| 研究課題/領域番号 |
23K11080
|
|---|
| 研究機関 | 茨城大学 |
|---|
研究代表者 |
王 瀟岩 茨城大学, 理工学研究科(工学野), 准教授 (10725667)
|
|---|
| 研究分担者 |
梅比良 正弘 南山大学, 理工学部, 教授 (00436239)
|
|---|
| 研究期間 (年度) |
2023-04-01 – 2026-03-31
|
| キーワード | split learning / in-network learning |
|---|
| 研究実績の概要 |
In FY2023, we have realized the efficient split in-network learning approach with fixed model and network settings. Specifically, we focused on the realization of BS-side model virtualization and UE-BS model’s gradient aggregation, under a given model partition and UE set. We validated the proposed approach’s convergence speed and model accuracy by simulations on PyTorch. Different from most of the previous studies that use MNIST dataset (i.e., handwritten digits from 0 to 9), we took into consideration the real mobile deep learning applications, and thus used an aerial view human action detection dataset for evaluation.
|
| 現在までの達成度 (区分) |
現在までの達成度 (区分)
2: おおむね順調に進展している
理由
We consider that the research progresses smoothly. We have published 1 journal papers, 2 internal conference papers and multiple domestic conference papers in FY2023 under the support of this funding.
|
| 今後の研究の推進方策 |
In the following year, we will consider to refine the approach and clarify the performance tradeoff curve under various resource budget. Furthermore, we will consider to implement the proposed approach on testbed.
|
| 次年度使用額が生じた理由 |
Parts of the conferences were attended online, by which the grant originally planed has not been used. These used grant will be used to purchase the workstation and for the participation fee for conferences in FY2024.
|
