Why do amorphous materials flow so slowly when temperature or density is mildly changed? Competing explanations typically fall into two pictures, where the dynamics is either controlled by the static ...

Abstract: We present a tree-based reparameterization (TRP) framework that provides a new conceptual view of a large class of algorithms for computing approximate marginals in graphs with cycles. This ...

Efficient convolutional neural network (CNN) architecture designs have attracted growing research interests. However, they usually apply single receptive field (RF), small asymmetric RFs, or pyramid ...

Heavy Quark Effective Theory splits a heavy quark momentum into a large fixed momentum and a variable residual momentum, p = m_Q v + k. It thereby suffers a redundancy of description corresponding to ...

Adaptive Rounding has emerged as an alternative to round-to-nearest (RTN) for post-training quantization by enabling cross-element error cancellation. Yet, dense and ...

Deploying large language models (LLMs) on resource-constrained devices presents significant challenges due to their extensive parameters and reliance on dense multiplication operations. This results ...