{"paper_id":"2411.02375","title":"Whack-a-mole Online Learning: Physics-Informed Neural Network for Intraday Implied Volatility Surface","authors":["Kentaro Hoshisashi","Carolyn E. Phelan","Paolo Barucca"],"paper_date":"2024-11-04","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2411.02375v1","page_url":"https://thequant.space/flowcharts/whack-a-mole-online-learning-physics-informed-neural-networ/","scores":{"empirical_rigor":7,"math_complexity":8.5,"quadrant":"Holy Grail"},"proposes":{"status":"available","statement":"The paper proposes Whack-a-mole Online Learning (WamOL), a calibration algorithm that combines self‑adaptive and loss‑balancing techniques within PINNs to calibrate time‑dependent implied volatility surfaces in real‑time.","quote":"We introduce the Whack-a-mole Online Learning (WamOL) algorithm, which combines self-adaptive and loss-balancing learning techniques to calibrate time-dependent IVS using PINNs.","summary":"Calibrating the time-dependent Implied Volatility Surface (IVS) using sparse market data is an essential challenge in computational finance, particularly for real-time applications. This task requires not only fitting market data but also satisfying a specified partial differential equation (PDE)…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"Option prices must satisfy no‑arbitrage inequalities on first and second derivatives.","type":"model","stated":"explicit","quote":"These principles state that market prices should not allow for guaranteed returns exceeding the risk‑free rate. We consider the necessary and sufficient conditions for no‑arbitrage in [2, 11], which are represented as non‑strict inequalities involving several first and second derivatives, −1"},{"assumption":"Market data are sparse, uneven, and arrive intraday, requiring online learning updates.","type":"data","stated":"explicit","quote":"WamOL leverages online learning techniques to facilitate real‑time intraday calibration, enabling the model to adapt swiftly to evolving market conditions."},{"assumption":"Only out‑of‑the‑money (OTM) S&P 500 options are used, with forward prices derived via put‑call parity linear regression.","type":"market","stated":"explicit","quote":"The data source provided spot prices, selecting out‑of‑the‑money (OTM) options, and we derived forward prices using linear regression based on put‑call parity."},{"assumption":"Training and validation errors are measured on 15‑minute intervals throughout the trading day.","type":"execution","stated":"explicit","quote":"calibration time (solid lines) and validated using the calibrated model on data from the rest of the day (dotted lines) at 15‑minute intervals"}]},"datasets":{"status":"available","items":[{"name":"Intraday S&P 500 options dataset","provider":"CBOE DataShop","kind":"options","access":"commercial","period":"2022‑10‑01 to 2023‑09‑30 (248 business d","universe":"S&P 500 options (OTM)","quote":"We conducted backtesting on extensive intraday options data for WamOL using a dataset of S&P 500 options. The historical data comprised intraday traded prices of S&P 500 options for 248 business days from October 1, 2022, to September 30, 2023. This dataset contains approximately 500,000 data points per day with 1 ms p","known_id":"cboe"}],"known":[{"id":"cboe","name":"Cboe","provider":"Cboe Global Markets","kind":"options","access":"commercial","quote":"This dataset contains approximately 500, 000 data points per day with 1 ms precision, obtained from CBOE DataShop2.","mentions":1,"used_in":1}],"sample":{"start":null,"end":null,"market":null,"asset_class":null,"frequency":null},"data_resolution":"unclear"},"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":"WamOL consistently achieves lower calibration and validation errors than competing models (MLP, standard PINNs) across training and prediction phases.","kind":"performance","quote":"WamOL consistently achieves lower calibration and validation errors than the other models overall in both training and predictions."},{"finding":"WamOL exhibits faster and more scalable computation times than baseline models, with sub‑linear growth as dataset size increases.","kind":"performance","quote":"Table 3 presents the computation times required by WamOL and the baseline models for calibrating the IVS on datasets of varying sizes. The computational efficiency of WamOL is evident from its ability to handle these complex optimisation challenges without significant overhead"}],"reported_performance":null,"out_of_sample":{"method":"unclear","headline_out_of_sample":null}},"independent_evidence":{"status":"none_found","supports":[],"contradicts":[],"qualifies":[],"compared_with":28,"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":"2026-11-14T17:55:57Z"},"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/"}