AIQF Daily digest — 7 new papers

Quant Paper Radar · 5 of 7

1. 1

   MLP-Initialized Reinforcement Learning for Exit Timing in SPX Short-Straddle Portfolios

   Jianning Wang, Robert Ślepaczuk, Paweł Sakowski · ARXIV · Options exit timing / reinforcement learning

   quant relevance 1.00

   The paper tests supervised-learning and reinforcement-learning policies for closing SPX short straddles and reports out-of-sample trading performance. Its intraday option-quote and execution requirements exceed the platform's data coverage.

   original paper

2. 2

   GCN-FinBERT-LSTM: a graph convolutional deep fusion model for multimodal stock market prediction

   PinYu Chen, Roberto Corizzo · OPENALEX · Multimodal ML equity prediction

   quant relevance 0.96

   The paper uses stock prices, technical indicators, inter-stock correlations, and news headlines to predict next-day stock trends. Its portfolio trading simulations make the proposed signal directly relevant to investing.

   original paper

3. 3

   Building crypto portfolios with agentic AI

   Paolo Giudici, Alessandro Piergallini, Antonino Castelli · OPENALEX · Crypto portfolio optimization

   quant relevance 0.95

   The paper compares static and rolling cryptocurrency portfolio optimization using daily returns and reports better out-of-sample risk-adjusted performance for the rolling approach. Agents orchestrate the workflow, while deterministic optimizers, not LLMs, calculate portfolio weights.

   original paper

The rest of the batch

• Market Impact Estimation via Causal Foundation Models

  Dennis Thumm, Ying Chen · ARXIV · relevance 0.94

  The paper estimates the causal price impact of large orders, a quantity directly relevant to execution-cost modeling and trade scheduling.

• FOMO: FLEX Options and Market Outcomes

  Shivesh Prakash, Frank Ferstler, Nicholas Pezolano · ARXIV · relevance 0.92

  The paper tests whether single-name FLEX option events identify opportunities to sell listed-option volatility through short straddles.

and 2 more · see the full radar