World Models Comparison Hub
4 papers - avg viability 5.8
Top Papers
- Causal-JEPA: Learning World Models through Object-Level Latent Interventions(9.0)
C-JEPA offers an efficient object-centric world model enhancing visual question answering and agent control with latent interventions.
- Self-Supervised Multi-Modal World Model with 4D Space-Time Embedding(8.0)
DeepEarth is a self-supervised multi-modal world model with a novel 4D space-time positional encoder, achieving state-of-the-art performance on ecological forecasting.
- The Trinity of Consistency as a Defining Principle for General World Models(4.0)
CoW-Bench provides a principled benchmark for evaluating multi-frame reasoning and generation in unified world models.
- Specification-Driven Generation and Evaluation of Discrete-Event World Models via the DEVS Formalism(2.0)
Develop discrete-event world models using DEVS formalism for enhanced reliability and flexibility in agentic systems.