arXiv:2606.09430v1 Announce Type: new 
Abstract: Online task-free continual learning (TFCL) requires intelligent agents to sequentially accumulate knowledge from an unbounded, non-stationary data stream under strict single-pass constraints and without any explicit task identifiers. Existing online TFCL paradigms primarily rely on parameter-efficient prompt tuning or dynamic structure expansion driven by training-coupled optimization dynamics, such as empirical loss fluctuations or evolving latent distances. As a result, these training-coupled solvers remain agnostic to the structural origins of distribution drift, mechanically enforcing a fixed strategy across fundamentally distinct streaming variations. To address this gap, we propose LargeMonitor, a framework that leverages large pretrained foundation models to autonomously orchestrate task-free continuous adaptation. Specifically, LargeMonitor introduces a decoupled detection module utilizing the frozen, stable representation space of large vision models (LVMs) to achieve robust, zero-shot drift detection without training-dependent interference or brittle threshold tuning. Upon a confirmed drift, the framework activates a context-aware diagnostic module driven by large multimodal models (LMMs) to interpret the precise semantic etiologies of the stream variation (e.g., novel class emergence vs. environmental domain shift). This dual-stage capability empowers the continuous learner to dynamically deploy adaptive and shift-specific optimization strategies. Extensive experiments across multiple TFCL settings and benchmarks demonstrate that LargeMonitor achieves precise, robust detection and diagnosis of complex data streams while consistently improving the performance of existing online TFCL algorithms.

يمثل تقديم LargeMonitor تقدمًا كبيرًا في التعلم المستمر بدون مهام عبر الإنترنت (TFCL)، مما يمكّن الوكلاء الذكيين من تجميع المعرفة بشكل تكيفي من تدفقات البيانات غير الثابتة دون معرفات مهام صريحة. يستخدم هذا الإطار نماذج كبيرة مدربة مسبقًا لتعزيز التكيف المستمر من خلال وحدة كشف مفصولة.

La introducción de LargeMonitor marca un avance significativo en el aprendizaje continuo sin tareas en línea (TFCL), permitiendo a los agentes inteligentes acumular conocimientos de manera adaptativa a partir de flujos de datos no estacionarios sin identificadores de tareas explícitos. Este marco utiliza grandes modelos preentrenados para mejorar la adaptación continua a través de un módulo de detección desacoplado.

L'introduction de LargeMonitor représente une avancée significative dans l'apprentissage continu sans tâche en ligne (TFCL), permettant aux agents intelligents d'accumuler de manière adaptative des connaissances à partir de flux de données non stationnaires sans identifiants de tâche explicites. Ce cadre utilise de grands modèles préentraînés pour améliorer l'adaptation continue grâce à un module de détection découplé.

The introduction of LargeMonitor marks a significant advancement in online task-free continual learning (TFCL), enabling intelligent agents to adaptively accumulate knowledge from non-stationary data streams without explicit task identifiers. This framework utilizes large pretrained models to enhance continuous adaptation through a decoupled detection module.