LARS: Memory-Efficient Fine-Tuning for On-Device LLMs
Discover how LARS reduces memory use in fine-tuning large language models on devices with limited resources, boosting efficiency without performance loss.
Stochastic KV Routing for Efficient Transformer Caching
Discover how stochastic KV routing enables adaptive depth-wise cache sharing to optimize transformer models, reducing memory use while preserving performan...
BiTA: Advanced Temporal Graph Model for Cyber Alert Prediction
Discover BiTA, a bidirectional GRU-Transformer model enhancing temporal graph networks for accurate cyber alert prediction and threat detection.
Generative Self-Supervised Learning for PPG-Based Health Estimation
Discover a novel generative self-supervised framework improving non-invasive physiological parameter estimation using photoplethysmography (PPG) data.
KARL: Reducing LLM Hallucinations with Knowledge-Aware RL
Discover KARL, a novel reinforcement learning method that reduces hallucinations in large language models by recognizing knowledge boundaries.