Reinforcement Learning: From Algorithms To Foundation Models

arXiv 2026 36.5 survey

TLDR

This thesis studies reinforcement learning from algorithms in games to foundation models, including generative and interactive world models.

Reasoning

The paper covers both multi-agent RL in games and RL with generative/foundation models, including world models. Strengths include direct mention of world models and interactive video world models. Weaknesses: limited explicit real-world evaluation and unclear experimental scope from abstract.

Read-first score

Read-first score 36.5, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 48.

Recency 5%

100

Uses a gentle age decay so recent papers surface without erasing older foundations. 2026

Topical relevance 25%

68.6

Uses existing LLM keyword relevance scores normalized to 0-100. world model,world simulator,generative world model,interactive world model,video world model,world dynamics prediction,model-based reinforcement learning world model

Methodology quality 15%

Screens visible abstract and analysis fields for experiment, dataset, baseline, metric, and limitation evidence. markers=experiment,result

Reproducibility 15%

Screens links and visible text for paper, code, dataset, artifact, and repository signals. pdf=True; code=False; dataset=False; markers=none

Graph prestige 15%

5.7

PageRank-style prestige over the local citation graph built from available OpenAlex references. local_pagerank=0.0016

Citation impact 15%

Uses OpenAlex-shaped citation metadata as a bibliometric attention signal, separate from paper quality. cited_by_count=0

Citation velocity 10%

Citation velocity estimates citations per publication-year to reduce old-paper bias. velocity=0.00

Field roles

Frontier

Rank sensitivity

Stability: volatile; rank range: 416.

Keyword Scores

world model

interactive world model

video world model

generative world model

model-based reinforcement learning world model

world dynamics prediction

world simulator

Deep Analysis

Innovations

Multi-agent RL algorithms for two-player zero-sum, large-scale video games, and multi-player general-sum settings
Diffusion-based world models as structured priors for planning, control, and policy optimization
Generative models as policy classes for decision making
Interactive video world models where actions shape future observations
Memory-augmented architectures for long-horizon modeling
Unified view of RL as objective-driven adaptation connecting decision making, environment modeling, and foundation-model capabilities

Methodology

The thesis develops algorithms for multi-agent RL in competitive and general-sum games, and integrates pretrained generative models and learned world models as representation tools and structured priors. It investigates diffusion-based world models, RL for efficient video generation, generative models as policies, and interactive video world models with memory for long-horizon tasks.

Key Results

No specific quantitative experimental results are reported in the abstract; the thesis presents a unified perspective on RL as objective-driven adaptation across strategic games and generative foundation models.

Resources

Paper