Efficient Model Training in Edge Networks With Hierarchical Split Learning
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
In this paper, we propose an efficient model training scheme, named <underline xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">G</u>roup-based <underline xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">H</u>ierarchical <underline xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">S</u>plit <underline xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">L</u>earning (GHSL), which can accelerate the artificial intelligence (AI) training process in edge networks in a “first-sequential-then-parallel” manner. Specifically, the proposed scheme hierarchically splits an AI model into a user-side and server-side model, while dividing a number of users into multiple groups. Users in each group train user-side models with the interaction of the shared server-side model sequentially; different groups perform the above training process parallelly; the AI models of each group are aggregated into a global model. We also carry out the convergence analysis for the proposed scheme over non-independent and identically distributed data, which reveals that the convergence rate depends on user grouping. Furthermore, we propose a data-driven two-stage user grouping algorithm to minimize the overall training delay, taking user resource heterogeneity and the black-box training process into account. The proposed algorithm first utilizes the Gaussian process regression approach to determine the number of groups, and then employs the coalition game theory to determine the optimal user grouping decision. Comprehensive simulation results demonstrate that the proposed scheme can reduce training delay, user-side computational workload, and communication overhead by up to 19%, 53%, and 54%, respectively, comparing to state-of-the-art benchmarks.
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.
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 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.001 |
| 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 it