{"paper_id":"2504.18600","title":"QuantBench: Benchmarking AI Methods for Quantitative Investment","authors":["Saizhuo Wang","Hao Kong","Jiadong Guo","Fengrui Hua","Yiyan Qi","Wanyun Zhou","Jiahao Zheng","Xinyu Wang","Lionel M. Ni","Jian Guo"],"paper_date":"2025-04-24","paper_type":"dataset-benchmark","paper_url":"https://arxiv.org/pdf/2504.18600v1","page_url":"https://thequant.space/flowcharts/quantbench-benchmarking-ai-methods-for-quantitative-investm/","scores":{"empirical_rigor":7,"math_complexity":3.5,"quadrant":"Street Traders"},"proposes":{"status":"available","statement":"The paper proposes QuantBench, an industrial‑grade benchmark platform for AI in quantitative investment that standardizes evaluation, supports flexible AI integration, and covers the full investment pipeline.","quote":"We present QuantBench, an industrial‑grade benchmark platform designed to address this critical need. QuantBench offers three key strengths: (1) standardization that aligns with quantitative investment industry practices, (2) flexibility to integrate various AI algorithms, and (3) full‑pipeline coverage of the entire q","summary":"The field of artificial intelligence (AI) in quantitative investment has seen significant advancements, yet it lacks a standardized benchmark aligned with industry practices. This gap hinders research progress and limits the practical application of academic innovations. We present QuantBench, an…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"The benchmark assumes a ranking‑based stock selection strategy that picks the top 300 stocks each cross‑section without feature selection.","type":"execution","stated":"implied","quote":"The backtest used a ranking‑based stock selection strategy, selecting the top 300 stocks at each cross‑section without applying feature selection."},{"assumption":"The market simulation framework can vary in realism depending on the specific trading scenario.","type":"market","stated":"implied","quote":"QuantBench offers a detailed market simulation framework that can vary in realism depending on the specific trading scenario."},{"assumption":"Financial data exhibits low signal‑to‑noise ratios and non‑stationarity.","type":"data","stated":"implied","quote":"task‑agnostic metrics to address the unique challenges presented by financial data, including low signal‑to‑noise ratios and non‑stationarity."}]},"datasets":{"status":"available","items":[{"name":"Chinese stock market data","provider":null,"kind":"market-prices","access":"unknown","period":null,"universe":null,"quote":"Using Chinese stock market data, we employed two feature sets—Alpha101 and Alpha158—while comparing XGBoost (tree‑based) and LSTM (DNN) models."}],"known":[],"sample":{"start":null,"end":null,"market":null,"asset_class":null,"frequency":null},"data_resolution":"unclear"},"transaction_costs":{"status":"available","modeled":"partial","headline_net_of_costs":null,"detail":"Execution metrics such as slippage, market impact, commissions and transaction fees are considered in the evaluation.","quote":"Other costs such as commissions and transaction fees are also considered.","source":"evidence_card","rule":"R2","candidates":{"front_matter":"unclear","evidence_card":"partial","evidence_card_quoted":true},"card_detail":"Execution metrics such as slippage, market impact, commissions and transaction fees are considered in the evaluation.","card_quote":"Other costs such as commissions and transaction fees are also considered.","meaning":"costs appear only as a robustness check, a flat fee without spread or impact, or outside the headline numbers"},"results":{"status":"available","evaluation_type":"trading-pnl","findings":[{"finding":"DNN models achieve higher Information Coefficient (IC) than tree‑based models, but tree models deliver better returns and Sharpe ratios on the Alpha101 feature set.","kind":"predictive-accuracy","quote":"The results in Table 2 indicate that DNNs generally outperform tree‑based models in IC. However, tree models produced better returns and Sharpe ratios with Alpha101."},{"finding":"Random validation set construction outperforms the traditional tail validation method, likely due to more diverse data patterns.","kind":"performance","quote":"The random validation set method outperformed the tail method, likely because it introduced more diverse data patterns into the validation process."}],"reported_performance":null,"out_of_sample":{"method":"unclear","headline_out_of_sample":null}},"independent_evidence":{"status":"none_found","supports":[],"contradicts":[],"qualifies":[],"compared_with":14,"citations":{"cited_by_in_corpus":4,"by_independent_authors":4,"how_cited":{"used_as_baseline":0,"uses_method_or_data":0,"extends":0,"discusses_findings":0,"mentions":3,"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":"2026-11-14T17:39:57Z"},"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/"}