New framework uses LLMs to align AI portfolio optimization with ESG preferences

Researchers have developed a new framework for portfolio optimization that integrates environmental, social, and governance (ESG) factors using Multi-Objective Reinforcement Learning (MORL). This approach addresses the challenge of differing ESG rating methodologies and the difficulty of manually weighting multiple objectives by incorporating a preference elicitation system. The system infers user utility functions through pairwise comparisons of portfolios based on Sharpe ratios and ESG scores. Simulations using Large Language Model personas revealed that regional backgrounds significantly influence these preferences, with European personas prioritizing ESG and Texas personas favoring financial returns. AI

IMPACT This research could lead to more sophisticated AI-driven investment strategies that better align with diverse sustainability preferences.

RANK_REASON This is a research paper detailing a novel framework for portfolio optimization. [lever_c_demoted from research: ic=1 ai=0.7]

Read on arXiv cs.LG →

• environmental, social and corporate governance
• Europe
• Gaussian Processes
• large language model
• Multi-objective reinforcement learning
• Portfolio Managers
• reinforcement learning
• Texas

• paper
• product
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →