Minkey Chang releases new paper on neural filtering for statistical arbitrage.
Get up to speed
Minkey Chang has released a new paper titled Amortized Neural Filtering for Calibrated Posteriors in Statistical Arbitrage, proposing the Neural Filter Estimator (NFE). The paper has been accepted for oral presentation at the 7th ACM International Conference on AI in Finance (ICAIF '26). NFE is trained on simulated paths of mean-reverting processes, and a single forward pass outputs the posterior distribution of the implied state for each spread in the cross-section, incorporating estimation uncertainty into trading decisions. On simulated paths with known ground truth, the method achieves accuracy comparable to other practical estimators, with its core advantage being the ability to correctly characterize its own uncertainty. In backtests on the Korean stock market, after deducting costs, traditional point estimation methods overpredicted returns by approximately 1000 basis points per holding period relative to actual realized returns; this amortized method reduces this prediction bias by roughly 2 to 8 times.
AI-generated from reports · 3 hr agoUpdated
Coverage timeline
Follow the coverage to understand different sides of the event.
- arXiv q-finAmortized Neural Filtering for Calibrating Posteriors in Statistical Arbitrage
The paper proposes a Neural Filter Estimator (NFE), trained on simulated paths of mean-reverting processes, that outputs the posterior distribution of the implied state for each spread in the cross-section in a single forward pass, incorporating estimation uncertainty into trading decisions. On simulated paths with known ground truth, the method achieves accuracy comparable to other practical estimators, with its advantage lying in correctly characterizing its own uncertainty.
Trend of this event
Not enough consecutive observations yet; trend is not drawn.