Scaling Laws for Pre-training Agents and World Models
This work provides insights for optimizing model and data sizing in embodied AI, though it is incremental in extending scaling laws to new domains.
The paper investigates how scaling model parameters, dataset size, and compute affects performance in pre-training agents and world models, showing that power laws similar to those in language modeling emerge in world modeling and imitation learning, with coefficients influenced by tokenizer, task, and architecture.
The performance of embodied agents has been shown to improve by increasing model parameters, dataset size, and compute. This has been demonstrated in domains from robotics to video games, when generative learning objectives on offline datasets (pre-training) are used to model an agent's behavior (imitation learning) or their environment (world modeling). This paper characterizes the role of scale in these tasks more precisely. Going beyond the simple intuition that `bigger is better', we show that the same types of power laws found in language modeling also arise in world modeling and imitation learning (e.g. between loss and optimal model size). However, the coefficients of these laws are heavily influenced by the tokenizer, task \& architecture -- this has important implications on the optimal sizing of models and data.