Age of Learning: Temporal Persistence of Prediction Errors as a Learning Signal

2026-09-26

AI Summary

This study addresses the limitation of existing algorithms that rely solely on instantaneous signals, which struggle to capture the duration of prediction errors. To overcome this, we introduce a "learning age" variable that quantifies error persistence, effectively distinguishing sustained under-learning from transient mistakes. For the first time, temporal error persistence is incorporated as an independent state variable to complement conventional difficulty metrics. This formulation is integrated with offline sample-level accumulation, streaming class-level maintenance, and adaptive resampling techniques to optimize training dynamics. Empirical evaluations on long-tailed classification tasks demonstrate that the proposed approach outperforms or matches existing baselines. These results validate the efficacy of leveraging temporal persistence in regulating learning dynamics within non-stationary environments.

Abstract

Current machine learning algorithms primarily rely on instantaneous signals such as loss, margin, and prediction confidence to characterize model behavior. These signals indicate how difficult a prediction is at the current optimization step, but they do not capture how long the model has remained incorrect. We study this temporal dimension of learning and introduce Age of Learning (AoL), a learning-state variable that measures the persistence of prediction errors over time. AoL increases while an error remains unresolved and resets when a correct prediction is achieved, thereby distinguishing persistent under-learning from transient mistakes. We develop AoL-based training strategies for both offline and streaming settings. In offline learning, sample-level AoL is accumulated over training and aggregated into class-level states that guide adaptive reweighting and resampling. In streaming learning, where full historical access is unavailable, we maintain lightweight class-level AoL states using current and buffered observations. Across long-tailed classification settings, AoL improves or matches standard training baselines, with larger benefits when learning difficulty persists over time. Multi-seed streaming experiments further show reproducible gains under temporally stable imbalance. Analysis of class frequency, loss, and margin shows that AoL is related to conventional difficulty measures but captures additional information about error duration. These results suggest that temporal persistence provides a useful complementary signal for characterizing and controlling learning dynamics in imbalanced and non-stationary environments.

Problem

• prediction error persistence
• temporal learning dynamics
• long-tailed classification
• streaming learning
• learning signal

Innovation

• Age of Learning
• Temporal Persistence
• Prediction Errors
• Long-tailed Classification
• Streaming Learning