BetaZero: Belief-State Planning for Long-Horizon POMDPs using Learned Approximations
This work addresses planning challenges in stochastic, partially observable domains like autonomous driving and resource exploration, offering a novel method that improves performance over existing solvers, though it builds incrementally on prior techniques from fully observable domains.
The authors tackled the problem of solving high-dimensional partially observable Markov decision processes (POMDPs) for long-horizon tasks by proposing BetaZero, a belief-state planning algorithm that combines online Monte Carlo tree search with offline neural network approximations. Experiments demonstrated that BetaZero outperforms state-of-the-art POMDP solvers on various benchmarks and a real-world critical mineral exploration problem.
Real-world planning problems, including autonomous driving and sustainable energy applications like carbon storage and resource exploration, have recently been modeled as partially observable Markov decision processes (POMDPs) and solved using approximate methods. To solve high-dimensional POMDPs in practice, state-of-the-art methods use online planning with problem-specific heuristics to reduce planning horizons and make the problems tractable. Algorithms that learn approximations to replace heuristics have recently found success in large-scale fully observable domains. The key insight is the combination of online Monte Carlo tree search with offline neural network approximations of the optimal policy and value function. In this work, we bring this insight to partially observable domains and propose BetaZero, a belief-state planning algorithm for high-dimensional POMDPs. BetaZero learns offline approximations that replace heuristics to enable online decision making in long-horizon problems. We address several challenges inherent in large-scale partially observable domains; namely challenges of transitioning in stochastic environments, prioritizing action branching with a limited search budget, and representing beliefs as input to the network. To formalize the use of all limited search information, we train against a novel $Q$-weighted visit counts policy. We test BetaZero on various well-established POMDP benchmarks found in the literature and a real-world problem of critical mineral exploration. Experiments show that BetaZero outperforms state-of-the-art POMDP solvers on a variety of tasks.