cs.LGcs.AIcs.CLcs.CVEmpirical
Introduces Looped World Models, a looped architecture for world modelling that iteratively refines latent environment states for up to 100x parameter efficiency.
First looped architectures for world modelling with adaptive computation that scales depth.
Before reading this…
Applications
To understand this paper, make sure you know these concepts first:
Current world models face a fundamental tension: faithful long-horizon simulation demands deep computation, but deeper models are expensive to deploy and prone to compounding errors. We resolve this by introducing Looped World Models (LoopWM), which are the first looped architectures for world modelling. Our method iteratively refines latent environment states through a parameter-shared transformer block. This yield up to 100x parameter efficiency over conventional approaches with adaptive computation that automatically scales depth to match the complexity of each prediction step. Orthogonal to scaling model size and training data, LoopWM establishes iterative latent depth as a new scaling axis for world simulation, which might significantly push the community forward.
Fixed-Point Reasoners: Stable and Adaptive Deep Looped Transformers
This paper proposes FPRM, a Transformer-based model using fixed-point convergenc…
Safety, Security, and Cognitive Risks in World Models
This paper surveys the risks associated with world models, proposing a unified t…
Variable-Width Transformers
This paper proposes a new architecture for transformer-based language models cal…
Efficient Encrypted Computation in Convolutional Spiking Neural Networks with TFHE
The paper introduces FHE-DiCSNN, a novel framework that uses the TFHE scheme to…
AEGIS: Scaling Long-Sequence Homomorphic Encrypted Transformer Inference via Hybrid Parallelism on M…
AEGIS is a novel system that significantly improves the scalability of running l…
Scratched Lenses, Shifted Depth: Passive Camera-Side Optical Attacks
This paper identifies and explores a new type of physical adversarial attack on…
When Safe Models Merge into Danger: Exploiting Latent Vulnerabilities in LLM Fusion
The paper introduces TrojanMerge, a framework demonstrating that model merging c…
Evaluating Privilege Usage of Agents with Real-World Tools
The paper introduces GrantBox, a new security sandbox that evaluates how well LL…