“I'm not a planning guy but I work in a closely related community so I'm a least somewhat familar with the area. Looking at the first paper (https://arxiv.org/pdf/1707.06170.pdf), it seems surprisingly shallow and light on details. So they have a learning system for continuous planning. So what? The AI Planning community has been doing this for ages with …”
Learning model-based planning from scratch
Conventional wisdom holds that model-based planning is a powerful approach to sequential decision-making. It is often very challenging in practice, however, because while a model can be used to evaluate a plan, it does not prescribe how to construct a plan. Here we introduce the "Imagination-based Planner", the first model-based, sequential decision-making agent that can learn to construct, evaluate, and execute plans. Before any action, it can perform a variable number of imagination steps, which involve proposing an imagined action and evaluating it with its model-based imagination. All imagined actions and outcomes are aggregated, iteratively, into a "plan context" which conditions future real and imagined actions. The agent can even decide how to imagine: testing out alternative imagined actions, chaining sequences of actions together, or building a more complex "imagination tree" by navigating flexibly among the previously imagined states using a learned policy. And our agent can learn to plan economically, jointly optimizing for external rewards and computational costs associated with using its imagination. We show that our architecture can learn to solve a challenging continuous control problem, and also learn elaborate planning strategies in a discrete maze-solving task. Our work opens a new direction toward learning the components of a model-based planning system and how to use them.
候选层:至少 2 个不同的人、在 2 个不同月份提到过,还没到主书架的 3 人 3 个月。没有做立场标注,所以只写“被提及”。
HN 书评
推荐在前,同类按时间 · 每篇最多保留 20 条,统计用全部提及“I think it is doable in under 5 years, but this critically depends on the resources invested by DM and other DL orgs. Deep RL is hugely demanding of computational resources to iterate your designs - for example, the first AlphaGo took something like 3 GPU-years to train it once (2 or 3 months parallelized); however, with much more iteration, DM was able t…”