Instance-Conditional Timescales of Decay for Non-Stationary Learning

Slow concept drift is a ubiquitous, yet under-studied problem in practical machine learning systems. In such settings, although recent data is more indicative of future data, naively prioritizing recent instances runs the risk of losing valuable information from the past. We propose an optimization-...

Full description

Saved in:
Bibliographic Details
Published inarXiv.org
Main Authors Jain, Nishant, Shenoy, Pradeep
Format Paper
LanguageEnglish
Published Ithaca Cornell University Library, arXiv.org 20.12.2023
Subjects
Algorithms
Datasets
Machine learning
Optimization
Training
Online AccessGet full text

Cover

More Information
Summary:Slow concept drift is a ubiquitous, yet under-studied problem in practical machine learning systems. In such settings, although recent data is more indicative of future data, naively prioritizing recent instances runs the risk of losing valuable information from the past. We propose an optimization-driven approach towards balancing instance importance over large training windows. First, we model instance relevance using a mixture of multiple timescales of decay, allowing us to capture rich temporal trends. Second, we learn an auxiliary scorer model that recovers the appropriate mixture of timescales as a function of the instance itself. Finally, we propose a nested optimization objective for learning the scorer, by which it maximizes forward transfer for the learned model. Experiments on a large real-world dataset of 39M photos over a 9 year period show upto 15% relative gains in accuracy compared to other robust learning baselines. We replicate our gains on two collections of real-world datasets for non-stationary learning, and extend our work to continual learning settings where, too, we beat SOTA methods by large margins.
ISSN:2331-8422