MuZero matches or exceeds AlphaZero-level performance in Go, Chess, Shogi and sets a new state of the art on 57 Atari games by learning a model that directly supports planning rather than reconstructing full environment dynamics.
hub
Model- based reinforcement learning for atari
17 Pith papers cite this work. Polarity classification is still indexing.
17
Pith papers citing it
hub tools
citation-role summary
background 4
citation-polarity summary
roles
background 4polarities
background 4representative citing papers
JEDI is the first online end-to-end latent diffusion world model that trains latents from denoising loss rather than reconstruction, achieving competitive Atari100k results with 43% less VRAM and over 3x faster sampling than pixel diffusion baselines.
Synthetic data augmentation helps channel-mixing time series models but degrades channel-independent ones, with reliable gains only from seasonal-trend generators and gradual schedules in low-resource settings.
NEO is a probabilistic neural model that induces compositional programs as a learned Language of Thought from non-textual observations and executes them via a shared transition model to enable explanation-driven generalization.
World models succeed when their latent states are built to meet task-specific sufficiency constraints rather than preserving the maximum amount of information.
Advantage-guided diffusion (SAG and EAG) steers sampling in diffusion world models to higher-advantage trajectories, enabling policy improvement and better sample efficiency on MuJoCo tasks.
DreamerV3 uses world models and robustness techniques to solve over 150 tasks across domains with a single configuration, including Minecraft diamond collection from scratch.
DreamerV2 reaches human-level performance on 55 Atari games by learning behaviors inside a separately trained discrete-latent world model.
Dreamer learns to control from images by imagining and optimizing behaviors in a learned latent world model, outperforming prior methods on 20 visual tasks in data efficiency and final performance.
POPLIN combines policy networks with model-predictive planning by optimizing either action sequences or policy parameters, yielding 3x better sample efficiency than PETS, TD3 and SAC on MuJoCo locomotion tasks.
FlowWM trains a flow-matching model in frozen DINOv3 feature space, using a one-step projection for temporal and task-driven losses, improving stochastic future prediction on a Waymo-derived benchmark.
Odysseus adapts PPO with a turn-level critic and leverages pretrained VLM action priors to train agents achieving at least 3x average game progress over frontier models in long-horizon Super Mario Land.
A zero-shot visual world model trained on one child's experience achieves broad competence on physical understanding benchmarks while matching developmental behavioral patterns.
A two-stage framework learns a world graph of pivotal states task-agnostically via joint training of a latent model and curiosity-driven policy, then uses the graph to accelerate hierarchical RL on maze tasks.
AIDA augments scarce target data for sim-to-real visual RL by adaptively truncating unreliable imagined rollouts via a distribution-shift-aware discriminator and applying self-consistency loss on reliable state reconstructions.
A literature review of safe RL using Lyapunov and barrier functions that identifies a shift to model-free methods since 2017, well-defined open problems per approach class, and high-dimensional scalability as the main barrier.
Offline RL promises to extract high-utility policies from static datasets but faces fundamental challenges that current methods only partially address.
citing papers explorer
-
Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model
MuZero matches or exceeds AlphaZero-level performance in Go, Chess, Shogi and sets a new state of the art on 57 Atari games by learning a model that directly supports planning rather than reconstructing full environment dynamics.
-
JEDI: Joint Embedding Diffusion World Model for Online Model-Based Reinforcement Learning
JEDI is the first online end-to-end latent diffusion world model that trains latents from denoising loss rather than reconstruction, achieving competitive Atari100k results with 43% less VRAM and over 3x faster sampling than pixel diffusion baselines.
-
Does Synthetic Data Help? Empirical Evidence from Deep Learning Time Series Forecasters
Synthetic data augmentation helps channel-mixing time series models but degrades channel-independent ones, with reliable gains only from seasonal-trend generators and gradual schedules in low-resource settings.
-
Learning to Theorize the World from Observation
NEO is a probabilistic neural model that induces compositional programs as a learned Language of Thought from non-textual observations and executes them via a shared transition model to enable explanation-driven generalization.
-
Latent State Design for World Models under Sufficiency Constraints
World models succeed when their latent states are built to meet task-specific sufficiency constraints rather than preserving the maximum amount of information.
-
Advantage-Guided Diffusion for Model-Based Reinforcement Learning
Advantage-guided diffusion (SAG and EAG) steers sampling in diffusion world models to higher-advantage trajectories, enabling policy improvement and better sample efficiency on MuJoCo tasks.
-
Mastering Diverse Domains through World Models
DreamerV3 uses world models and robustness techniques to solve over 150 tasks across domains with a single configuration, including Minecraft diamond collection from scratch.
-
Mastering Atari with Discrete World Models
DreamerV2 reaches human-level performance on 55 Atari games by learning behaviors inside a separately trained discrete-latent world model.
-
Dream to Control: Learning Behaviors by Latent Imagination
Dreamer learns to control from images by imagining and optimizing behaviors in a learned latent world model, outperforming prior methods on 20 visual tasks in data efficiency and final performance.
-
Exploring Model-based Planning with Policy Networks
POPLIN combines policy networks with model-predictive planning by optimizing either action sequences or policy parameters, yielding 3x better sample efficiency than PETS, TD3 and SAC on MuJoCo locomotion tasks.
-
Flow Matching in Feature Space for Stochastic World Modeling
FlowWM trains a flow-matching model in frozen DINOv3 feature space, using a one-step projection for temporal and task-driven losses, improving stochastic future prediction on a Waymo-derived benchmark.
-
Odysseus: Scaling VLMs to 100+ Turn Decision-Making in Games via Reinforcement Learning
Odysseus adapts PPO with a turn-level critic and leverages pretrained VLM action priors to train agents achieving at least 3x average game progress over frontier models in long-horizon Super Mario Land.
-
Zero-shot World Models Are Developmentally Efficient Learners
A zero-shot visual world model trained on one child's experience achieves broad competence on physical understanding benchmarks while matching developmental behavioral patterns.
-
Learning World Graphs to Accelerate Hierarchical Reinforcement Learning
A two-stage framework learns a world graph of pivotal states task-agnostically via joint training of a latent model and curiosity-driven policy, then uses the graph to accelerate hierarchical RL on maze tasks.
-
Domain Adaptation with Adaptive Imagination for Visual Reinforcement Learning under Limited Target Data
AIDA augments scarce target data for sim-to-real visual RL by adaptively truncating unreliable imagined rollouts via a distribution-shift-aware discriminator and applying self-consistency loss on reliable state reconstructions.
-
A Review On Safe Reinforcement Learning Using Lyapunov and Barrier Functions
A literature review of safe RL using Lyapunov and barrier functions that identifies a shift to model-free methods since 2017, well-defined open problems per approach class, and high-dimensional scalability as the main barrier.
-
Offline Reinforcement Learning: Tutorial, Review, and Perspectives on Open Problems
Offline RL promises to extract high-utility policies from static datasets but faces fundamental challenges that current methods only partially address.