling strategy based on hypothesis testing.","quote":"We propose a novel contrastive learning framework to generate asset embeddings from financial time series data.","summary":"Representation learning has emerged as a powerful paradigm for extracting valuable latent features from complex, high-dimensional data. In financial domains, learning informative representations for assets can be used for tasks like sector classification, and risk management. However, the complex…","strategy":"contrastive learning approach with individual sigmoid loss"},"assumptions":{"status":"available","items":[{"assumption":"The null hypothesis assumes equal likelihood of co‑occurrence between assets (p0 = 1/N).","type":"statistical","stated":"explicit","quote":"is the expected co-occurrence probability under the assumption of equal likelihood, i.e."}]},"datasets":{"status":"available","items":[{"name":"daily returns dataset","provider":null,"kind":"market-prices","access":"public","period":"2000-2018","universe":"611 U.S. stocks","quote":"For this evaluation we use a publicly available, real-world financial dataset consisting of daily returns for a universe of 611 U.S. stocks from 2000 to 20181."}],"known":[],"sample":{"start":"2013","end":"2018","market":"multiple assets (real-world datasets)","asset_class":"equities","frequency":"daily"},"data_resolution":"daily"},"transaction_costs":{"status":"not_extracted","modeled":"unclear","headline_net_of_costs":null,"detail":null,"quote":null,"source":"none","rule":"R0","candidates":{"front_matter":"unclear","evidence_card":"unclear","evidence_card_quoted":false},"meaning":"the classifier was not confident"},"results":{"status":"available","evaluation_type":"unclear","findings":[{"finding":"The contrastive learning approach with the sigmoid‑softmax loss achieves the best sector classification performance, with an F1‑score of 0.69 and accuracy of 69%.","kind":"performance","quote":"The results show that our contrastive learning approach using the sigmoid-softmax loss achieves the best performance across all metrics, with an F1-score of 0.69 and an accuracy of 69%."},{"finding":"In the hedging experiment, the individual sigmoid loss yields the lowest average out‑of‑sample volatility of 19.1%, outperforming the Pearson‑correlation baseline with statistical significance.","kind":"performance","quote":"The results show that our contrastive learning approach with the individual sigmoid loss achieves the lowest average volatility of 19.1%, outperforming the baseline method based on Pearson correlation with statistical significance."}],"reported_performance":{"strategy":"contrastive learning approach with individual sigmoid loss","sharpe":null,"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":7,"citations":{"cited_by_in_corpus":1,"by_independent_authors":1,"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":true,"code_url":"https://github.com/rian-dolphin/stock-embeddings","note":null},"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":"2026-11-14T18:05:14Z"},"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/"}