Knowledge-driven Augmentation and Retrieval for Integrative Temporal Adaptation
For practitioners deploying models on time-evolving data, KARITA provides a method to handle temporal shifts by leveraging domain knowledge, though the gains are incremental over existing adaptation techniques.
KARITA addresses temporal shifts in model deployment by integrating knowledge sources (e.g., medical ontology) with retrieval-augmented learning, achieving consistent improvements across clinical, legal, and scientific classification tasks.
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.