MIND
2026 Datasets
Dataset Analysis
MIND is a benchmark for evaluating memory consistency and action control in world models using 250 high-quality videos across diverse scenes and action spaces.
Provenance
Collected from papers.
Derived from paper: MIND: Benchmarking Memory Consistency and Action Control in World Models
Related papers
ARB4WM: An Adversarial Robustness Benchmark for World Models in Continuous ControlAligning Agentic World Models via Knowledgeable Experience LearningDreaming Of Others: Latent Teammate Modeling In World Models For Multi-Agent Reinforcement LearningDreaming Smoothly and Sample Efficiently with Gradient Penalized Latent DynamicsFrom Observations to Events: Event-Aware World Model for Reinforcement LearningHaM-World: Soft-Hamiltonian World Models with Selective Memory for PlanningMIND: Benchmarking Memory Consistency and Action Control in World ModelsMind Dreamer: Untethering Imagination via Active Causal Intervention on Latent ManifoldsOut of Sight but Not Out of Mind: Hybrid Memory for Dynamic Video World ModelsPhysics-IQ VerifiedPhysicsMind: Sim and Real Mechanics Benchmarking for Physical Reasoning and Prediction in Foundational VLMs and World ModelsProbing the effectiveness of World Models for Spatial Reasoning through Test-time ScalingR2-Dreamer: Redundancy-Reduced World Models without Decoders or AugmentationResWM: Residual-Action World Model for Visual RLTeaching Video Generators to Remember: Eliciting Dynamic Memory for Out-of-Sight State EvolutionThinking with Imagination: Agentic Visual Spatial Reasoning with World SimulatorsWorld2VLM: Distilling World Model Imagination into VLMs for Dynamic Spatial ReasoningDMWM: Dual-Mind World Model with Long-Term ImaginationDreamerV3-XP: Optimizing exploration through uncertainty estimationDual-Mind World Models: A General Framework for Learning in Dynamic Wireless NetworksKAN-Dreamer: Benchmarking Kolmogorov-Arnold Networks as Function Approximators in World ModelsLarge Emotional World ModelLatent Action World Models for Control with Unlabeled TrajectoriesMinD: Unified Visual Imagination and Control via Hierarchical World ModelsMindDrive: An All-in-One Framework Bridging World Models and Vision-Language Model for End-to-End Autonomous DrivingMindJourney: Test-Time Scaling with World Models for Spatial ReasoningDreamSmooth: Improving Model-based Reinforcement Learning via Reward SmoothingIs Your LLM Secretly a World Model of the Internet? Model-Based Planning for Web AgentsWeb Agents with World Models: Learning and Leveraging Environment Dynamics in Web NavigationDreamerPro: Reconstruction-Free Model-Based Reinforcement Learning with Prototypical RepresentationsOrca: The World is in Your MindScaling Agent Learning via Experience Synthesis