解锁持续少样本学习中的组合泛化
Unlocking Compositional Generalization in Continual Few-Shot Learning
AI总结：
本文提出了一种新的持续少样本学习范式，通过严格解耦表示学习与组合推理，实现对新概念的高效泛化，并在多个基准测试中取得最佳性能。
摘要：
基于对象的表示方法在少样本学习中具有关键属性：而不是将场景视为单一单元，模型可以将其分解为个体对象级别的部分，这些部分可以在不同概念之间进行匹配和比较。在实践中，这种潜力很少被实现。持续学习者要么将场景压缩成全局嵌入，要么通过部分级匹配目标进行训练，这使表示过于紧密地依赖于已见过的模式，从而无法泛化到真正的新概念。在本文中，我们识别出这种根本性的结构冲突，并开创了一种新的范式，严格解耦表示学习与组合推理。利用自监督视觉变换器（ViTs）固有的片段级语义几何，我们的框架采用双阶段策略。在训练期间，槽表示完全优化为整体类别身份，保留高度可泛化的对象级几何结构。在推理期间，保留的槽被动态组合以匹配新场景。我们证明了这种范式提供了双重结构优势：冻结的主干自然防止了表示漂移，而我们的轻量级、整体优化保持了特征对新概念转移的能力。广泛的实验验证了这种方法，在标准持续学习基准中实现了最佳的未见概念泛化和最小的遗忘。
英文摘要：
Object-centric representations promise a key property for few-shot learning: Rather than treating a scene as a single unit, a model can decompose it into individual object-level parts that can be matched and compared across different concepts. In practice, this potential is rarely realized. Continual learners either collapse scenes into global embeddings, or train with part-level matching objectives that tie representations too closely to seen patterns, leaving them unable to generalize to truly novel concepts. In this paper, we identify this fundamental structural conflict and pioneer a new paradigm that strictly decouples representation learning from compositional inference. Leveraging the inherent patch-level semantic geometry of self-supervised Vision Transformers (ViTs), our framework employs a dual-phase strategy. During training, slot representations are optimized entirely toward holistic class identity, preserving highly generalizable, object-level geometries. At inference, preserved slots are dynamically composed to match novel scenes. We demonstrate that this paradigm offers dual structural benefits: The frozen backbone naturally prevents representation drift, while our lightweight, holistic optimization preserves the features' capacity for novel-concept transfer. Extensive experiments validate this approach, achieving state-of-the-art unseen-concept generalization and minimal forgetting across standard continual learning benchmarks.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
