{"slug": "inference-time-prior-adaptation-for-imbalanced-data-streams-under-concept-drift", "title": "Inference-Time Prior Adaptation for Imbalanced Data Streams under Concept Drift", "summary": "Researchers posted arXiv:2610.33126v1, describing PAI-KNN, a k-nearest neighbor classifier that adapts to concept drift and class imbalance through inference-time prior adjustment instead of memory modification. PAI-KNN keeps a simple sliding-window buffer and dynamically adjusts its decision rule using class prior estimates derived from recent observations. Evaluated on multiple synthetic data streams spanning various imbalance and drift scenarios, PAI-KNN achieved competitive performance with state-of-the-art methods and performed particularly well under dynamic class imbalance and concept drift, suggesting memory-based stream classifiers need not rely exclusively on memory modification.", "body_md": "arXiv:2610.33126v1 Announce Type: new \nAbstract: Data stream classification under concept drift and class imbalance requires models to adapt continuously while maintaining strong minority-class performance. Many existing approaches rely on active memory management, in- cluding instance removal, memory restructuring, resampling, and synthetic data generation. We present PAI-KNN, a k-nearest neighbor classifier that adapts via inference-time prior adjustment rather than memory modification. PAI-KNN main- tains a simple sliding-window buffer and dynamically adjusts its decision rule using class prior estimates derived from recent observations. Evaluated on a mul- titude of synthetic data streams covering various imbalance and drift scenarios, PAI-KNN achieves competitive performance with state-of-the-art methods, per- forming particularly well under scenarios involving dynamic class imbalance and concept drift. These findings highlight the potential of inference-time adaptation in challenging non-stationary environments and suggest that adaptation in memory- based stream classifiers need not rely exclusively on memory modification.", "url": "https://wpnews.pro/news/inference-time-prior-adaptation-for-imbalanced-data-streams-under-concept-drift", "canonical_source": "https://arxiv.org/abs/2610.33126", "published_at": "2026-10-28 04:00:00+00:00", "updated_at": "2026-10-28 04:13:41.434021+00:00", "lang": "en", "topics": ["machine-learning", "artificial-intelligence", "ai-research"], "entities": ["PAI-KNN", "arXiv"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/inference-time-prior-adaptation-for-imbalanced-data-streams-under-concept-drift", "markdown": "https://wpnews.pro/news/inference-time-prior-adaptation-for-imbalanced-data-streams-under-concept-drift.md", "text": "https://wpnews.pro/news/inference-time-prior-adaptation-for-imbalanced-data-streams-under-concept-drift.txt", "jsonld": "https://wpnews.pro/news/inference-time-prior-adaptation-for-imbalanced-data-streams-under-concept-drift.jsonld"}}