JaxMARL-BC accelerated simulator for multi-agent reinforcement learning macroeconomic modeling released
Get up to speed
Daniele Montagnani, Andrea Coletta, Antonietta Mira, and Aldo Glielmo released and open-sourced JaxMARL-BC, a JAX-accelerated simulator for multi-agent reinforcement learning macroeconomic modeling, used to train multi-agent reinforcement learning business cycle models. The simulator trains 75 to 724 times faster than reference implementations and can scale to 200,000 independently learning households; the authors verified its correctness under known DSGE limits and characterized scaling behavior through simple cost laws. The authors completed a proof-of-concept calibration of an extended model with heterogeneous capital returns in a two-hour experiment, whereas the equivalent experiment previously required more than a month.
AI-generated from reports · 1 hr agoUpdated
Coverage timeline
Follow the coverage to understand different sides of the event.
- arXiv q-finJaxMARL-BC: A Multi-Agent Reinforcement Learning Macroeconomic Simulator Trained in Seconds
This paper releases the open-source JAX-accelerated simulator JaxMARL-BC for training multi-agent reinforcement learning business cycle models, with training speeds 75 to 724 times faster than reference implementations, scalable to 200,000 independent learning households. The authors validate the simulator's correctness under known DSGE limits, characterize its scaling properties through a simple cost law, and complete a proof-of-concept calibration of an extended model with heterogeneous capital returns in a two-hour experiment, whereas the equivalent experiment previously required over a month.
Trend of this event
Not enough consecutive observations yet; trend is not drawn.