Skip to main navigation Skip to search Skip to main content

One-shot atomic detection

  • University of British Columbia

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Feature selection in data science involves identifying the most prominent and uncorrelated features in the data, which can be useful for compression and interpretability. If these feature can be easily extracted, then a model can be trained over a reduced set of weights, which leads to more efficient training and possibly more robust classifiers. There are many approaches to feature selection; in this work, we propose screening the 'atoms' of a gradient of a loss function taken at a random point. We illustrate this approach on sparse and low-rank optimization problems. Despite the simplicity of the approach, we are often able to select the dominant features easily, and greatly improve the runtime and robustness in training overparametrized models.

Original languageEnglish
Title of host publication2019 IEEE 8th International Workshop on Computational Advances in Multi-Sensor Adaptive Processing, CAMSAP 2019 - Proceedings
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages1-5
Number of pages5
ISBN (Electronic)9781728155494
DOIs
StatePublished - Dec 2019
Event8th IEEE International Workshop on Computational Advances in Multi-Sensor Adaptive Processing, CAMSAP 2019 - Le Gosier, Guadeloupe
Duration: Dec 15 2019Dec 18 2019

Publication series

Name2019 IEEE 8th International Workshop on Computational Advances in Multi-Sensor Adaptive Processing, CAMSAP 2019 - Proceedings

Conference

Conference8th IEEE International Workshop on Computational Advances in Multi-Sensor Adaptive Processing, CAMSAP 2019
Country/TerritoryGuadeloupe
CityLe Gosier
Period12/15/1912/18/19

Fingerprint

Dive into the research topics of 'One-shot atomic detection'. Together they form a unique fingerprint.

Cite this