Search Results for author:
Found 4 papers, 1 papers with code
Advancing Direct Convolution using Convolution Slicing Optimization and ISA Extensions
The speed-up over an Im2Col + BLAS method based on current BLAS implementations for end-to-end machine-learning model inference is in the range of 9% - 25% for Intel x86 and 10% - 42% for IBM POWER10 architectures.
A general framework for multi-step ahead adaptive conformal heteroscedastic time series forecasting
This paper introduces a novel model-agnostic algorithm called adaptive ensemble batch multi-input multi-output conformalized quantile regression (AEnbMIMOCQR} that enables forecasters to generate multi-step ahead prediction intervals for a fixed pre-specified miscoverage rate in a distribution-free manner.
Improved conformalized quantile regression
To show that this improved version outperforms the classic version of conformalized quantile regression and is more adaptive to heteroscedasticity, we extensively compare the prediction intervals of both in open datasets.
Enabling Massive Deep Neural Networks with the GraphBLAS
The performance of the GraphBLAS implementation is measured relative to a standard dense linear algebra library implementation.
