dorsal/arxiv
View SchemaDeePM: Regime-Robust Deep Learning for Systematic Macro Portfolio Management
| Authors | Kieran Wood, Stephen J. Roberts, Stefan Zohren |
|---|---|
| Categories | |
| ArXiv ID | 2601.05975vv1 |
| URL | https://arxiv.org/abs/2601.05975 |
| License | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ |
Abstract
We propose DeePM (Deep Portfolio Manager), a structured deep-learning macro portfolio manager trained end-to-end to maximize a robust, risk-adjusted utility. DeePM addresses three fundamental challenges in financial learning: (1) it resolves the asynchronous "ragged filtration" problem via a Directed Delay (Causal Sieve) mechanism that prioritizes causal impulse-response learning over information freshness; (2) it combats low signal-to-noise ratios via a Macroeconomic Graph Prior, regularizing cross-asset dependence according to economic first principles; and (3) it optimizes a distributionally robust objective where a smooth worst-window penalty serves as a differentiable proxy for Entropic Value-at-Risk (EVaR) - a window-robust utility encouraging strong performance in the most adverse historical subperiods. In large-scale backtests from 2010-2025 on 50 diversified futures with highly realistic transaction costs, DeePM attains net risk-adjusted returns that are roughly twice those of classical trend-following strategies and passive benchmarks, solely using daily closing prices. Furthermore, DeePM improves upon the state-of-the-art Momentum Transformer architecture by roughly fifty percent. The model demonstrates structural resilience across the 2010s "CTA (Commodity Trading Advisor) Winter" and the post-2020 volatility regime shift, maintaining consistent performance through the pandemic, inflation shocks, and the subsequent higher-for-longer environment. Ablation studies confirm that strictly lagged cross-sectional attention, graph prior, principled treatment of transaction costs, and robust minimax optimization are the primary drivers of this generalization capability.
{
"annotation_id": "4834bb96-d1f0-41c2-9d48-52417e9777f1",
"date_created": "2026-02-17T05:53:04.961000Z",
"date_modified": "2026-02-17T05:53:04.961000Z",
"file_hash": "42eead976a3d016c8f145f7932cf858934938fdff3442b1397e046087583d50d",
"private": false,
"record": {
"abstract": "We propose DeePM (Deep Portfolio Manager), a structured deep-learning macro portfolio manager trained end-to-end to maximize a robust, risk-adjusted utility. DeePM addresses three fundamental challenges in financial learning: (1) it resolves the asynchronous \"ragged filtration\" problem via a Directed Delay (Causal Sieve) mechanism that prioritizes causal impulse-response learning over information freshness; (2) it combats low signal-to-noise ratios via a Macroeconomic Graph Prior, regularizing cross-asset dependence according to economic first principles; and (3) it optimizes a distributionally robust objective where a smooth worst-window penalty serves as a differentiable proxy for Entropic Value-at-Risk (EVaR) - a window-robust utility encouraging strong performance in the most adverse historical subperiods. In large-scale backtests from 2010-2025 on 50 diversified futures with highly realistic transaction costs, DeePM attains net risk-adjusted returns that are roughly twice those of classical trend-following strategies and passive benchmarks, solely using daily closing prices. Furthermore, DeePM improves upon the state-of-the-art Momentum Transformer architecture by roughly fifty percent. The model demonstrates structural resilience across the 2010s \"CTA (Commodity Trading Advisor) Winter\" and the post-2020 volatility regime shift, maintaining consistent performance through the pandemic, inflation shocks, and the subsequent higher-for-longer environment. Ablation studies confirm that strictly lagged cross-sectional attention, graph prior, principled treatment of transaction costs, and robust minimax optimization are the primary drivers of this generalization capability.",
"arxiv_id": "2601.05975",
"authors": [
"Kieran Wood",
"Stephen J. Roberts",
"Stefan Zohren"
],
"categories": [
"q-fin.TR",
"cs.LG",
"stat.ML"
],
"license": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
"title": "DeePM: Regime-Robust Deep Learning for Systematic Macro Portfolio Management",
"url": "https://arxiv.org/abs/2601.05975",
"version": "v1"
},
"schema_id": "dorsal/arxiv",
"source": {
"execution_id": "e4617d05-05e6-4b7f-9d6c-df8731105212",
"id": "arXiv Dataset",
"type": "Model",
"variant": "snapshot-2026-01-17",
"version": "0.1.0"
},
"user_id": 1000002
}