DRL-Based Workload Allocation for Distributed Coded Machine Learning
Bibliographic record
Abstract
Over the past years, Distributed Machine Learning (DML) has been employed to tackle the high complexity problem with many Machine Learning (ML) algorithms. With DML, the original computation task involved in an ML algorithm is first split into multiple subtasks, which are forwarded to a group of computing devices in a distributed environment. Thereafter, the subtask results are collected in order to arrive at the final result for the original task. Despite the advantages of DML, the overall computation time can be seriously increased if some subtask results cannot be collected in a timely manner due to device malfunctions or network glitches. Recently, Distributed Coded Machine Learning (DCML) has been proposed to mitigate the problem with DML. Specifically, DCML employs coding techniques to inject redundancy into the original computation task. With the injected redundancy, DCML does not need to collect all subtask results to construct the result for the original task. So far, how to split the original task and thereafter assign an appropriate workload to each computing device in DCML has been a challenging problem. In this paper, we propose a novel work allocation scheme for DCML, DWA, to tackle the challenging problem. Our experimental results indicate that DWA outperforms the existing DCML 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 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 | 0.016 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.012 | 0.016 |
| 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; both teacher heads agree on what is shown here.
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".