{"paper_id":"2505.05121","title":"Error Analysis of Deep PDE Solvers for Option Pricing","authors":["Jasper Rou"],"paper_date":"2025-05-08","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2505.05121v1","page_url":"https://thequant.space/flowcharts/error-analysis-of-deep-pde-solvers-for-option-pricing/","scores":{"empirical_rigor":6.5,"math_complexity":7,"quadrant":"Holy Grail"},"proposes":{"status":"available","statement":"The paper empirically evaluates two neural network PDE solvers, the Deep Galerkin Method and the Time Deep Gradient Flow method, for option pricing in Black‑Scholes and Heston models.","quote":"Through comparative experiments in both the Black--Scholes and the Heston model, we assess the empirical performance of two neural network algorithms to solve PDEs: the Deep Galerkin Method and the Time Deep Gradient Flow method (TDGF).","summary":"Option pricing often requires solving partial differential equations (PDEs). Although deep learning-based PDE solvers have recently emerged as quick solutions to this problem, their empirical and quantitative accuracy remain not well understood, hindering their real-world applicability. In this…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"Option prices follow the Black‑Scholes PDE with known analytical solution.","type":"model","stated":"explicit","quote":"Indeed, let u : [0, T] × Ω→R denote the price of this derivative, with Ω⊆R and t the time to maturity. Then, u solves the Black–Scholes PDE"},{"assumption":"The Heston model does not have a closed‑form price, but a reference price can be obtained via the COS method using its characteristic function.","type":"model","stated":"explicit","quote":"This PDE does not have an exact solution. The characteristic function of the Heston model does have an analytical representation [13], from which a reference price can be determined using the COS method [8]."}]},"datasets":{"status":"available","items":[{"name":"Black‑Scholes synthetic option data","provider":null,"kind":"synthetic","access":"unknown","period":null,"universe":null,"quote":"This PDE has an exact solution: u(t, S) = SΦ(d1) −Ke−rtΦ(d2), with Φ the standard normal cumulative distribution function,"},{"name":"Heston synthetic option data (COS reference prices)","provider":null,"kind":"synthetic","access":"unknown","period":null,"universe":null,"quote":"The characteristic function of the Heston model does have an analytical representation [13], from which a reference price can be determined using the COS method [8]."}],"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 L2‑error decreases almost linearly with the number of sampling stages for both methods, but the benefit stops after a certain point.","kind":"performance","quote":"Since the training time grows linearly with the the number of sampling stages, it would be optimal to stop at this point."},{"finding":"Increasing the number of layers tends to reduce error, though without a clear rate; two or three layers are recommended.","kind":"performance","quote":"Since the number of layers has a big influence on the training time, we recommend using two or three layers."},{"finding":"For TDGF, more time steps lower the L2‑error, with second‑order stepping reducing error faster than first‑order.","kind":"performance","quote":"For the TDGF, the number of time steps decreases the L2-error. Using a second-order time‑stepping method, the error decreases quicker than using a first‑order method."},{"finding":"For TDGF, larger numbers of nodes per layer improve accuracy; for DGM, error does not decrease with nodes per layer, so a smaller size is advised.","kind":"performance","quote":"For the DGM, the L2‑error does not decrease with the number of nodes per layer. We recommend choosing a smaller number of nodes per layer like thirty."},{"finding":"TDGF benefits slightly from more training samples, while DGM shows no error reduction with additional samples.","kind":"performance","quote":"For the TDGF the L2‑error decreases slightly with the number of samples"}],"reported_performance":null,"out_of_sample":{"method":"none","headline_out_of_sample":null}},"independent_evidence":{"status":"available","supports":[],"contradicts":[{"paper":{"paper_id":"2304.09750","title":"Application of Tensor Neural Networks to Pricing Bermudan Swaptions","paper_date":"2023-04-18","authors":["Raj G. Patel","Tomas Dominguez","Mohammad Dib","Samuel Palmer","Andrea Cadarso","Fernando De Lope Contreras"],"empirical_rigor":7,"page_url":"https://thequant.space/flowcharts/application-of-tensor-neural-networks-to-pricing-bermudan-swaptions/"},"question":"How does increasing the number of layers in neural networks affect pricing accuracy/error for option pricing models?","this_paper_finding":"Increasing the number of layers tends to reduce error, though without a clear rate; two or three layers are recommended.","other_paper_finding":"Dense Neural Networks suffer from over‑fitting when depth is increased to 20 layers, leading to worse price accuracy and stability.","other_paper_quote":"the Dense Neural Network price for 20 layers falls subject to over‑fitting and is considerably worse than the price observed in the 10 layer setting both in terms of accuracy and stability.","note":"A reports over‑fitting and worse accuracy at 20 layers, while B finds error generally decreases with more layers and recommends 2‑3 layers.","confidence":0.487}],"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":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":"2026-11-14T17:39:03Z"},"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/"}