Unsupervised Modeling of Acquisition Variability in Connectomes
Discover a hybrid latent space model that reduces acquisition variability in structural connectomes using unsupervised deep learning techniques.
Optimizing Online Multiple Testing with Weighted Regret
Discover how Weighted Regret and Decoupled-OMT improve online multiple testing by reducing false positives and negatives in dynamic data streams.
Multi-Scale Dequant for Faster Efficient LLM Inference
Discover Multi-Scale Dequant, a novel method that eliminates dequantization bottlenecks to boost large language model inference efficiency and reduce laten...
TERMS-Bench: Advanced Evaluation of LLM Negotiation Agents
Discover TERMS-Bench, a novel framework diagnosing LLM negotiation agents beyond deal rates for deeper insights and improved AI negotiation strategies.
Adaptive Importance Sampling for Efficient Quantized RL
Discover how Adaptive Importance Sampling improves training efficiency and stability in quantized reinforcement learning for large language models.