{"paper_id":"2403.11738","title":"A path-dependent PDE solver based on signature kernels","authors":["Alexandre Pannier","Cristopher Salvi"],"paper_date":"2024-03-18","paper_type":"theoretical","paper_url":"https://arxiv.org/pdf/2403.11738v3","page_url":"https://thequant.space/flowcharts/a-path-dependent-pde-solver-based-on-signature-kernels/","scores":{"empirical_rigor":4,"math_complexity":8.5,"quadrant":"Lab Rats"},"proposes":{"status":"available","statement":"The paper proposes a kernel-based solver for path‑dependent PDEs using signature kernels, with a convergence theory and closed‑form solution in the linear case.","quote":"We develop a kernel-based solver for path-dependent PDEs (PPDEs) along with a convergence theory.","summary":"We develop a provably convergent kernel-based solver for path-dependent PDEs (PPDEs). Our numerical scheme leverages signature kernels, a recently introduced class of kernels on path-space. Specifically, we solve an optimal recovery problem by approximating the solution of a PPDE with an element of…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"Strict assumptions guaranteeing consistency and convergence of the scheme as the number of collocation points increases.","type":"other","stated":"explicit","quote":"Under strict assumptions, we prove consistency of the proposed scheme, guaranteeing convergence to the PPDE solution as the number of collocation points increases."},{"assumption":"Injectivity of the signature (point‑separating property) ensured by paths having one strictly monotone coordinate.","type":"model","stated":"explicit","quote":"This is a technical assumption needed to ensure injectivity of the signature and consequently the point separating property assumption needed to establish universality of the signature kernel."},{"assumption":"The static kernel g is four times differentiable with bounded derivatives up to order 4.","type":"statistical","stated":"explicit","quote":"Assumption 3.10. g : V × V →R is four times differentiable and, for all 0 ≤n, m ≤4, we have ... < ∞."},{"assumption":"State‑dependent payoffs are used in numerical experiments (though the method allows path‑dependent payoffs).","type":"data","stated":"explicit","quote":"For clarity of exposition and because well‑posedness results are scarcer in the path‑dependent case, we will however stick to state‑dependent payoffs in our numerical experiments."},{"assumption":"The kernel‑based solver does not require additional training costs; the Gram matrix is estimated once offline.","type":"execution","stated":"explicit","quote":"Our kernel‑based solver does not require additional “training” costs: the estimation of the Gram matrix is performed once, offline, and can then be reused…"}]},"datasets":{"status":"available","items":[{"name":"Synthetic rough‑volatility option pricing data","provider":null,"kind":"synthetic","access":"unknown","period":null,"universe":null,"quote":"We consider two values of H = 0.1 and H = 0.3 and two strike values K = 0.1 and K = 1.0. We take the same number of sample paths and collocation points for Monte Carlo and the signature kernel solver respectively..."}],"known":[],"sample":{"start":null,"end":null,"market":null,"asset_class":null,"frequency":null},"data_resolution":"synthetic"},"transaction_costs":{"status":"not_applicable","modeled":"n/a","headline_net_of_costs":null,"detail":null,"quote":null,"source":"front_matter","rule":"R3","candidates":{"front_matter":"n/a","evidence_card":"n/a","evidence_card_quoted":false},"meaning":"the paper reports no trading or execution result"},"results":{"status":"available","evaluation_type":"pricing-calibration","findings":[{"finding":"The numerical performance (MSE and MAE) of the signature‑kernel PPDE solver is comparable to Monte Carlo across different numbers of paths and collocation points.","kind":"performance","quote":"As it can be observed from the tables, the numerical performance in terms of MSE and MAE of the two algorithms are comparable across the different settings."},{"finding":"The solver’s computational complexity scales as O(h⁻² ĤN d²) and can be reduced to O(h⁻¹ ĤN d²) on a GPU.","kind":"theoretical","quote":"The complexity of the resulting solver scales as O(h−2 ˆNd2) where this time ˆN is the number of collocation paths. This cost can be further reduced to O(h−1 ˆNd2) if the operations are carried out on a GPU."}],"reported_performance":null,"out_of_sample":{"method":"none","headline_out_of_sample":null}},"independent_evidence":{"status":"available","supports":[],"contradicts":[],"qualifies":[],"compared_with":20,"citations":{"cited_by_in_corpus":4,"by_independent_authors":3,"how_cited":{"used_as_baseline":0,"uses_method_or_data":1,"extends":0,"discusses_findings":0,"mentions":1,"unclassified":2},"independent_papers":[{"paper":{"paper_id":"2411.19317","title":"Deep learning interpretability for rough volatility","paper_date":"2024-11-28","authors":["Bo Yuan","Damiano Brigo","Antoine Jacquier","Nicola Pede"],"empirical_rigor":3,"page_url":"https://thequant.space/flowcharts/deep-learning-interpretability-for-rough-volatility/"},"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/crispitagorico/sigppde","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":"2026-11-14T18:12:56Z"},"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/"}