• 文献检索
  • 文档翻译
  • 深度研究
  • 学术资讯
  • Suppr Zotero 插件Zotero 插件
  • 邀请有礼
  • 套餐&价格
  • 历史记录
应用&插件
Suppr Zotero 插件Zotero 插件浏览器插件Mac 客户端Windows 客户端微信小程序
定价
高级版会员购买积分包购买API积分包
服务
文献检索文档翻译深度研究API 文档MCP 服务
关于我们
关于 Suppr公司介绍联系我们用户协议隐私条款
关注我们

Suppr 超能文献

核心技术专利:CN118964589B侵权必究
粤ICP备2023148730 号-1Suppr @ 2026

文献检索

告别复杂PubMed语法,用中文像聊天一样搜索,搜遍4000万医学文献。AI智能推荐,让科研检索更轻松。

立即免费搜索

文件翻译

保留排版,准确专业,支持PDF/Word/PPT等文件格式,支持 12+语言互译。

免费翻译文档

深度研究

AI帮你快速写综述,25分钟生成高质量综述,智能提取关键信息,辅助科研写作。

立即免费体验

智能体群体中共同概念、对称性和一致性的出现——一种信息论模型

Emergence of common concepts, symmetries and conformity in agent groups-an information-theoretic model.

作者信息

Möller Marco, Polani Daniel

机构信息

Theory of Complex Systems Group, Institute of Solid State Physics, Technical University of Darmstadt, Germany.

Adaptive Systems Research Group, Department of Computer Science, University of Hertfordshire, Hatfield, UK.

出版信息

Interface Focus. 2023 Apr 14;13(3):20230006. doi: 10.1098/rsfs.2023.0006. eCollection 2023 Jun 6.

DOI:10.1098/rsfs.2023.0006
PMID:37065261
原文链接:https://pmc.ncbi.nlm.nih.gov/articles/PMC10102731/
Abstract

The paper studies principles behind structured, especially symmetric, representations through enforced inter-agent conformity. For this, we consider agents in a simple environment who extract individual representations of this environment through an information maximization principle. The representations obtained by different agents differ in general to some extent from each other. This gives rise to ambiguities in how the environment is represented by the different agents. Using a variant of the information bottleneck principle, we extract a 'common conceptualization' of the world for this group of agents. It turns out that the common conceptualization appears to capture much higher regularities or symmetries of the environment than the individual representations. We further formalize the notion of identifying symmetries in the environment both with respect to 'extrinsic' (birds-eye) operations on the environment as well as with respect to 'intrinsic' operations, i.e. subjective operations corresponding to the reconfiguration of the agent's embodiment. Remarkably, using the latter formalism, one can re-wire an agent to conform to the highly symmetric common conceptualization to a much higher degree than an unrefined agent; and that, without having to re-optimize the agent from scratch. In other words, one can 're-educate' an agent to conform to the de-individualized 'concept' of the agent group with comparatively little effort.

摘要

本文研究了通过强制智能体之间的一致性来实现结构化(尤其是对称)表示的原理。为此,我们考虑处于简单环境中的智能体,它们通过信息最大化原理提取该环境的个体表示。不同智能体获得的表示通常在某种程度上彼此不同。这就导致了不同智能体对环境的表示方式存在模糊性。我们使用信息瓶颈原理的一个变体,为这组智能体提取出世界的“共同概念化”。结果表明,与个体表示相比,共同概念化似乎捕捉到了环境中更高的规律性或对称性。我们进一步形式化了在环境中识别对称性的概念,既涉及对环境的“外在”(鸟瞰)操作,也涉及“内在”操作,即与智能体自身形态重新配置相对应的主观操作。值得注意的是,使用后一种形式主义,与未经过优化的智能体相比,人们可以将一个智能体重连,使其在更高程度上符合高度对称的共同概念化;而且,无需从头重新优化智能体。换句话说,人们可以相对轻松地“重新训练”一个智能体,使其符合智能体群体的去个性化“概念”。

https://cdn.ncbi.nlm.nih.gov/pmc/blobs/11cc/10102731/94ba29a59571/rsfs20230006f11.jpg
https://cdn.ncbi.nlm.nih.gov/pmc/blobs/11cc/10102731/00401d2019b7/rsfs20230006f04.jpg
https://cdn.ncbi.nlm.nih.gov/pmc/blobs/11cc/10102731/fa30bf3182ad/rsfs20230006f08.jpg
https://cdn.ncbi.nlm.nih.gov/pmc/blobs/11cc/10102731/d25ae90dcf7a/rsfs20230006f10.jpg
https://cdn.ncbi.nlm.nih.gov/pmc/blobs/11cc/10102731/94ba29a59571/rsfs20230006f11.jpg
https://cdn.ncbi.nlm.nih.gov/pmc/blobs/11cc/10102731/00401d2019b7/rsfs20230006f04.jpg
https://cdn.ncbi.nlm.nih.gov/pmc/blobs/11cc/10102731/fa30bf3182ad/rsfs20230006f08.jpg
https://cdn.ncbi.nlm.nih.gov/pmc/blobs/11cc/10102731/d25ae90dcf7a/rsfs20230006f10.jpg
https://cdn.ncbi.nlm.nih.gov/pmc/blobs/11cc/10102731/94ba29a59571/rsfs20230006f11.jpg

相似文献

1
Emergence of common concepts, symmetries and conformity in agent groups-an information-theoretic model.智能体群体中共同概念、对称性和一致性的出现——一种信息论模型
Interface Focus. 2023 Apr 14;13(3):20230006. doi: 10.1098/rsfs.2023.0006. eCollection 2023 Jun 6.
2
Representation in natural and artificial agents: an embodied cognitive science perspective.自然与人工智能体中的表征:一种具身认知科学视角
Z Naturforsch C J Biosci. 1998 Jul-Aug;53(7-8):480-503. doi: 10.1515/znc-1998-7-804.
3
TEST: a tropic, embodied, and situated theory of cognition.测试:一种具身、情境化的认知理论。
Top Cogn Sci. 2014 Jul;6(3):442-60. doi: 10.1111/tops.12024. Epub 2013 Apr 24.
4
Macromolecular crowding: chemistry and physics meet biology (Ascona, Switzerland, 10-14 June 2012).大分子拥挤现象:化学与物理邂逅生物学(瑞士阿斯科纳,2012年6月10日至14日)
Phys Biol. 2013 Aug;10(4):040301. doi: 10.1088/1478-3975/10/4/040301. Epub 2013 Aug 2.
5
Representations of space and time in the maximization of information flow in the perception-action loop.感知-行动循环中信息流最大化时的空间与时间表征。
Neural Comput. 2007 Sep;19(9):2387-432. doi: 10.1162/neco.2007.19.9.2387.
6
Quantifying Reinforcement-Learning Agent's Autonomy, Reliance on Memory and Internalisation of the Environment.量化强化学习智能体的自主性、对记忆的依赖以及环境内化程度。
Entropy (Basel). 2022 Mar 13;24(3):401. doi: 10.3390/e24030401.
7
How Morphological Computation Shapes Integrated Information in Embodied Agents.形态计算如何塑造具身智能体中的整合信息。
Front Psychol. 2021 Nov 29;12:716433. doi: 10.3389/fpsyg.2021.716433. eCollection 2021.
8
The embodied nature of medical concepts: image schemas and language for PAIN.医学概念的具身本质:疼痛的意象图式与语言
Cogn Process. 2014 Aug;15(3):283-96. doi: 10.1007/s10339-013-0594-9. Epub 2014 Jan 5.
9
Accurate lattice parameters from 2D-periodic images for subsequent Bravais lattice type assignments.从二维周期图像中获取准确的晶格参数,用于后续的布拉维晶格类型分配。
Adv Struct Chem Imaging. 2018;4(1):5. doi: 10.1186/s40679-018-0051-z. Epub 2018 Mar 28.
10
Building compressed causal models of the world.
Cogn Psychol. 2024 Dec;155:101682. doi: 10.1016/j.cogpsych.2024.101682. Epub 2024 Oct 19.

引用本文的文献

1
Exact and Soft Successive Refinement of the Information Bottleneck.信息瓶颈的精确与软渐进细化
Entropy (Basel). 2023 Sep 19;25(9):1355. doi: 10.3390/e25091355.

本文引用的文献

1
Identifying regulation with adversarial surrogates.利用对抗代理识别调控。
Proc Natl Acad Sci U S A. 2023 Mar 21;120(12):e2216805120. doi: 10.1073/pnas.2216805120. Epub 2023 Mar 15.
2
Machine Learning Hidden Symmetries.机器学习中的隐藏对称性。
Phys Rev Lett. 2022 May 6;128(18):180201. doi: 10.1103/PhysRevLett.128.180201.
3
Epistemic Communities under Active Inference.主动推理下的认知共同体
Entropy (Basel). 2022 Mar 29;24(4):476. doi: 10.3390/e24040476.
4
Embodiment and Its Influence on Informational Costs of Decision Density-Atomic Actions vs. Scripted Sequences.具身性及其对决策密度信息成本的影响——原子动作与脚本序列
Front Robot AI. 2021 Apr 12;8:535158. doi: 10.3389/frobt.2021.535158. eCollection 2021.
5
Efficient compression in color naming and its evolution.颜色命名中的高效压缩及其演变。
Proc Natl Acad Sci U S A. 2018 Jul 31;115(31):7937-7942. doi: 10.1073/pnas.1800521115. Epub 2018 Jul 18.
6
Representations of space and time in the maximization of information flow in the perception-action loop.感知-行动循环中信息流最大化时的空间与时间表征。
Neural Comput. 2007 Sep;19(9):2387-432. doi: 10.1162/neco.2007.19.9.2387.
7
Color naming, unique hues, and hue cancellation predicted from singularities in reflection properties.根据反射特性中的奇点预测颜色命名、独特色调和色调消除。
Vis Neurosci. 2006 May-Aug;23(3-4):331-9. doi: 10.1017/S0952523806233182.
8
Finite state automata resulting from temporal information maximization and a temporal learning rule.
Neural Comput. 2005 Oct;17(10):2258-90. doi: 10.1162/0899766054615671.
9
Least effort and the origins of scaling in human language.省力原则与人类语言中的缩放起源。
Proc Natl Acad Sci U S A. 2003 Feb 4;100(3):788-91. doi: 10.1073/pnas.0335980100. Epub 2003 Jan 22.
10
Energy as a constraint on the coding and processing of sensory information.能量作为感觉信息编码与处理的一种限制因素。
Curr Opin Neurobiol. 2001 Aug;11(4):475-80. doi: 10.1016/s0959-4388(00)00237-3.