{"paper_id":"2508.02759","title":"Hedging with memory: shallow and deep learning with signatures","authors":["Eduardo Abi Jaber","Louis-Amand Gérard"],"paper_date":"2025-08-03","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2508.02759v1","page_url":"https://thequant.space/flowcharts/hedging-with-memory-shallow-and-deep-learning-with-signatur/","scores":{"empirical_rigor":4,"math_complexity":7.5,"quadrant":"Lab Rats"},"proposes":{"status":"available","statement":"The paper proposes using path signatures as features in feed‑forward neural networks for hedging exotic derivatives, showing they outperform LSTMs with far less training compute.","quote":"In a deep learning setting, we use signatures as features in feedforward neural networks and show that they outperform LSTMs in most cases, with orders of magnitude less training compute.","summary":"We investigate the use of path signatures in a machine learning context for hedging exotic derivatives under non-Markovian stochastic volatility models. In a deep learning setting, we use signatures as features in feedforward neural networks and show that they outperform LSTMs in most cases, with…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"The spot price follows a geometric Brownian motion for the polynomial payoff experiment.","type":"model","stated":"explicit","quote":"where S is a geometric Brownian motion in order to have an explicit solution to compare the methods to."},{"assumption":"The European call payoff assumes the underlying follows the Heston model with CIR variance dynamics.","type":"model","stated":"explicit","quote":"where S follows the Heston model with its variance Vt = Σ2 t having CIR dynamics (5.2)."},{"assumption":"The Asian call payoff assumes a geometric price process with delayed‑equation volatility.","type":"model","stated":"explicit","quote":"where S will be geometric with delayed-equation (DE) volatility (5.3)."},{"assumption":"The methods assume access to the expected signature of the price or volatility, either exactly or via Monte‑Carlo simulation.","type":"data","stated":"explicit","quote":"We assume throughout this section that we are given such data (either exactly or through Monte Carlo simulations)."},{"assumption":"Linear signature regressions are agnostic to the underlying dynamics and can be applied in general contexts.","type":"other","stated":"explicit","quote":"By contrast, the linear signature regressions are agnostic to the underlying dynamics and can be applied in more general contexts."},{"assumption":"Training and evaluation are performed over multiple random initialisations (50 for SNN, 10 for RNN).","type":"execution","stated":"explicit","quote":"All numerical experiments have been averaged over 50 iterations for the SNN and 10 for the RNN, lower for obvious time reasons."}]},"datasets":{"status":"none_found","items":[],"known":[],"sample":{"start":null,"end":null,"market":null,"asset_class":null,"frequency":null},"data_resolution":"synthetic"},"transaction_costs":{"status":"available","modeled":"no","headline_net_of_costs":null,"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":"pricing-calibration","findings":[{"finding":"Signature‑based feed‑forward networks achieve better hedging performance than LSTMs while requiring orders of magnitude less training compute.","kind":"performance","quote":"In a deep learning setting, we use signatures as features in feedforward neural networks and show that they outperform LSTMs in most cases, with orders of magnitude less training compute."},{"finding":"Calibrating a signature volatility model and solving the hedging problem within it yields more accurate and stable results across payoffs and volatility dynamics.","kind":"performance","quote":"Solving the hedging problem in the calibrated signature volatility model yields more accurate and stable results across different payoffs and volatility dynamics."},{"finding":"Both signature and log‑signature neural networks plateau after truncation order 2, with the signature NN achieving lower mean‑squared P&L than the log‑signature NN under the Heston model.","kind":"performance","quote":"Figure 12 shows that the signature and log‑signature have a significantly different impact on training with the Log‑Signature NN plateauing relatively high and early in the truncation, whereas the Signature NN goes lower and does not seems to plateau under the Heston model."},{"finding":"Most of the hedging performance can be attributed to the first two or three orders of the signature.","kind":"theoretical","quote":"conclusions are much less clear but we can still see that most of the performance can be attributed to the first 2 or 3 orders of the signature."},{"finding":"Higher truncation orders beyond 3 provide little to no additional benefit and may even delay convergence.","kind":"negative","quote":"However, apart from a small improvement between truncation 2 and 3 for the look‑back call option, it seems that higher orders either don’t help or actually delay the convergence."}],"reported_performance":null,"out_of_sample":{"method":"holdout","headline_out_of_sample":null}},"independent_evidence":{"status":"available","supports":[],"contradicts":[{"paper":{"paper_id":"2508.15237","title":"Option pricing under non-Markovian stochastic volatility models: A deep signature approach","paper_date":"2025-08-21","authors":["Jingtang Ma","Xianglin Wu","Wenyuan Li"],"empirical_rigor":4,"page_url":"https://thequant.space/flowcharts/option-pricing-under-non-markovian-stochastic-volatility-models-a-deep/"},"question":"How does increasing truncation order affect model convergence or approximation error?","this_paper_finding":"Higher truncation orders beyond 3 provide little to no additional benefit and may even delay convergence.","other_paper_finding":"Approximation errors for option prices decrease monotonically with higher truncation levels for both SDE and PDE schemes.","other_paper_quote":"Both numerical schemes show monotonically decreasing errors with higher truncation levels","note":"A finds orders >3 give little benefit and may delay convergence, while B finds errors decrease monotonically with higher truncation levels.","confidence":0.601},{"paper":{"paper_id":"2307.09767","title":"Sig-Splines: universal approximation and convex calibration of time series generative models","paper_date":"2023-07-19","authors":["Magnus Wiese","Phillip Murray","Ralf Korn"],"empirical_rigor":3,"page_url":"https://thequant.space/flowcharts/sig-splines-universal-approximation-and-convex-calibration/"},"question":"How does truncation order affect model performance?","this_paper_finding":"Higher truncation orders beyond 3 provide little to no additional benefit and may even delay convergence.","other_paper_finding":"Higher orders of truncation improve autocorrelation metrics, indicating better capture of serial independence and volatility clustering.","other_paper_quote":"Moreover, it is observed that higher orders of truncation lead to improved performance in autocorrelation metrics, which detect serial independence and volatility clustering.","note":"A finds higher orders improve autocorrelation metrics, while B finds orders >3 give little benefit and may delay convergence.","confidence":0.463}],"qualifies":[],"compared_with":25,"citations":{"cited_by_in_corpus":2,"by_independent_authors":2,"how_cited":{"used_as_baseline":0,"uses_method_or_data":1,"extends":0,"discusses_findings":0,"mentions":1,"unclassified":0},"independent_papers":[{"paper":{"paper_id":"2609.34474","title":"Deep kernel hedging","paper_date":"2026-09-28","authors":["Jean-Loup Dupret","Donatien Hainaut","Edouard Motte"],"empirical_rigor":7,"page_url":"https://thequant.space/flowcharts/deep-kernel-hedging/"},"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":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":"2026-11-14T17:30:47Z"},"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/"}