LogSumExp
PulseAugur coverage of LogSumExp — every cluster mentioning LogSumExp across labs, papers, and developer communities, ranked by signal.
1 day(s) with sentiment data
-
New Mahalanobis-based attention mechanism boosts AI model efficiency
Researchers have introduced Mahalanobis-Based Multi-Head Attention (MHA-CSP), a novel attention mechanism that replaces the standard dot-product with a Mahalanobis distance-based RBF kernel. This approach allows for att…
-
GaugeQuant optimizes LLM quantization by learning from model symmetries
Researchers have developed GaugeQuant, a novel method for optimizing the quantization of large language models (LLMs) by leveraging their internal symmetries. This technique introduces a LogSumExp term to the training l…
-
LogSumExp Smoothing Near-Optimal for Max Function Approximation
A new paper published on arXiv presents an elementary proof that LogSumExp smoothing is nearly optimal for approximating the max function in $\mathbb{R}^d$. The research establishes a lower bound for overestimating smoo…
-
New Safe KL Divergence Improves LogSumExp Optimization
Researchers have developed a novel approximation for the LogSumExp function, which is crucial for optimization problems like entropy-regularized optimal transport and distributionally robust optimization. This new appro…