Hailing Cheng
2 papers · Latest:
Artificial Intelligence
Learning to Rotate: Temporal and Semantic Rotary Encoding for Sequential Modeling
This paper introduces SIREN-RoPE, a novel approach that treats the rotation manifold in Rotary Positional Embeddings as a learnable, signal-conditioned space, improving sequential modeling.
2604.24717
Machine LearningScalable Hyperparameter-Divergent Ensemble Training with Automatic Learning Rate Exploration for Large Models
HDET repurposes data-parallel replicas for simultaneous, automatic learning rate exploration, improving large model training without extra cost.
2604.24708
📬 Weekly AI Paper Digest
Get the top 10 AI/ML arXiv papers from the week — summarized, scored, and delivered to your inbox every Monday.