Sediment/DiscoveryUnder development
Back to home
CONCEPT
SUB-FIELDS
PAPERS
ExploringAI for Games
RL6 papersReinforcement Learning
MCTS3 papersSearch & Planning
PCG3 papersProcedural Content Gen
MAS4 papersMulti-Agent Systems
LLMs5 papersLarge Language Models
RL2013Human-level control (DQN)Mnih et al.

Deep Q-networks reach human-level play on Atari from pixels.

RL · MCTS2016AlphaGoSilver et al.

Policy/value networks plus tree search defeat a Go world champion.

Reinforcement LearningSearch & Planning
RL · MCTS2019MuZeroSchrittwieser et al.

Plans with a learned model, no rules given.

Reinforcement LearningSearch & Planning
PCG · RL2020PCGRLKhalifa et al.

Frames level generation as an RL control problem.

Procedural Content GenReinforcement Learning
RL · MAS2019OpenAI FiveOpenAI

Team of RL agents reaches pro level at Dota 2.

Reinforcement LearningMulti-Agent Systems
PCG2016WaveFunctionCollapseGumin

Constraint-based tile generation from a single example image.

RL · LLMs2023VoyagerWang et al.

An LLM-driven agent that writes skills to explore Minecraft.

Reinforcement LearningLarge Language Models
LLMs · MAS · MCTS2022CICERO (Diplomacy)Meta AI

Language + planning to negotiate and play Diplomacy.

Large Language ModelsMulti-Agent SystemsSearch & Planning
PCG · LLMs2023MarioGPTSudhakaran et al.

Generates Mario levels from text prompts with a language model.

Procedural Content GenLarge Language Models
LLMs · MAS2023Generative AgentsPark et al.

LLM agents with memory simulate believable social behaviour.

Large Language ModelsMulti-Agent Systems
LLMs · MAS2024SIMADeepMind

A generalist agent following language instructions across 3D games.

Large Language ModelsMulti-Agent Systems
Drag to explore · Select sub-fields to find shared papers
100%
RL
MCTS
PCG
MAS
LLMs