Ledge Co., Ltd.
-Event report-Idea x PoC contest “ELCON” using image recognition solution “ELFE” ~ More than 150 unique ideas gathered, and 5 companies conducted PoC for 1 month ~
With the mission of “smoothly spreading AI and other cutting-edge technologies into society”, it operates Japan’s largest AI-specialized web media “Ledge.ai” and plans and develops AI solutions. Ledge Co., Ltd., in cooperation with Sony Network Communications Inc. and Sony Biz Networks Inc., will hold a competition event “ELCON” on October 11, 2022 (Tuesday) to solicit new ideas using the image recognition solution “ELFE”. ) to November 8, 2022 (Tuesday).
■ About “Elcon”
ELCON is a contest to solicit ideas about what can be solved by the image discrimination solution “ELFE” under the theme of “If there was an AI that can distinguish ○○” that I feel in my daily life and work. The purpose of this project is to freely imagine ELFE usage scenes through “ELCON” and discover the seeds of problems in the world. This time, we held the “Idea Section” to solicit ideas from the general public, and the “PoC Section” to practice PoC (proof of concept) using ELFE for business companies at the same time. In the PoC section, participating companies used ELFE licenses free of charge and conducted PoC for one month with the support of Sony Biz Networks Inc., a sponsor company.
◎ Event site: https://elfe-concours.studio.site
■The total number of ideas in the Idea Division exceeded 120. Excellence Award: “AI that can distinguish the color of traffic lights”
A total of 127 ideas were submitted to the Idea Division, and many applicants submitted multiple ideas. Among them, the following three ideas won the award.
·Outstanding performance award
Ideas that have a large social impact and can be realized with current technology
Mr. Ken Yamazaki “AI that can distinguish the color of traffic lights” ・Idea Award
Ideas that are new or unexpected and look forward to being realized Shinya Ono “AI that can identify when to eat cantaloupe”
Ideas that are novel, have a large impact on society, and can be realized with current technology
Mr. Kazuhiko Matsuoka “AI to determine the degree of dirt on the hull of a ship” There were also many unique ideas such as “AI that finds someone similar to your first love”, although they unfortunately missed out on the prize. A summary report of the idea division will be posted on the AI-related media “Ledge.ai” at a later date.
◎ Idea section report article: https://ledge.ai/elfe_idea_report/ ■ Five companies participated in the PoC section, and the “Strawberry disease inspection” model creation project won the Excellence Award. In the PoC section, there were more applications than the capacity, and 5 companies out of them conducted PoC for one month with the support of Sony Biz Networks Inc. The themes that the participating companies worked on are as follows.
・Advantest Co., Ltd.
“Distinguish between normal and abnormal boards from the external image of the board”
・aba Co., Ltd.
“AI to distinguish the amount of excrement (feces/urine) in diapers” ・Geological Research Institute Co., Ltd.
“Earthquake Waveform False Detection Judgment”
・Niigata University of International and Information Studies “Counting swans from drone images”
The Excellence Award went to welzo Co., Ltd. (former Nichiryu Nagase Co., Ltd.) for its “strawberry disease inspection” model creation project.
Mr. Tsubasa Sugawara of Sony Biz Networks talks about the reason why he was selected for the Excellence Award.
“It was hard to decide which project was better and which was better, but someone who had no prior experience in using technology was able to improve the accuracy in a month, achieved the original benchmark, and completed the PoC with a purpose. It was very impressive that they enjoyed the process of improving accuracy while acquiring knowledge, such as using open data without being tied down and changing tuning. I have selected this company for the Excellence Award because I want more companies to work on AI in the future.”
Regarding the summary report of the PoC section and this project, which won the Excellence Award, we plan to publish an article on the efforts on the AI-related media “Ledge.ai” at a later date.
■ About the award ceremony
The award ceremony was held online on Wednesday, December 14th. [Image 2
■ About image recognition solution “ELFE”
ELFE is an image discrimination AI solution that employs “Evolutional Learning for Feature Extraction”, a technology that automatically develops discrimination models. Sony’s proprietary technology enables AI development with about 1/100 of the training data compared to general development methods. In addition, AI can be automatically generated simply by uploading data, so even AI beginners can develop without code. In addition to images, sensor data and sound data can be used, and the optimal algorithm and usage environment can be selected. [Image 3
ELFE is a technology developed by Sony Group Corporation based on joint research between Sony Computer Science Laboratories, Inc. and Sony Group Corporation.
Since 2018, Sony Network Communications Inc. has been providing the service as an image recognition solution. From 2021, Sony Biz Networks Inc. will start offering “ELFE on AWS” as an AWS version of the image recognition solution on “Managed Cloud with AWS”.
■ Ledge Co., Ltd.
Representative: Kenta Kotaki
Location: NMF Gotanda Ekimae Building 7F, 1-1-8 Nishi-Gotanda, Shinagawa-ku, Tokyo
Operation of AI-specialized web media “Ledge.ai (https://ledge.ai/)” AI project consulting
Holding and managing various offline/online events related to AI Operation of “e.g. (https://ledge-eg.com/)” search platform for AI use cases ■ Sony Biz Networks Inc.
Representative: Yasutaka Ogasawara
Location: 1-12-1 Dogenzaka, Shibuya-ku, Tokyo Shibuya Mark City West 23F (general reception 21F)
Telecommunications business based on the Telecommunications Business Law (Notification number/A-24-12875)
Network integration business (planning, design, operation, management, maintenance, sales and rental of network equipment and related equipment)
■ Sony Network Communications Inc.
Representative: Jun Watanabe
Location: 1-7-1 Konan, Minato-ku, Tokyo
Solution service business
Details about this release:
Ledge Co., Ltd.