{"paper_id":"2408.10368","title":"Deep-MacroFin: Informed Equilibrium Neural Network for Continuous Time Economic Models","authors":["Yuntao Wu","Jiayuan Guo","Goutham Gopalakrishna","Zissis Poulos"],"paper_date":"2024-08-19","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2408.10368v4","page_url":"https://thequant.space/flowcharts/deep-macrofin-informed-equilibrium-neural-network-for-conti/","scores":{"empirical_rigor":2.5,"math_complexity":9.5,"quadrant":"Lab Rats"},"proposes":{"status":"available","statement":"The paper proposes Deep-MacroFin, a comprehensive deep learning framework for solving high‑dimensional PDEs in continuous‑time economic models.","quote":"In this paper, we present Deep-MacroFin, a comprehensive framework designed to solve partial differential equations, with a particular focus on models in continuous time economics.","summary":"In this paper, we present Deep-MacroFin, a comprehensive framework designed to solve partial differential equations, with a particular focus on models in continuous time economics. This framework leverages deep learning methodologies, including Multi-Layer Perceptrons and the newly developed…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"Time is continuous and the horizon is infinite.","type":"model","stated":"explicit","quote":"Time is continuous and the horizon is infinite."},{"assumption":"The model assumes the presence of two state variables: η (experts’ wealth share) and σ (volatility).","type":"model","stated":"explicit","quote":"Figure 1: System overview. This model assumes the presence of two state variables: η (experts’ wealth share) and σ (volatility), two agents: Ve (experts) and Vh (households), and one endogenous variable: q (price of the capital)."},{"assumption":"Agents can short‑sell when η is below a critical share ηψ.","type":"market","stated":"explicit","quote":"When η ∈(0, ηψ), agents can short-sell, leading to a nonlinear rise in the capital price q."},{"assumption":"Loss weights are manually selected for best empirical performance.","type":"other","stated":"explicit","quote":"we manually select them for best empirical performance, as automatic methods proved less robust."},{"assumption":"The neural network approximations satisfy the equilibrium equations without any assumption about the solution itself.","type":"model","stated":"explicit","quote":"Models that minimize L(θ, T ) ensures the approximated functions satisfy the specified equations and constraints."}]},"datasets":{"status":"available","items":[{"name":"100‑dimensional Laplace equation","provider":null,"kind":"synthetic","access":"unknown","period":null,"universe":null,"quote":"Specifically, we demonstrate the ability to solve a 100‑dimensional Laplace equation using 4.3GB CUDA memory, 12 GFLOPs, with an average epoch time of 0.16 seconds"},{"name":"Lucas orchard benchmark","provider":null,"kind":"benchmark","access":"unknown","period":null,"universe":null,"quote":"Figure 5: Lucas orchard benchmark"}],"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":"unclear","findings":[{"finding":"Deep‑MacroFin uses 5× less CUDA memory and 40× fewer FLOPs than existing libraries on 100‑dimensional problems.","kind":"performance","quote":"fewer FLOPs in 100D problems) and user‑friendly implementation than existing libraries."},{"finding":"The framework can solve a 100‑dimensional Laplace equation with 4.3 GB memory and 12 GFLOPs, averaging 0.16 s per epoch.","kind":"performance","quote":"Specifically, we demonstrate the ability to solve a 100‑dimensional Laplace equation using 4.3GB CUDA memory, 12 GFLOPs, with an average epoch time of 0.16 seconds"},{"finding":"A time‑stepping scheme enables convergence of a 50‑dimensional economic model within 10 minutes.","kind":"performance","quote":"With time‑stepping and ergodic sampling, the 50‑tree model converges within 10 minutes."},{"finding":"KANs provide analytical expressions for critical wealth share ηψ, e.g., ψ = −2.2823 + 3.7254 exp … implying ηψ ≈0.3197.","kind":"theoretical","quote":"A key advantage of KAN is its ability to yield an analytical expression for ηψ. For example: ψ = −2.2823 + 3.7254 exp −1.0082(0.6741 −η)2 implies ηψ ≈0.3197 for ψ(ηψ) = 1."},{"finding":"Scaling KANs to higher dimensions is challenging due to numerical instability and limited parallelism.","kind":"negative","quote":"scaling KANs to higher dimensions remains challenging due to numerical instability with deeper or wider networks and limited parallelism."}],"reported_performance":null,"out_of_sample":{"method":"none","headline_out_of_sample":null}},"independent_evidence":{"status":"none_found","supports":[],"contradicts":[],"qualifies":[],"compared_with":25,"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":"2026-11-14T18:03:55Z"},"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/"}