Generalized Incremental Learning under Concept Drift across Evolving Data Streams

TL;DR

This paper introduces a novel framework called CSFA for generalized incremental learning in data streams with concept drift, effectively handling evolving label spaces and distributions without extensive supervision.

Contribution

It formalizes GILCD and proposes CSFA, a training-free, source-free adaptation method combining prototype calibration and sharpness-aware optimization for robust open-world streaming learning.

Findings

CSFA outperforms state-of-the-art methods in experiments.

It effectively handles label space co-evolution and distribution shifts.

The framework maintains high accuracy with limited supervision.

Abstract

Real-world data streams exhibit inherent non-stationarity characterized by concept drift, posing significant challenges for adaptive learning systems. While existing methods address isolated distribution shifts, they overlook the critical co-evolution of label spaces and distributions under limited supervision and persistent uncertainty. To address this, we formalize Generalized Incremental Learning under Concept Drift (GILCD), characterizing the joint evolution of distributions and label spaces in open-environment streaming contexts, and propose a novel framework called Calibrated Source-Free Adaptation (CSFA). First, CSFA introduces a training-free prototype calibration mechanism that dynamically fuses emerging prototypes with base representations, enabling stable new-class identification without optimization overhead. Second, we design a novel source-free adaptation algorithm, i.e., Reliable Surrogate Gap Sharpness-aware (RSGS) minimization. It integrates sharpness-aware perturbation loss optimization with surrogate gap minimization, while employing entropy-based uncertainty filtering to discard unreliable samples. This mechanism ensures robust distribution alignment and mitigates generalization degradation caused by uncertainties. Thus, CSFA establishes a unified framework for stable adaptation to evolving semantics and distributions in open-world streaming scenarios. Extensive experiments validate the superior performance and effectiveness of CSFA compared to SOTA approaches.