Autocurricula and the Emergence of Innovation from Social Interaction: A Manifesto for Multi-Agent Intelligence Research
arXiv:1903.00742
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.
arXiv comment16 pages, 2 figures
References in corpus (9)
- Human-level performance in first-person multiplayer games with population-based deep reinforcement learning
- IMPALA: Scalable Distributed Deep-RL with Importance Weighted Actor-Learner Architectures
- Deep Neuroevolution: Genetic Algorithms Are a Competitive Alternative for Training Deep Neural Networks for Reinforcement Learning
- Episodic Curiosity through Reachability
- Unsupervised Predictive Memory in a Goal-Directed Agent
- Reinforcement Learning Neural Turing Machines - Revised
- Inequity aversion improves cooperation in intertemporal social dilemmas
- Psychlab: A Psychology Laboratory for Deep Reinforcement Learning Agents
- Multi-task Deep Reinforcement Learning with PopArt
Cited by in corpus (22)
- A Survey and Critique of Multiagent Deep Reinforcement Learning
- Emergent Tool Use From Multi-Agent Autocurricula
- Machine Culture
- Creative Problem Solving in Artificially Intelligent Agents: A Survey and Framework
- BADGER: Learning to (Learn [Learning Algorithms] through Multi-Agent Communication)
- Open Problems in Cooperative AI
- Exploration with Unreliable Intrinsic Reward in Multi-Agent Reinforcement Learning
- Variational Automatic Curriculum Learning for Sparse-Reward Cooperative Multi-Agent Problems
- Diverse Auto-Curriculum is Critical for Successful Real-World Multiagent Learning Systems
- Options as responses: Grounding behavioural hierarchies in multi-agent RL
- Arena: A General Evaluation Platform and Building Toolkit for Multi-Agent Intelligence
- Emergent Road Rules In Multi-Agent Driving Environments
- UneVEn: Universal Value Exploration for Multi-Agent Reinforcement Learning
- Grounding Artificial Intelligence in the Origins of Human Behavior
- Adversarial Environment Generation for Learning to Navigate the Web
- A game-theoretic analysis of networked system control for common-pool resource management using multi-agent reinforcement learning
- Learning in Matrix Games can be Arbitrarily Complex
- Neural Auto-Curricula
- Explore and Control with Adversarial Surprise
- Behaviour-conditioned policies for cooperative reinforcement learning tasks
- Individual and Collective Autonomous Development
- Meta-control of social learning strategies