[1903.00742] Autocurricula and the Emergence of Innovation from Social Interaction: A Manifesto for Multi-Agent Intelligence Research

Abstract:Evolution has produced a multi-scale mosaic of interacting adaptive units. Innovations arise when perturbations push parts of the system away from stable equilibria into new regimes where previously well-adapted solutions no longer work. Here we explore the hypothesis that multi-agent systems sometimes display intrinsic dynamics arising from competition and cooperation that provide a naturally emergent curriculum, which we term an autocurriculum. The solution of one social task often begets new social tasks, continually generating novel challenges, and thereby promoting innovation. Under certain conditions these challenges may become increasingly complex over time, demanding that agents accumulate ever more innovations.

March 2nd, 2019 Autocurricula and the Emergence of Innovation from Social Interaction: A Manifesto for Multi-Agent Intelligence Research Joel Z. Leibo1 , Edward Hughes1 , Marc Lanctot1 and Thore Graepel1 1 DeepMind Evolution has produced a multi-scale mosaic of interacting adaptive units. Innovations…

related reading

• Séb Krier - Marginal REVOLUTIONmarginalrevolution.com
• s9909.pdfassets.press.princeton.edu
• Harness Engineering for Self-Improvement | Lil'Loglilianweng.github.io
• Patterns and problems in multiagent systemsanthropic.com
• Towards self-driving codebases · Cursorcursor.com
• [2411.00114] Project Sid: Many-agent simulations toward AI civilizationarxiv.org
• The social AI hypothesissubstack.com
• Evolution as Backstop for Reinforcement Learning · Gwern.netgwern.net
• Towards a scale-free theory of intelligent agencymindthefuture.info
• PhD_thesis_Shirley_Wu_final.pdfcs.stanford.edu
• Multi-agent safety — AI Alignment Forumalignmentforum.org
• Of Swarms and Sand Godsblog.cosmos-institute.org