按年被提到
20172017 · 2 次2026
强化学习与智能体›博弈与游戏 AI·arXiv:1705.08439 · cs.AI

Thinking Fast and Slow with Deep Learning and Tree Search

Thomas Anthony, Zheng Tian, David Barber · 2017
开始阅读 ↗PDFalphaXiv 讨论
被提及 · 3 人首次 2017-12-07 · 最近 2024-03-10
3不同的人
2活跃月份
3讨论串
198被引(OpenAlex)
arXiv 摘要(候选层不写导读)

Sequential decision making problems, such as structured prediction, robotic control, and game playing, require a combination of planning policies and generalisation of those plans. In this paper, we present Expert Iteration (ExIt), a novel reinforcement learning algorithm which decomposes the problem into separate planning and generalisation tasks. Planning new policies is performed by tree search, while a deep neural network generalises those plans. Subsequently, tree search is improved by using the neural network policy to guide search, increasing the strength of new plans. In contrast, standard deep Reinforcement Learning algorithms rely on a neural network not only to generalise plans, but to discover them too. We show that ExIt outperforms REINFORCE for training a neural network to play the board game Hex, and our final tree search agent, trained tabula rasa, defeats MoHex 1.0, the most recent Olympiad Champion player to be publicly released.

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
评论sanxiyn · 2017-12

“Quick, someone apply AlphaZero algorithm to this MCTS prover for math: https://arxiv.org/abs/1611.05990 I actually prefer "expert iteration" and "tree-policy target" term instead of "AlphaZero algorithm", because they emphasize what is novel about it among the entire system. Terms are from https://arxiv.org/abs/1705.08439”

Google's AlphaZero Beats Stockfish In 100-Game Match原评论 ↗
评论nlperguiy · 2017-12

“https://arxiv.org/abs/1705.08439 The original paper. The references in the paper paint a much clearer picture of where exactly the idea behind reinforcement learning with optimal, suboptimal, random oracles comes from. There are also mathematical proofs that these setups work. I was quite shocked to not see [6, 16] references in any of the recent MCTS …”

Learning from Scratch by Thinking Fast and Slow原评论 ↗
评论fabmilo · 2024-03

“I believe this kind of graph exploration is what we need to progress reasoning in AI. Plain LLMS will fail. The link has tons of good references, including the Zobrist hashing https://en.wikipedia.org/wiki/Zobrist_hashing for game tables. We need to find a good hashing for language based state description so that graph exploration doesn't explode computat…”

Monte-Carlo graph search from first principles原评论 ↗