{"paper_id":"2502.07393","title":"FinRL-DeepSeek: LLM-Infused Risk-Sensitive Reinforcement Learning for Trading Agents","authors":["Mostapha Benhenda"],"paper_date":"2025-02-11","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2502.07393v1","page_url":"https://thequant.space/flowcharts/finrl-deepseek-llm-infused-risk-sensitive-reinforcement-lea/","scores":{"empirical_rigor":6,"math_complexity":6.5,"quadrant":"Holy Grail"},"proposes":{"status":"available","statement":"The paper proposes a novel risk-sensitive trading agent that combines reinforcement learning with large language models, extending CPPO with LLM-generated risk and recommendation signals.","quote":"We extend the Conditional Value-at-Risk Proximal Policy Optimization (CPPO) algorithm, by adding risk assessment and trading recommendation signals generated by a LLM from financial news.","summary":"This paper presents a novel risk-sensitive trading agent combining reinforcement learning and large language models (LLMs). We extend the Conditional Value-at-Risk Proximal Policy Optimization (CPPO) algorithm, by adding risk assessment and trading recommendation signals generated by a LLM from…","strategy":"CPPO-DeepSeek"},"assumptions":{"status":"available","items":[{"assumption":"A 10% LLM infusion is used in the tests, meaning the original PPO and CPPO are modified up to 10%.","type":"model","stated":"explicit","quote":"In these tests, we use a 10% infusion of LLM, which means that the original PPO and CPPO up to 10%"},{"assumption":"One representative news article per stock per day is randomly selected, reducing the dataset to 2 million records.","type":"data","stated":"explicit","quote":"To reduce LLM API costs, we randomly select a representative news article per stock per day, reducing the dataset to 2 million records."},{"assumption":"CVaR confidence level is set to 0.05, representing the worst 5% of outcomes.","type":"statistical","stated":"explicit","quote":"α is the CVaR confidence level (e.g., 0.05 for the worst 5%)."}]},"datasets":{"status":"available","items":[{"name":"FNSPID dataset","provider":"Dong et al., 2024","kind":"text-news","access":"unknown","period":"1999–2023","universe":"Nasdaq-100 constituents","quote":"We use the FNSPID dataset (Dong et al., 2024), which contains 15.7 million time-aligned financial news records spanning 1999–2023.","known_id":"fnspid"}],"known":[{"id":"fnspid","name":"FNSPID","provider":"Dong et al.","kind":"text-news","access":"public","quote":"Data and LLM Prompts We use the FNSPID dataset (Dong et al., 2024), which contains 15.7 million time-aligned financial news records spanning 1999–2023.","mentions":6,"used_in":3}],"sample":{"start":"2019","end":"2023","market":"Nasdaq-100 index","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":"Integrating LLM stock recommendations consistently improves PPO cumulative returns, but does not yet beat the Nasdaq‑100 benchmark.","kind":"performance","quote":"These preliminary results in figure 1 with Qwen 2.5 show that integrating LLM stock recommendations consistently improves PPO in terms of cumulative returns. However, it’s not sufficient to beat Nasdaq 100 benchmark so far."},{"finding":"PPO and CPPO‑DeepSeek outperform other methods, including the Nasdaq‑100 benchmark, in stochastic runs.","kind":"performance","quote":"In these two runs (RL agents are stochastic), PPO and CPPO-DeepSeek outperform other methods, including the Nasdaq 100 benchmark."},{"finding":"Stronger LLM infusion generally degrades PPO‑DeepSeek performance, even with very small (0.1%) perturbations.","kind":"negative","quote":"For PPO-DeepSeek, stronger LLM infusion mostly degrades performance, even for tiny (0.1%) perturbations."},{"finding":"Stronger LLM infusion improves CPPO‑DeepSeek performance.","kind":"performance","quote":"For CPPO-DeepSeek, stronger LLM infusion improves performance"}],"reported_performance":{"strategy":"CPPO-DeepSeek","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":"available","supports":[],"contradicts":[],"qualifies":[],"compared_with":24,"citations":{"cited_by_in_corpus":3,"by_independent_authors":3,"how_cited":{"used_as_baseline":0,"uses_method_or_data":1,"extends":0,"discusses_findings":0,"mentions":1,"unclassified":1},"independent_papers":[{"paper":{"paper_id":"2410.14927","title":"Hierarchical Reinforced Trader (HRT): A Bi-Level Approach for Optimizing Stock Selection and Execution","paper_date":"2024-10-19","authors":["Zijie Zhao","Roy E. Welsch"],"empirical_rigor":6,"page_url":"https://thequant.space/flowcharts/hierarchical-reinforced-trader-hrt-a-bi-level-approach-fo/"},"how":"uses_method_or_data","quote":null}],"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/benstaf/FinRL_DeepSeek","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":"2026-11-14T17:45:48Z"},"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/"}