English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

In several real world applications, machine learning models are deployed tomake predictions on data whose distribution changes gradually along time,leading to a drift between the train and test distributions. Such models areoften re-trained on new data periodically, and they hence need to generalize todata not too far into the future. In this context, there is much prior work onenhancing temporal generalization, e.g. continuous transportation of past data,kernel smoothed time-sensitive parameters and more recently, adversariallearning of time-invariant features. However, these methods share severallimitations, e.g, poor scalability, training instability, and dependence onunlabeled data from the future. Responding to the above limitations, we proposea simple method that starts with a model with time-sensitive parameters butregularizes its temporal complexity using a Gradient Interpolation (GI) loss.GI allows the decision boundary to change along time and can still preventoverfitting to the limited training time snapshots by allowing task-specificcontrol over changes along time. We compare our method to existing baselines onmultiple real-world datasets, which show that GI outperforms more complicatedgenerative and adversarial approaches on the one hand, and simpler gradientregularization methods on the other.

Training for the Future: A Simple Gradient Interpolation Loss to Generalize Along Time