神经演替：持续学习中入侵、共存与稳定的介观理论
Neural Succession: A Mesoscopic Theory of Invasion, Coexistence, and Stabilization in Continual Learning
浏览论文内容
中文总结 AI 辅助
本文提出演替学习理论，以介观生态学视角解释持续学习中的遗忘与共存，通过入侵前兼容性预测遗忘，并形式化稳定条件，为持续学习提供预适应诊断方法。
中文摘要 AI 辅助
持续学习通常通过保留旧知识的机制来研究。我们提出了演替学习理论（SLT），这是一种介观解释，其中当前表示是常驻群落，新任务是入侵者，遗忘是常驻位移，联合保留是共存，重放是常驻强化，训练从建立走向稳定。其实证坐标是定向入侵前兼容性，在训练新任务之前对常驻模型进行测量。在八个实验中，兼容性在20个有向Split-CIFAR-10转换上预测了后续遗忘（三次重复r=-0.789，新任务簇95%置信区间[-0.90,-0.72]，每次单独重复r<=-0.67），以比无信息基线低24%的误差预测了保留集遗忘，并在受控MNIST排列和CIFAR-10旋转下复现（r=-0.804, -0.718）。在84个转换的套件上，兼容性在每个保留阈值下区分共存与排斥（AUC 0.93-0.97）。重放以最多325个存储示例修复每个转换，并且在位移最大的地方效率最高。兼容性达到|r|=0.720，而激活、表示、雅可比和固定系数Lotka-Volterra特化则不然。可塑性和特征更新率从训练早期到后期可靠下降（15/15和14/15次运行）。我们形式化了最小栖息地修改界限、位移下限、充分共存条件、可识别性定律及其范围限制推论、演替稳定性和局部强化。可识别性定律还预测了坐标失去效力的位置，该预测与三个CIFAR-100分区和五个优化器机制相匹配。SLT是一种预适应诊断，补充了重放、正则化和投影方法。
英文摘要
Continual learning is usually studied through mechanisms that preserve old knowledge. We develop Successional Learning Theory (SLT), a mesoscopic account in which the current representation is a resident community, the incoming task is an invader, forgetting is resident displacement, joint retention is coexistence, replay is resident reinforcement, and training moves from establishment toward stabilization. Its empirical coordinate is directional pre-invasion compatibility, measured on the resident model before the incoming task is learned. Across eight experiments, compatibility orders later forgetting on the 20 directed Split-CIFAR-10 transitions (three-repeat r=-0.789, incoming-task cluster 95% CI [-0.90,-0.72], every repeat alone r<=-0.67), forecasts held-out forgetting with 24% lower error than a no-information baseline, and reproduces under controlled MNIST permutations and CIFAR-10 rotations (r=-0.804, -0.718). On an 84-transition suite, compatibility separates coexistence from exclusion at every retention threshold (AUC 0.93-0.97). Replay repairs every transition with at most 325 stored examples and is most efficient where displacement is largest. Compatibility reaches |r|=0.720, while activation, representation, Jacobian, and fixed-coefficient Lotka-Volterra specializations do not. Plasticity and feature turnover fall reliably from early to late training (15/15 and 14/15 runs). We formalize a minimum habitat-modification bound, a displacement floor, a sufficient coexistence condition, an identifiability law with a range-restriction corollary, successional stabilization, and local reinforcement. The identifiability law also predicts where the coordinate loses leverage, and the prediction matches three CIFAR-100 partitions and five optimizer regimes. SLT is a pre-adaptation diagnostic that complements replay, regularization, and projection methods.
发表机构
- Indiana University Indianapolis（印第安纳大学印第安纳波利斯分校）
- North South University（南北大学）
机构由 AI 辅助整理，请以论文原文为准。
