Knowledge-driven Augmentation and Retrieval for Integrative Temporal Adaptation

TL;DR

KARITA is a novel method that captures diverse temporal shifts and integrates rich knowledge sources to improve model adaptation across various domains.

Contribution

It introduces a knowledge-driven approach for temporal augmentation and retrieval, effectively handling semantic and knowledge shifts in dynamic environments.

Findings

Consistent improvements in classification tasks across multiple domains.

Knowledge integration enhances temporal augmentation and learning.

Effective handling of semantic and feature shifts over time.

Abstract

Time introduces fundamental challenges in model development and deployment: models are usually trained on historical data while deployed on future data where semantic distributions and domain knowledge may evolve. Unfortunately, existing studies either overlook temporal shifts or hardly capture rich shifting patterns of both semantic and knowledge. We develop Knowledge-driven Augmentation and Retrieval for Integrative Temporal Adaptation (KARITA) to capture diverse temporal shifts (e.g., uncertainty and feature shift), construct and integrate rich knowledge sources (e.g., medical ontology like MeSH), and leverage shifting insights for selecting-retrieval augmented learning. We evaluate KARITA on classification tasks across multiple domains, clinical, legal, and scientific corpora, demonstrating consistent improvements across multiple domains with temporal adaptation. Our results show that knowledge integration can be more critical and effective in temporal augmentation and learning.