Multi-step Prediction of Worker Resource Usage at the Extreme Edge
Bibliographic record
Abstract
Democratizing the edge by leveraging the prolific yet underutilized computational resources of end devices, referred to as Extreme Edge Devices (EEDs), can open a new edge computing tech market that is people-owned, democratically managed, and accessible/lucrative to all. Parallel computing at EEDs can also move the computing service much closer to end-users, which can help satisfy the stringent Quality-of-Service (QoS) requirements of delay-critical and/or data-intensive IoT applications. However, EEDs are heterogeneous user-owned devices, and are thus subject to a highly dynamic user access behavior (i.e., dynamic resource usage). This makes the process of determining the computational capability of EEDs increasingly challenging. Estimating the dynamic resource usage of EEDs (i.e., workers) has been mostly overlooked. The complexity of Machine Learning (ML)-based models renders them impractical for deployment at the edge for the purpose of such estimations. In this paper, we propose the Resource Usage Multi-step Prediction (RUMP) scheme to estimate the dynamic resource usage of workers over multiple steps ahead in a computationally efficient way while providing a relatively high prediction accuracy. Towards that end, RUMP exploits the use of the Hierarchical Dirichlet Process-Hidden Semi-Markov Model (HDP-HSMM) to estimate the dynamic resource usage of workers in EED-based computing paradigms. Extensive evaluations on a real testbed of heterogeneous workers for multi-step sizes show an 87.5% prediction accuracy for the starting point of 2-steps and coming to as little as a 16% average difference in prediction error compared to a representative of state-of-the-art ML-based schemes.
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 | 0.001 | 0.004 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 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 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".