arXivDaily arXiv每日学术速递 周一至周五更新
北美退伍军人节(Veterans Day),12 日无更新
arXiv 2110.02450cs.AI
收藏

奖惩对称的通用智能

Reward-Punishment Symmetric Universal Intelligence

The U.S. Securities and Exchange Commission(美国证券交易委员会);DeepMind;AMU

Samuel Allen Alexander, Marcus Hutter

AI总结:

该研究扩展Legg-Hutter智能体-环境框架纳入惩罚机制,证明满足特定柯尔莫哥洛夫复杂度对称性的通用图灵机下,Legg-Hutter智能测度关于原点对称,解释了负智能的存在性。

摘要:

智能体的智能水平可以为负吗?我们对Legg-Hutter智能体-环境框架进行扩展,纳入惩罚机制,并论证该问题的答案是肯定的。我们证明,若背景编码与通用图灵机(UTM)满足特定的柯尔莫哥洛夫复杂度对称性,则所得的Legg-Hutter智能测度关于原点对称。这尤其意味着,在此类UTM下,忽略奖励的智能体其Legg-Hutter智能值为0。

英文摘要:

Can an agent's intelligence level be negative? We extend the Legg-Hutter agent-environment framework to include punishments and argue for an affirmative answer to that question. We show that if the background encodings and Universal Turing Machine (UTM) admit certain Kolmogorov complexity symmetries, then the resulting Legg-Hutter intelligence measure is symmetric about the origin. In particular, this implies reward-ignoring agents have Legg-Hutter intelligence 0 according to such UTMs.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑