InvestESG
收藏资源简介:
InvestESG是一个用于研究气候投资作为社会困境的多智能体强化学习基准数据集。该数据集由华盛顿大学的研究人员创建,旨在模拟公司在环境、社会和治理(ESG)披露要求下的气候投资行为。数据集包含公司和投资者两类智能体,通过模拟复杂的交互过程,评估ESG披露政策对公司气候投资的影响。数据集的创建过程结合了PyTorch和JAX框架,支持高效的并行计算。该数据集的应用领域主要集中在气候变化政策的研究和设计,旨在通过模拟实验提供对大规模社会经济挑战的洞察。
InvestESG is a multi-agent reinforcement learning benchmark dataset for studying climate investment as a social dilemma. Developed by researchers at the University of Washington, it aims to simulate corporate climate investment behaviors under environmental, social, and governance (ESG) disclosure requirements. The dataset includes two types of AI Agents: firms and investors, and evaluates the impact of ESG disclosure policies on corporate climate investment by simulating complex interactive processes. Built using both PyTorch and JAX frameworks, it supports efficient parallel computing. Its application scenarios primarily focus on the research and design of climate change policies, with the goal of providing insights into large-scale socio-economic challenges through simulation experiments.

- 1InvestESG: A multi-agent reinforcement learning benchmark for studying climate investment as a social dilemma华盛顿大学 · 2024年



