Coarse-Graining Recursive Expert Computation: Continuation-Relative Interfaces and Failure Modes
cs.LG, cs.AI, cs.CC, cs.CE
Submitted: 2026-10-10
Updated: 2026-10-10
Terminology
Sources
- Task-Conditioned Routing Signatures in Sparse Mixture-of-Experts Transformers
- Deep Equilibrium Models
- Identifying Functionally Important Features with End-to-End Sparse Dictionary Learning
- LoopMoE: Unifying Iterative Computation with Mixture-of-Experts for Language Modeling
- Streamlining Redundant Layers to Compress Large Language Models
- Training Verifiers to Solve Math Word Problems
- Similarity and Matching of Neural Network Representations
- Universal Transformers
- LiFT: Loop Flow Transformers
- MoEcho: Exploiting Side-Channel Attacks to Compromise User Privacy in Mixture-of-Experts LLMs
- Switch Transformers: Scaling to Trillion Parameter Models with Simple and Efficient Sparsity
- One Step Diffusion via Shortcut Models
- Mean Flows for One-step Generative Modeling
- Knowledge Distillation under Teacher Misspecification: An Order-Parameter Analysis of the Gap between Teacher Mimicry and Task Performance
- Recursive Flow Matching
- Flow Map Language Models: One-step Language Modeling via Continuous Denoising
- Leave It to the Experts: Detecting Knowledge Distillation via MoE Expert Signatures
- Consistency Deep Equilibrium Models
- Flow Matching for Generative Modeling
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: The Predictive Cost of ELBO Binding
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks