English 
Kioxia Corporation said it has developed an AI image recognition tech that can automatically identify products moving through logistics workflows. This is being done together with Tsubakimoto Chain Co. and EAGLYS Inc.
The idea is simple. Logistics needs faster automation. Needs efficiency. Needs something that lets companies deal with changing market conditions without blowing up costs or hurting service quality. KIOXIA AiSAQ and Memory Centric AI sit at the center of this development. These are meant to deal with the problem of scaling AI when product types keep expanding and getting more diverse. The tech will be shown at the 2025 International Robot Exhibition.
Because e commerce keeps rising, logistics networks are carrying more products. And not only more but a much wider mix of them. At the same time, labor shortages stay a problem. Companies are pushed to rely on AI to handle operations efficiently. Traditional image recognition AI uses deep learning models. These models need tuning and retraining every time a new product or a seasonal product shows up. That takes time. It increases power use. It raises operational costs. And the bigger the product catalog, the worse it gets.
こちらもお読みください： 日本、サイバー規制強化でAI、核融合、宇宙を後押し
KIOXIA AiSAQ software and Memory Centric AI tackle these issues. The system stores large volumes of new product data in high capacity storage. Images. Labels. Feature information. All of it. This lets companies add new product data quickly without retraining the original model. But when data grows, searches slow down and memory demands increase. To deal with that, the system indexes the stored data and then moves those indexed sets into SSD storage. This setup makes it quicker to pull the data and it keeps the whole thing running fine even when the amount of data keeps going up.
キオキシア, Tsubakimoto Chain そして EAGLYS will be showing this together at the 2025 International Robot Exhibition. It runs December 3 to 6. The location is Tokyo Big Sight. Their demo will be at the Tsubakimoto Chain booth, which is E6 23. This exhibition is one of the major global events for automation in manufacturing and logistics. Visitors will see the full image recognition system running. As products move on a conveyor, the system captures the image data. Then it classifies the products using the stored features and labels. The demo shows how logistics sites can handle a huge and constantly changing range of products with better accuracy and efficiency.
日本語