Y
Yusuke Koda
Researcher at Kyoto University
Publications - 41
Citations - 515
Yusuke Koda is an academic researcher from Kyoto University. The author has contributed to research in topics: Computer science & Reinforcement learning. The author has an hindex of 8, co-authored 36 publications receiving 213 citations.
Papers
More filters
Journal ArticleDOI
Distillation-Based Semi-Supervised Federated Learning for Communication-Efficient Collaborative Training with Non-IID Private Data
TL;DR: A distillation-based semi-supervised FL (DS-FL) algorithm that exchanges the outputs of local models among mobile devices, instead of model parameter exchange employed by the typical frameworks to overcome largely incremental communication costs due to model sizes in typical frameworks.
Journal ArticleDOI
Proactive Received Power Prediction Using Machine Learning and Depth Images for mmWave Networks
Takayuki Nishio,Hironao Okamoto,Kota Nakashima,Yusuke Koda,Koji Yamamoto,Masahiro Morikura,Yusuke Asai,Ryo Miyatake +7 more
TL;DR: In this article, the authors leverage camera imagery and machine learning to construct a prediction model from a dataset of sequential images labeled with received power in several hundred milliseconds ahead of the time at which each image is obtained.
Journal ArticleDOI
Handover Management for mmWave Networks With Proactive Performance Prediction Using Camera Images and Deep Reinforcement Learning
TL;DR: In this paper, the authors proposed a proactive handover framework for millimeter-wave networks, where handover timings are optimized while obstacle-caused data rate degradations are predicted before the degradation occurs.
Proceedings ArticleDOI
Reinforcement learning based predictive handover for pedestrian-aware mmWave networks
TL;DR: Reinforcement learning is applied to learn the optimal handover policy maximizing the future throughput expected under the locations and velocities of a pedestrian to outperform the heuristic handover decisions in terms of throughput performance.
Proceedings ArticleDOI
Differentially Private AirComp Federated Learning with Power Adaptation Harnessing Receiver Noise
TL;DR: In this paper, a differentially private over-the-air computation (AirComp)-based federated learning (FL) was designed, where the key idea is to harness receiver noise perturbation injected to aggregated global models inherently, thereby preventing the inference of clients' private data.