Continual Learning for Streaming Data with Dynamic Memory Allocation and Drift-Aware Rehearsal
Bibliographic record
Abstract
The capability of artificial neural networks to learn continuously from non-stationary data streams without catastrophic forgetting remains a paramount challenge in machine learning. While Experience Replay (ER) has emerged as a robust strategy, standard approaches often rely on fixed-size memory buffers and uniform sampling, which are suboptimal for streaming data exhibiting significant concept drift and class imbalance. In this paper, we propose a novel framework titled Dynamic Memory Allocation with Drift-Aware Rehearsal (DM-DAR). Our approach introduces a dynamic budgeting mechanism that adjusts the memory quota per class based on real-time estimation of forgetting and feature space drift. Furthermore, we implement a dual-criterion retrieval strategy that prioritizes samples located near the decision boundary and those representing the centroid of drifting distributions. Extensive experiments on split-CIFAR100 and sequential TinyImage Net benchmarks demonstrate that DM-DAR significantly outperforms state-of-the-art baselines in terms of average accuracy and backward transfer, while maintaining computational efficiency. The results suggest that adaptive resource management is crucial for the longevity and plasticity of continual learning systems deployed in volatile environments.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.007 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.001 | 0.003 |
| Open science | 0.003 | 0.002 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.001 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".