Website :  https://twinsense.ai/

“Neuromorphic computing” is a computing technology that uses a model inspired by the function of nerve cells such as the brain of a living organism. It is attracting worldwide attention as the next generation of cutting-edge AI technology, and Intel Corporation in the United States has begun large-scale LSI development, including SNN (spiking neural network). In particular, in video applications, it refers to a system composed of “neuromorphic sensing (surveying = the role of the eyes)” and “neuromorphic processing (data analysis = the role of the brain)”. TwinSense Co., Ltd. was established in December 2023 in order to be the first in Japan to embark on these two fields, which are expected to play a central role in the next generation of computers and to speed up data processing and significantly reduce power consumption.

[Background of the Development]

Currently, the development of neuromorphic computing and the associated cutting-edge AI technology SNN is progressing around the world. Neuromorphic sensing has already been developed by French startup company Prophesee and Sony semiconductor Solutions, and  enabling precise surveying. However, there is currently no globally effective development environment or software for processing and analyzing surveying data obtained by neuromorphic sensing, and it has not yet been put to practical use. In addition, there is no development environment that utilizes neuromorphic computing technology, and no environment for conducting demonstration experiments aimed at social implementation. In order to solve these issues, we are working on the following initiatives.

(1) Development of “Neuromorphic Processing” (already developed)

TwinSense has been developing “Neuromorphic Processing (Data Analysis = Brain’s Role)” to bring the technology of its already developed “Neuromorphic Sensing (surveying = eye role)” product closer to practical application as a solution. In order to advance AI edge processing, we are developing not only CNN (Convolutional Neural Network), the conventional AI technology for image analysis, but also SNN (Spiking Neural Network), the next-generation cutting-edge AI technology, at the same time, with the aim of putting it to practical use.

(2) Preparation of an “integrated development environment platform” (already developed)

In order to be the first in the world to implement neuromorphic computing in society, we have developed an “integrated development environment platform” based on the concept of EdgeSmart Technology™.

The platform integrates four technologies: event sensor technology, AI technology, 3D graphics technology, and computer vision technology, making it possible to develop edge terminals that can perform advanced real-time video analysis with low power consumption, which was previously considered difficult. Specifically, by using this platform, things that were previously considered impossible to detect using monitoring, such as river water levels and wave conditions at night, and sudden drone intrusions, can be clearly detected.

Integrated Development Environment Platform

(3) Solution of cloud services business (Accelerating development with this funding)

With this funding, we will develop a cloud services business that runs various applications developed using an integrated development environment platform. This cloud service will enable labor-saving and energy-saving applications such as traffic monitoring, people flow measurement, disaster monitoring, railway control, and monitoring. It will also enable customers to significantly reduce the costs of the systems they currently operate.

We will soon start evaluating the functions of the beta version together with partners who is leading companies in each field that are already collaborating with us, and achieve higher quality cloud services that is unable to do with existing solutions.

[Future outlook and recruitment]

In order to further evolve AI edge processing in the future, TwinSense is conducting research and development of Spiking Neural Network (SNN), which is attracting global attention, and aims for early practical application. In the future, we will continue to raise funds in this field and are globally recruiting research and development personnel in the following fields.

• Development of sensing technology and multi-sensing integration technology in the field of autonomous driving applications
• AI (CNN, SNN)

[Next-generation advanced technology SNN]

The history of AI research has focused on modeling the functions of the brain and building practical learning models. Since 2000, especially in the 2010s, folding neural networks (CNNs) and recurrent neural networks (RNNs) have made great progress, creating breakthroughs in the fields of image recognition and natural language processing as the first wave. The second wave was the emergence of the Transformer model, with startup OpenAI putting ChatGPT into practical use. Since then, generative AI services such as Microsoft’s Copilot, Google’s Gemini, and Meta’s Llama have been released and are rapidly expanding the market. In the future, the third wave is expected to see the practical use of spiking neural networks (SNNs), which engineer operations that are even closer to the brain. SNNs consume less energy and may be able to recognize complex phenomena at high speeds with ultra-low power consumption that is several hundredths to one hundred billionth of the current edge AI (※).

*Reference:

 Intel Newsroom: “Intel builds world’s largest neuromorphic system to realize highly sustainable AI”
https://www.intel.co.jp/content/www/jp/ja/newsroom/news/intel-builds-world-largest-neuromorphic-system.html

・AI-SCHOLAR
https://ai-scholar.tech/articles/survey/neuromorphics_loihi

[Investor comments]

 Energy & Environment Investment, Inc. Nishikawa Toru and Ozaki Chihiro:

“TwinSense is developing software that enables the social implementation of neuromorphic technology, a mechanism that mimics biological nerve cells using electronic circuits.

As this technology is implemented in society, we expect that it will enable information exchange at overwhelmingly higher speeds and with lower power consumption than conventional solutions, and will also contribute to reducing the load on data centers.

We have invested in this company because we expect that this technology will have a major impact if it is implemented in society in the smart city field and for autonomous product-related applications in the future.　As EEI, we will fully support the company’s growth.”

【TwinSense Funders】

CEO, Okamura Kazuhisa

After Intel Japan, he served as the Japan manager for over six European and American venture companies, and was the Japan representative for Prophesee, a French venture company that develops Event Vision Sensors that is neuromorphic sensing technology.

CTO, Benjamin Schmitt

He has lived in Japan for over 20 years and is a specialist in 3DCG, CV, and AI, having served as director at Digital Media Professionals and other positions, before going on to serve as head of applied technology at Prophesee Japan.

VP & Board Member, Shingo Kokudo

After working at Intel Japan, he was co-founder of the US venture NuCore Technology and executive officer at MediaTek (Taiwan), and is currently the representative director of TX Entrepreneur Partners, an organization that supports deep tech startups.

Company Overview
Company Name   : TwinSense Inc.
Established : December 19, 2023
Location: Setagaya-Ku, Tokyo Japan
CEO : Kazuhisa Okamura
Business : Development of neuro-based sensing and processing

Contact on press release：marcom@twinsense.ai
Find position：hr@twinsense.ai