Author
Publications
A Graph Theoretic Framework of Recomputation Algorithms for Memory-Efficient Backpropagation
NeurIPS 2019
By : Mitsuru Kusumoto, Takuya Inoue, Gentaro Watanabe, Takuya Akiba, Masanori Koyama
Sampling Techniques for Large-Scale Object Detection from Sparsely Annotated Objects
CVPR 2019 (oral)
By : Yusuke Niitani, Takuya Akiba, Tommi Kerola, Toru Ogawa, Shotaro Sano, Shuji Suzuki
Variance-based Gradient Compression for Efficient Distributed Deep Learning
Invited to Workshop in ICLR 2018
By : Yusuke Tsuzuku, Hiroto Imachi, Takuya Akiba
ChainerMN: Scalable Distributed Deep Learning Framework
ML Systems Workshop in NIPS 2017