{"paper_id":"2506.09080","title":"FinHEAR: Human Expertise and Adaptive Risk-Aware Temporal Reasoning for Financial Decision-Making","authors":["Jiaxiang Chen","Mingxi Zou","Zhuo Wang","Qifan Wang","Dongning Sun","Chi Zhang","Zenglin Xu"],"paper_date":"2025-06-10","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2506.09080v2","page_url":"https://thequant.space/flowcharts/finhear-human-expertise-and-adaptive-risk-aware-temporal-re/","scores":{"empirical_rigor":6,"math_complexity":4,"quadrant":"Street Traders"},"proposes":{"status":"available","statement":"The paper proposes FinHEAR, a multi‑agent framework that integrates human expertise with adaptive risk‑aware temporal reasoning for financial decision‑making.","quote":"We propose FinHEAR, a multi-agent framework for Human Expertise and Adaptive Risk-aware reasoning.","summary":"Financial decision-making presents unique challenges for language models, demanding temporal reasoning, adaptive risk assessment, and responsiveness to dynamic events. While large language models (LLMs) show strong general reasoning capabilities, they often fail to capture behavioral patterns…","strategy":"FinHEAR"},"assumptions":{"status":"available","items":[{"assumption":"The backtest assumes a conservative transaction cost of 5 basis points per side.","type":"execution","stated":"explicit","quote":"We further conducted a comprehensive backtest incorporating conservative transaction costs (5 basis points per side)."},{"assumption":"The dataset covers daily market data from December 2019 to December 2023.","type":"data","stated":"explicit","quote":"we construct a multi-source financial dataset spanning December 2019 to December 2023"},{"assumption":"Expert‑guided retrieval is assumed to reduce information asymmetry by emulating bounded rationality.","type":"model","stated":"implied","quote":"To reduce information asymmetry, one agent retrieves relevant expert cases to emulate human‑guided decisions under bounded rationality"},{"assumption":"Position sizes are adjusted based on prediction confidence to model loss‑averse behavior.","type":"model","stated":"implied","quote":"sizes based on prediction confidence, modeling the loss‑averse behavior described by Prospect Theory"}]},"datasets":{"status":"available","items":[{"name":"Multi‑stock and commodity price dataset","provider":null,"kind":"market-prices","access":"unknown","period":"December 2019 – December 2023","universe":"AAPL, TSLA, XOM, XAUUSD","quote":"We primarily focus on three stocks from different sectors: Apple Inc. (AAPL), Tesla, Inc. (TSLA), and Exxon Mobil Corporation (XOM), as well as one commodity asset: Gold (XAUUSD)."},{"name":"Macro‑economic news and reports","provider":null,"kind":"macro","access":"unknown","period":"December 2019 – December 2023","universe":null,"quote":"daily public news and important macroeconomic literature and data such as the official minutes of the U.S. Federal Open Market Committee (FOMC) and periodic reports of the Consumer Price Index (CPI)."}],"known":[],"sample":{"start":"2023-01","end":"2023-12","market":"multiple financial assets","asset_class":"multi-asset","frequency":"daily"},"data_resolution":"daily"},"transaction_costs":{"status":"available","modeled":"partial","headline_net_of_costs":false,"detail":"Conservative transaction costs of 5 basis points per side are applied in a robustness backtest, not in the headline results.","quote":"We further conducted a comprehensive backtest incorporating conservative transaction costs (5 basis points per side).","source":"front_matter","rule":"R3","candidates":{"front_matter":"partial","evidence_card":"partial","evidence_card_quoted":true},"card_detail":"Conservative transaction costs of 5 basis points per side are applied in a robustness backtest, not in the headline results.","card_quote":"We further conducted a comprehensive backtest incorporating conservative transaction costs (5 basis points per side).","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":"FinHEAR consistently outperforms strong baselines across trend prediction and trading tasks, achieving higher accuracy and better risk‑adjusted returns.","kind":"performance","quote":"FinHEAR consistently outperforms strong baselines across trend prediction and trading tasks, achieving higher accuracy and better risk‑adjusted returns."},{"finding":"FinHEAR achieves superior forecasting accuracy (ACC) and Matthews Correlation Coefficient (MCC) compared to rule‑based, deep‑learning, and other LLM‑based baselines on all tested assets.","kind":"predictive-accuracy","quote":"FinHEAR consistently outperforms rule‑based, deep learning, and LLM‑based baselines across forecasting and decision‑making metrics."}],"reported_performance":{"strategy":"FinHEAR","sharpe":3.28,"sharpe_benchmark":null,"annual_return_pct":null,"total_return_pct":83.89,"max_drawdown_pct":15.77},"out_of_sample":{"method":"holdout","headline_out_of_sample":true}},"independent_evidence":{"status":"none_found","supports":[],"contradicts":[],"qualifies":[],"compared_with":36,"citations":{"cited_by_in_corpus":0,"by_independent_authors":0,"how_cited":{"used_as_baseline":0,"uses_method_or_data":0,"extends":0,"discusses_findings":0,"mentions":0,"unclassified":0},"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/Jackson-1102/FinHEAR","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":"2026-11-14T17:35:56Z"},"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/"}