Learning When and How to Forget in Variable Window LFU Caching
Bibliographic record
Abstract
In this paper, we introduce a novel caching approach to improve content delivery, particularly for local small-cell access points with limited computational and memory resources. We focus on improving cache hit ratios and minimizing the load on the fronthaul link with the aid of a change point detection module. Importantly, the whole process is done locally in the context of emerging self-organizing networks and preserving privacy. We present a modified least frequently used (LFU) caching strategy, called variable window LFU (VW-LFU), which adapts to changes in content popularity. We introduce a non-parametric change point detection method, which recognizes changes in the traffic distribution with less than a 12% delay and prompts the VW-LFU system to adjust its window size. This approach mitigates biases from previous environments, improving cache decisions and fronthaul load. We prove that in an environment with several countable stationary modes, the average regret with our algorithm can be as low as <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$O$</tex>(1). Implemented in a four-room conference center with nine access points, our results show more than a 40% improvement in the average hit ratio, in low-capacity regimes, demonstrating the effectiveness of our strategy. We note that VW-LFU has an efficient implementation of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$O$</tex>(1) with the aid of hashtable and linked frequency lists.
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 distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
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 teacher head, 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".