RESEARCH

Our research is divided into function-based research programs
named the Language program, the Customer program, the Vision program, and the Moonshot programs.

Rakuten Institute of Technology covers a broad range of research topics.
Please refer to the explanation of each domain, and the materials which explain the details of some of our research topics. Also, part of our contribution is available as open-source software.

Language program

Fully automated AI catalog

Rakuten is today one of the largest e-commerce platforms in Japan with more than 200 million items. One of the foundation technologies to innovate shopping experience is the clean, structured item data that capture the key features that lead to the purchase decision-making for the customres, such as size and color infromation for fashion and spec information for electronics. Once we have the data, we believe it is possible to make much better discovery and search experience based on AI. However, challenge is creating such structured data is very costly and manual with the current technology. In the language program, we challenge to develop "fully automated AI catalog" where all the necessary technologies for structures such as item categorization, feature extraction, query & customer behavior understanding and normalizations are done by automated AI with small inputs from annotations.
In addition, Rakuten aims to become one of the largest communication platform with Rakuten mobile and Viber. With this global ambition, we develop machine translation capability under the language program.

Customer program

Universal customer model

Rakuten group operates more than 70 different services, and e-commerce and Fintech services such as Rakuten card are the largest business in the industry with more than 100 million IDs in Japan and 1.4 billion customers across the globe. With this large and deep customer data, Rakuten is one of the best positioned company to innovate understanding of human being in terms of how people spend money and time. Customer program challenges to build universal customer models and its applications, motivated by the recent success of pre-trained models in language and vision fields. Such models can be adapted to improve diverse set of solutions for services including recommendation systems, customer targeting solutions and credit scoring functions to name a few. As initial and simpler applications, we work on predictions of customer features so-called customer DNA.

Vision program

Frictionless experience with images

Images play essential role as communication medium for shopping experience. Leveraging state-of-the-art computer vision technology, Vision Program builds novel systems to improve such experiences through technologies such as recognizing objects, extracting attributes, enhancing image quality or generating images, with the aim to deploy in real-world products. In addition, pictures such as drivers license are often used in critical identification process so-called KYC, and as one of the biggest Fintech companies in Japan, removing friction of such identification process is critical for the customer experience. We develop robust e-KYC algorithms. Across all the examples, we combine the technologies developed by language and vision programs to create a multi-modal approach combining different types of input data.

Moonshot programs

5G Experience

Together with Rakuten Mobile, we strive to create new user experiences using Rakuten’s 5G mobile network. We focus on technical management and research and development of proof-of-concept solutions that best demonstrate the use of high bandwidth, low latency, and multiple connections of 5G. We also support the development of new 5G business cases, and find new monetization opportunities, together with various departments within the Rakuten group, academic institutions, and other external partners.

Autonomous Network

We are conducting research to achieve a truly autonomous network by AI, together with Rakuten Mobile. Rakuten Mobile is the first company that can ever try to solve this problem, enabled by its software-defined architecture of the network.User demand drives networking, communication, and data center innovations; these factors revolutionize the Telco business. As new services and user devices are introduced daily, the demand constantly increases and the range of configuration and adaptation options grow exponentially. Thus, our goal is to build a network that handles and adapts itself to unforeseen changes autonomously.

Cure Cancer

Curing cancer is, arguably, one of the most challenging problems that human society is facing today. Rakuten is innovating this fight against cancer with photo-immunotherapy treatment called Illuminox in Rakuten medical. Rakuten Institute of Technology contributes to this moonshot goal by working together with Rakuten medical to create solutions such as multi-modal analysis to predict patient responses to treatment for personal medicine, and computer-vision algorithms for medical image processing.

NEWEST PROJECTS

PRESENTATION ARCHIVE

Filter by Research Area:
Search by Keyword: