{"paper_id":"2511.05519","title":"An uncertainty-aware physics-informed neural network solution for the Black-Scholes equation: a novel framework for option pricing","authors":["Sina Kazemian","Ghazal Farhani","Amirhessam Yazdi"],"paper_date":"2025-10-27","paper_type":"methodological","paper_url":"https://arxiv.org/pdf/2511.05519v1","page_url":"https://thequant.space/flowcharts/an-uncertainty-aware-physics-informed-neural-network-solutio/","scores":{"empirical_rigor":6,"math_complexity":8.5,"quadrant":"Holy Grail"},"proposes":{"status":"available","statement":"The paper proposes an uncertainty‑aware physics‑informed neural network (PINN) with an anchored‑ensemble fine‑tuning stage for solving the Black‑Scholes PDE and pricing options.","quote":"We present an uncertainty-aware, physics-informed neural network (PINN) for option pricing that solves the Black–Scholes (BS) partial differential equation (PDE) as a mesh-free, global surrogate over (S,t).","summary":"We present an uncertainty-aware, physics-informed neural network (PINN) for option pricing that solves the Black–Scholes (BS) partial differential equation (PDE) as a mesh-free, global surrogate over (S,t). The model embeds the BS operator and boundary/terminal conditions in a residual-based…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"The Black‑Scholes dynamics with constant risk‑free rate and volatility (no‑dividend) are assumed for the pricing PDE.","type":"model","stated":"explicit","quote":"To enforce no-arbitrage under a non-dividend assumption, labels y are clipped to instrument-specific bounds—American put: L = max(K −S, 0), U = K; European call: L = max(S −Ke^{−rτ}, 0), U = S."},{"assumption":"Early‑exercise feature of American options is handled via an obstacle‑style relaxation while keeping the BS residual in the continuation region.","type":"model","stated":"explicit","quote":"For American options, we handle early exercise via an obstacle‑style relaxation while retaining the BS residual in the continuation region."},{"assumption":"The PINN is trained without any labeled option price data, relying solely on the PDE residual and boundary/terminal conditions.","type":"data","stated":"explicit","quote":"The model embeds the BS operator and boundary/terminal conditions in a residual‑based objective and requires no labeled prices."}]},"datasets":{"status":"available","items":[{"name":"Yahoo Finance option chain dataset","provider":"Yahoo Finance","kind":"options","access":"public","period":null,"universe":"AAPL, MSFT, GOOGL, AMZN, META, F, T, NOK, CCL, SOFI, SNAP, NIO, AAL, PBR, RIVN","quote":"We build supervised datasets for American puts and European calls on the Yahoo Finance tickers AAPL, MSFT, GOOGL, AMZN, META, F, T, NOK, CCL, SOFI, SNAP, NIO, AAL, PBR, RIVN.","known_id":"yahoo-finance"}],"known":[],"sample":{"start":null,"end":null,"market":null,"asset_class":null,"frequency":null},"data_resolution":"unclear"},"transaction_costs":{"status":"available","modeled":"unclear","headline_net_of_costs":null,"detail":null,"quote":null,"source":"conflict","rule":"R4b","candidates":{"front_matter":"n/a","evidence_card":"no","evidence_card_quoted":true},"conflict_kind":"applicability","card_detail":null,"card_quote":"... transaction costs, stochastic volatility, and jumps.","meaning":"the platform classifier and the paper-text extraction disagree, so no value is asserted"},"results":{"status":"available","evaluation_type":"pricing-calibration","findings":[{"finding":"The anchored PINN achieves low pricing errors for European options (MAE ≈5×10⁻², RMSE ≈7×10⁻²) with explained variance ≈0.999.","kind":"predictive-accuracy","quote":"On European calls/puts, the approach attains low errors (e.g., MAE"},{"finding":"For American puts the method remains accurate (MAE/RMSE on the order of 10⁻¹ with EV ≈0.999) and avoids error accumulation of time‑marching schemes.","kind":"predictive-accuracy","quote":"with EV ≈0.999) and avoids the error accumulation typical of time-marching sc"},{"finding":"Across four parameter sets, the PINN‑anchor outperforms ANN and RNN and matches or slightly exceeds the KAN baseline, while training more stably.","kind":"performance","quote":"Consistent with the tables, PINN‑anchor and KAN generally outperform ANN/RNN, with visibly smaller deviations near the strike. Between PINN‑anchor and KAN, accuracy is similar across most parameter sets; however, KAN trains markedly slower (about 10×) and exhibits"},{"finding":"The model provides calibrated epistemic uncertainty bands that respect no‑arbitrage bounds and reflect model uncertainty.","kind":"predictive-accuracy","quote":"predictions are returned with calibrated uncertainty bands that (i) respect no-arbitrage bounds through our bounded-output mapping and (ii) reflect model uncertainty arising from model behaviour."}],"reported_performance":null,"out_of_sample":{"method":"unclear","headline_out_of_sample":null}},"independent_evidence":{"status":"none_found","supports":[],"contradicts":[],"qualifies":[],"compared_with":47,"citations":{"cited_by_in_corpus":1,"by_independent_authors":1,"how_cited":{"used_as_baseline":0,"uses_method_or_data":0,"extends":0,"discusses_findings":0,"mentions":1,"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":"2026-11-14T17:23:25Z"},"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/"}