124 citations · 666 across the 43 of their papers we have counts for
44 papers
Two-Way Coupled Dynamics of Quantized Vortex Bundle Formation in Spatially Periodic Normal-Fluid Turbulence
Hiromichi Kobayashi, Makoto Tsubota, Rio Yokota
Using fully coupled two-way simulations combining the vortex filament model with the Navier--Stokes equations, we discover that quantized vortices in superfluid He are systemat…
Proxy-surface-based fast direct solver for TE-mode scattering problems on distributed memory systems
Yasuhiro Matsumoto, Rio Yokota
This paper describes an MPI/OpenMP hybrid parallelized fast direct solver for the scattering problem of transverse electric (TE)-mode electromagnetic waves. Because TE-mode scatter…
Computing the k-th Eigenvalue of Symmetric -Matrices
M. Ridwan Apriansyah, Rio Yokota
The numerical solution of eigenvalue problems is essential in various application areas of scientific and engineering domains. In many problem classes, the practical interest is on…
Cache Optimization and Performance Modeling of Batched, Small, and Rectangular Matrix Multiplication on Intel, AMD, and Fujitsu Processors
Sameer Deshmukh, Rio Yokota, George Bosilca
Factorization and multiplication of dense matrices and tensors are critical, yet extremely expensive pieces of the scientific toolbox. Careful use of low rank approximation can dra…
SegRCDB: Semantic Segmentation via Formula-Driven Supervised Learning
Risa Shinoda, Ryo Hayamizu, Kodai Nakashima +3
Pre-training is a strong strategy for enhancing visual models to efficiently train them with a limited number of labeled images. In semantic segmentation, creating annotation masks…
Reducing shared memory footprint to leverage high throughput on Tensor Cores and its flexible API extension library
Hiroyuki Ootomo, Rio Yokota
NVIDIA Tensor Core is a mixed-precision matrix-matrix multiplication and addition computing unit, where the theoretical peak performance is more than 300 TFlop/s on NVIDIA A100 GPU…
Pre-training Vision Transformers with Very Limited Synthesized Images
Ryo Nakamura, Hirokatsu Kataoka, Sora Takashima +3
Formula-driven supervised learning (FDSL) is a pre-training method that relies on synthetic images generated from mathematical formulae such as fractals. Prior work on FDSL has sho…
DGEMM on Integer Matrix Multiplication Unit
Hiroyuki Ootomo, Katsuhisa Ozaki, Rio Yokota
Deep learning hardware achieves high throughput and low power consumption by reducing computing precision and specializing in matrix multiplication. For machine learning inference,…
ASDL: A Unified Interface for Gradient Preconditioning in PyTorch
Kazuki Osawa, Satoki Ishikawa, Rio Yokota +2
Gradient preconditioning is a key technique to integrate the second-order information into gradients for improving and extending gradient-based learning algorithms. In deep learnin…
Mixed-Precision Random Projection for RandNLA on Tensor Cores
Hiroyuki Ootomo, Rio Yokota
Random projection can reduce the dimension of data while capturing its structure and is a fundamental tool for machine learning, signal processing, and information retrieval, which…
Visual Atoms: Pre-training Vision Transformers with Sinusoidal Waves
Sora Takashima, Ryo Hayamizu, Nakamasa Inoue +2
Formula-driven supervised learning (FDSL) has been shown to be an effective method for pre-training vision transformers, where ExFractalDB-21k was shown to exceed the pre-training…
Quantum Circuit Simulation by SGEMM Emulation on Tensor Cores and Automatic Precision Selection
Hiroyuki Ootomo, Hidetaka Manabe, Kenji Harada +1
Quantum circuit simulation provides the foundation for the development of quantum algorithms and the verification of quantum supremacy. Among the various methods for quantum circui…
Informative Sample-Aware Proxy for Deep Metric Learning
Aoyu Li, Ikuro Sato, Kohta Ishikawa +2
Among various supervised deep metric learning methods proxy-based approaches have achieved high retrieval accuracies. Proxies, which are class-representative points in an embedding…
Empirical Study on Optimizer Selection for Out-of-Distribution Generalization
Hiroki Naganuma, Kartik Ahuja, Shiro Takagi +5
Modern deep learning systems do not generalize well when the test data distribution is slightly different to the training data distribution. While much promising work has been acco…
Scalable Linear Time Dense Direct Solver for 3-D Problems Without Trailing Sub-Matrix Dependencies
Qianxiang Ma, Sameer Deshmukh, Rio Yokota
Factorization of large dense matrices are ubiquitous in engineering and data science applications, e.g. preconditioners for iterative boundary integral solvers, frontal matrices in…
Parallel QR Factorization of Block Low-Rank Matrices
M. Ridwan Apriansyah, Rio Yokota
We present two new algorithms for Householder QR factorization of Block Low-Rank (BLR) matrices: one that performs block-column-wise QR, and another that is based on tiled QR. We s…
Pre-training Vision Transformers with Formula-driven Supervised Learning
Hirokatsu Kataoka, Sora Takashima, Ryo Hayamizu +6
In the present work, we show that the performance of formula-driven supervised learning (FDSL) can match or even exceed that of ImageNet-21k and can approach that of the JFT-300M d…
Recovering single precision accuracy from Tensor Cores while surpassing the FP32 theoretical peak performance
Hiroyuki Ootomo, Rio Yokota
Tensor Core is a mixed-precision matrix-matrix multiplication unit on NVIDIA GPUs with a theoretical peak performance of more than 300 TFlop/s on Ampere architectures. Tensor Cores…
OPIRL: Sample Efficient Off-Policy Inverse Reinforcement Learning via Distribution Matching
Hana Hoshino, Kei Ota, Asako Kanezaki +1
Inverse Reinforcement Learning (IRL) is attractive in scenarios where reward engineering can be tedious. However, prior IRL algorithms use on-policy transitions, which require inte…
Quantum Turbulence Coupled with Externally Driven Normal-Fluid Turbulence in Superfluid He
Satoshi Yui, Hiromichi Kobayashi, Makoto Tsubota +1
The coupled dynamics of quantum turbulence (QT) and normal-fluid turbulence (NFT) have been a central challenge in quantum hydrodynamics, since it is expected to cause the unsolved…
RePOSE: Fast 6D Object Pose Refinement via Deep Texture Rendering
Shun Iwase, Xingyu Liu, Rawal Khirodkar +2
We present RePOSE, a fast iterative refinement method for 6D object pose estimation. Prior methods perform refinement by feeding zoomed-in input and rendered RGB images into a CNN…
Epipolar-Guided Deep Object Matching for Scene Change Detection
Kento Doi, Ryuhei Hamaguchi, Shun Iwase +3
This paper describes a viewpoint-robust object-based change detection network (OBJ-CDNet). Mobile cameras such as drive recorders capture images from different viewpoints each time…
Scalable and Practical Natural Gradient for Large-Scale Deep Learning
Kazuki Osawa, Yohei Tsuji, Yuichiro Ueno +3
Large-scale distributed training of deep neural networks results in models with worse generalization performance as a result of the increase in the effective mini-batch size. Previ…
Effect of Mixed Precision Computing on H-Matrix Vector Multiplication in BEM Analysis
Rise Ooi, Takeshi Iwashita, Takeshi Fukaya +2
Hierarchical Matrix (H-matrix) is an approximation technique which splits a target dense matrix into multiple submatrices, and where a selected portion of submatrices are low-rank…