Efficient Identification of Approximate Best Configuration of Training in Large Datasets

Authors: Silu Huang, Chi Wang, Bolin Ding, Surajit Chaudhuri3862-3869

AAAI 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We conduct experiments with large datasets. We demonstrate that our ABC solution is tens to hundreds of times faster, while returning top configurations with no more than 1% accuracy loss.
Researcher Affiliation Collaboration 1University of Illinois, Urbana-Champaign, IL 2Microsoft Research, Redmond, WA 3Alibaba Group, Bellevue, WA shuang86@illinois.edu, {wang.chi, surajitc}@microsoft.com, bolin.ding@alibaba-inc.com
Pseudocode Yes Algorithm 1: ABC
Open Source Code No The paper does not provide a direct link to open-source code for its methodology or an explicit statement of code release.
Open Datasets Yes We evaluate with five large-scale machine learning benchmarks that are publicly available.
Dataset Splits No The paper states:
Hardware Specification Yes We conducted our evaluation on a VM with 8 cores and 56 GB RAM.
Software Dependencies No The paper mentions
Experiment Setup Yes The initial training sample size and testing sample size are 1000 and 2000 respectively. The geometry step size is set to be c = 2. ϵ = 0.01, δ = 0.5.