UTICA: Multi-Objective Self-Distillation Foundation Model Pretraining for Time Series Classification

Layer by layer, module by module: Choose both for optimal OOD probing of ViT

Closing the Zero-Shot Gap in Time Series Classification with Synthetic Data and Test-Time Strategies

AIF-GEN: Open-Source Platform and Synthetic Dataset Suite for Reinforcement Learning on Large Language Models

Emergent Reasoning via Recursive Latent Reinforcement Pretraining

Shared Gradient Discovery and Superposition: Learning Dynamics of Generalization in LLMs

μLO: Compute-Efficient Meta-Generalization of Learned Optimizers

CauKer: Classification Time Series Foundation Models Can Be Pretrained On Synthetic Data

The Window Theory: Physics as the topology of causality