Du Wu

ORCID: 0000-0002-4002-0837
Publications
Citations
Views
---
Saved
---
About
Contact & Profiles
Research Areas
  • Parallel Computing and Optimization Techniques
  • Advanced Data Storage Technologies
  • Advanced X-ray Imaging Techniques
  • Anomaly Detection Techniques and Applications
  • Cell death mechanisms and regulation
  • Color perception and design
  • Medical Imaging Techniques and Applications
  • Manufacturing Process and Optimization
  • Simulation and Modeling Applications
  • Software System Performance and Reliability
  • VLSI and FPGA Design Techniques
  • interferon and immune responses
  • Advanced Sensor and Control Systems
  • Human Motion and Animation
  • Advanced Neural Network Applications
  • Artificial Intelligence in Games
  • Video Surveillance and Tracking Methods
  • Medical Research and Treatments
  • Advanced MRI Techniques and Applications
  • Image and Video Quality Assessment
  • RNA and protein synthesis mechanisms
  • Industrial Vision Systems and Defect Detection
  • Hydraulic and Pneumatic Systems
  • Educational Reforms and Innovations
  • Matrix Theory and Algorithms

Tokyo Institute of Technology
2024

RIKEN Center for Computational Science
2024

Shenzhen Institutes of Advanced Technology
2021-2022

Jilin Province Science and Technology Department
2019

Jilin University
2019

National Cheng Kung University
1999

The TIPE2 (tumor necrosis factor-alpha-induced protein 8-like 2) is a major regulator of cancer and inflammatory diseases. availability its sequence structure, as well the critical amino acids involved in ligand binding, provides insights into function helps greatly identify novel drug candidates against protein. With current advances deep learning molecular dynamics simulation-based screening, large-scale exploration inhibitory for becomes possible. In this work, we apply learning-based...

10.3389/fphar.2021.772296 article EN cc-by Frontiers in Pharmacology 2021-11-23

Computed Tomography (CT) serves as a key imaging technology that relies on computationally intensive filtering and back-projection algorithms for 3D image reconstruction. While conventional high-resolution reconstruction (> 2K3) solutions provide quick results, they typically treat an offline workload to be performed remotely large-scale HPC systems. The growing demand post-construction AI-driven analytics the need real-time adjustments call are feasible local computing resources, i.e....

10.1145/3650200.3656634 article EN 2024-05-30

We present <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">FastConv</i> , a template-based code auto-generation open-source library that can automatically generate high-performance deep learning convolution kernels of arbitrary matrices/tensors shapes. FastConv is based on the Winograd algorithm, which reportedly highest performing algorithm for time-consuming layers convolutional neural networks. ARM CPUs cover wide range designs and...

10.1109/tpds.2022.3146257 article EN publisher-specific-oa IEEE Transactions on Parallel and Distributed Systems 2022-01-27

Crowd counting is an important research topic in the field of computer vision. The multi-column convolution neural network (MCNN) has been used this and achieved competitive performance. However, when crowd distribution uneven, accuracy based on MCNN still needs to be improved. In order adapt uneven distributions, global density feature taken into account paper. features are extracted added through cascaded learning method. Because some detailed during down-sampling process will lost it...

10.1109/access.2019.2926881 article EN cc-by IEEE Access 2019-01-01

In the existing network security protection system, border is most important line of defense. At present, boundary large enterprises mainly uses system based on traffic log analysis. This has many advantages after a long time development. However, increasing amount data had great impact efficiency analysis equipment and analysts.This paper proposes cleaning method FA-CNN, using principal component (PCA) filtering model as foundation. Firstly, an initial screening for abnormal logs...

10.1109/pandafpe57779.2023.10141222 article EN 2023-04-01

In this paper, we propose a highly efficient computing method for game character control with phase-functioned neural networks (PFNN). The primary challenge to accelerate PFNN on mobile platforms is that dynamically produces weight matrices an argument, phase, which individual each character. Therefore existing libraries generally assume frozen are inefficient PFNN. situation becomes even worse when multiple characters present. To address the challenges, reformulate equations and leverage...

10.1145/3545008.3545095 article EN 2022-08-29
Coming Soon ...