{"paper_id":"2308.08918","title":"IMM: An Imitative Reinforcement Learning Approach with Predictive Representation Learning for Automatic Market Making","authors":["Hui Niu","Siyuan Li","Jiahao Zheng","Zhouchi Lin","Jian Li","Jian Guo","Bo An"],"paper_date":"2023-08-17","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2308.08918v1","page_url":"https://thequant.space/flowcharts/imm-an-imitative-reinforcement-learning-approach-with-predi/","scores":{"empirical_rigor":6,"math_complexity":7.5,"quadrant":"Holy Grail"},"proposes":{"status":"available","statement":"The paper proposes Imitative Market Maker (IMM), a novel reinforcement learning framework that combines expert signals and imitation learning to efficiently learn multi-price level market making strategies.","quote":"we propose Imitative Market Maker (IMM), a novel RL framework leveraging both knowledge from suboptimal signal-based experts and direct policy interactions to develop multi-price level MM strategies efficiently.","summary":"Market making (MM) has attracted significant attention in financial trading owing to its essential function in ensuring market liquidity. With strong capabilities in sequential decision-making, Reinforcement Learning (RL) technology has achieved remarkable success in quantitative trading.…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"Buy limit orders are placed on the bid side and sell orders on the ask side.","type":"market","stated":"explicit","quote":"It is assumed that buy limit orders are placed on the bid side, and sell ones on the ask side."},{"assumption":"A stable reference price (pref) can be defined to encode the LOB and multi‑price level orders.","type":"model","stated":"explicit","quote":"enabling effective encoding of LOB and multi-price level orders."}]},"datasets":{"status":"available","items":[{"name":"four real‑world market datasets","provider":null,"kind":"order-book","access":"unknown","period":null,"universe":null,"quote":"Extensive experimental results on four real‑world market datasets demonstrate that IMM outperforms current RL‑based market making strategies in terms of several financial criteria."},{"name":"four futures markets","provider":null,"kind":"order-book","access":"unknown","period":null,"universe":null,"quote":"Experiments on four futures markets demonstrate that IMM outperforms the benchmarks, and further ablation studies verify the effectiveness of the components in the proposed method."}],"known":[],"sample":{"start":null,"end":null,"market":null,"asset_class":null,"frequency":null},"data_resolution":"l2-book"},"transaction_costs":{"status":"none_found","modeled":"unclear","headline_net_of_costs":null,"detail":null,"quote":null,"source":"none","rule":"R0b","candidates":{"front_matter":"unclear","evidence_card":"no","evidence_card_quoted":false},"meaning":"no cost treatment was found in the passages the extractor read; that is not evidence that costs are not modeled"},"results":{"status":"available","evaluation_type":"trading-pnl","findings":[{"finding":"IMM significantly outperforms benchmark methods in both profitability and risk management across all test datasets.","kind":"performance","quote":"These comparison results indicate that the proposed approach significantly outperforms the benchmarks in terms of both profitability and risk management."},{"finding":"On the RB dataset, IMM achieves the highest terminal wealth and risk‑adjusted return, with only a slightly higher MAP compared to the expert and other RL‑based methods.","kind":"performance","quote":"As demonstrated in Table 1, on the RB dataset, the proposed method attains the highest terminal wealth as well as risk‑ajusted return, albeit with a slightly elevated MAP in comparison to the expert and the two RL‑based methods."},{"finding":"On the FU dataset, IMM obtains the highest terminal wealth, the best return‑to‑risk performance, and superior spread‑capturing ability while maintaining the second‑lowest inventory level.","kind":"performance","quote":"On the FU dataset, IMM not only achieves the highest terminal wealth but also demonstrates the most favorable return‑to‑risk performance and spread‑capturing ability, while maintaining the second‑lowest inventory level."},{"finding":"Ablation studies confirm that the state/action representations, the SL‑based prediction model, and the TCSA network each contribute positively to IMM’s performance.","kind":"performance","quote":"The findings of the ablation study substantiate the effectiveness of the model components."}],"reported_performance":null,"out_of_sample":{"method":"holdout","headline_out_of_sample":null}},"independent_evidence":{"status":"none_found","supports":[],"contradicts":[],"qualifies":[],"compared_with":19,"citations":{"cited_by_in_corpus":2,"by_independent_authors":2,"how_cited":{"used_as_baseline":0,"uses_method_or_data":0,"extends":0,"discusses_findings":0,"mentions":1,"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":"2026-11-14T18:23:15Z"},"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/"}