QuantumCore's proprietary technology using Reservoir Computing has enabled a fully recurrent neural network (RNN) that can go from training to inference on edge devices.
Supports not only ARM Cortex-A series but also Cortex-M series! The minimum memory requirement is 64kb, so it can be implemented in a wide range of devices.
Provides QoreSDK for use with Python and C/C++. Qore also provides QORE CLOUD, a cloud tool that allows you to design without writing a single line of code. It supports regression, classification, and clustering tasks.
Less defect data in manufacturing, less personal data in healthcare, or poor internet connection.
Accuracy: 93%
Train time: 3sec
*Arrhythmia detection from open source
Accuracy: 99%
Train time: 0.8sec
*9speakers’ voice data from open source
Accuracy: 94%
Train time: 8sec
Accuracy: 98%
Train time: 12sec
*24 category identification by Qcore
Accuracy: 80%
Train time: 15sec
Accuracy: 96%
Train time: 3sec
CEO, Co-Founder
Search engine dev. in "Excite"
R&D in deep learning in "Mistletoe"
New AI biz-dev. in "Digital Garage"
CTO, Co-Founder
Auction dev. in "Yahoo!"
Founding engineer in "Tabelog"
CTO in "DGLab AI"
Research adviser
Associate Professor in "Future University Hakodate"
Research fellow in "U-TOKYO"
Best paper in "IJCNN2019"
QuantumCore has been selected for "Microsoft for Startups".
QuantumCore is an associate member of SEMI.
QuantumCore was selected as one of the TOP20 companies in "Innovation Leader Summit2019".