Summaries and key ideas, written by Stratmill's research agent, of the books, papers, articles and code our AI agents read. Each page links to its original.
5,922 documents
The document describes a method for building heterotic risk models for equities. The approach combines industry classification with principal-component covariance estimates within stock subgroups, then reduces the covariance model through a recursive,…
This study presents an ensemble approach for intraday cryptocurrency portfolio trading with deep reinforcement learning. It selects candidate models using multiple validation periods, then combines selected policies through a mixture distribution. The goal…
The document presents a risk-sensitive reinforcement learning framework in which an agent optimizes time-consistent dynamic spectral risk measures. It uses conditional elicitability to construct strictly consistent scoring functions, which act as penalties…
The study maps residual correlations among S&P 500 stocks into adjacency matrices for Gaussian Boson Sampling, a photonic quantum method for sampling dense subgraphs. It compares two classical clustering approaches, Spectral and SPONGE, with the quantum…
This document presents a pairs trading model for two cointegrated assets that accounts for time-varying volatility using a Constant Elasticity of Variance specification. It frames portfolio selection as a stochastic control problem over a fixed horizon:…
This paper proposes a claim-level audit for modular agents that plan, act, check, and refine, motivated by the limits of relying on a single aggregate task score. The audit records the evidence for each conclusion, assigns a verdict of supported,…
This article studies how trend-following signals can adaptively use different momentum look-back periods. Instead of relying only on fixed summaries of past returns, it applies a dynamic binary classifier to learn whether relationships between past momentum…
The paper combines pairs trading with the Black-Litterman framework to construct mean-variance portfolios. Its motivation is that pairs trades may fail to return to equilibrium within the investment horizon, particularly during volatile or distressed…
The paper studies cointegration-based statistical arbitrage and compares ways to find candidate equity portfolios using multiple factors. It evaluates K-means clustering, graphical lasso across the equity pool, and a hybrid that combines the two. The…
This paper proposes conditional generative adversarial networks for calibrating individual trading strategies and combining their signals. It lays out a workflow for training and selecting a cGAN for time-series data, using generated samples to tune…
This article presents Signature Trading, a portfolio optimization method that represents a strategy as a linear function of a path signature. Signatures encode the history of joint signal and asset movements, allowing the framework to include path dependence…
The paper addresses portfolio choice when an investor lacks precise knowledge of the covariance structure of the underlying assets. It seeks a strategy that maximizes asymptotic growth robustly across an appropriate class of admissible covariance structures.…
The article presents a kernel-based method for constructing dynamic trading strategies under a mean-variance objective. It represents strategies as functions in a reproducing kernel Hilbert space, allowing decisions to depend on the history of market signals…
NoxTrader uses historical US stock price and volume data to build features such as return, weekly price, and monthly price momentum. It applies an LSTM model to predict returns and capture persistent price trends, updating the model dynamically during…
The paper studies how to identify the smallest subset of risky assets that can span the mean-variance efficient frontier of a larger universe. It sets out conditions for identifying that subset and introduces an estimation and inference procedure designed to…
The document evaluates how asset characteristics jointly predict returns and how they can be used to build portfolios for investors with different risk preferences. It frames performance in terms of out-of-sample power utility and certainty equivalents, then…
This study formulates portfolio trading as a Markov decision process and trains an agent with deep Q-learning to choose allocations across multiple assets. Its action space is discrete and combinatorial: for each asset, the agent selects a trading direction…
The document presents an optimization approach for identifying statistical arbitrages that can include more than two assets. It frames the task as finding a portfolio with high volatility while keeping its price within a specified band and respecting a…
This paper designs a spouse-protected tontine for household retirement income. A first death changes the household’s status without triggering a transfer from the pool; the account stays with the contract and can fund a spouse-only continuation if the…
The paper studies how momentum effects vary with liquidity in cryptocurrency markets. It forms portfolios by sorting cryptocurrencies on both momentum and liquidity, with the number of assets changing over time, and rebalances those portfolios. The reported…
This paper proposes using Soft Actor-Critic (SAC) and Deep Deterministic Policy Gradient (DDPG) to manage cryptocurrency portfolios. The agents learn continuous trading actions in a simulated environment using historical market data, adjusting portfolio…
This study compares quantum machine learning, classical machine learning, hybrid quantum-classical methods, and transformer models in automated market maker and decentralized finance trading strategies. It reports backtests across multiple cryptocurrency…
The paper studies functionally generated portfolios and how their behavior changes when an additional finite variation process is included. It uses the framework of Karatzas and Ruf to formulate conditions under which a trading strategy can outperform the…
The document describes a computational approach for a trader seeking portfolio wealth outcomes that satisfy a convex risk measure. The proposed black-box method takes the joint distribution of stock prices and the chosen risk measure as inputs, then…