Verification before the introduction of “Azure Machine Learning” in pig shipment forecasting

Small-scale pig farmers gathered to do the sixth pig industrialization

Headquartered in Shibukawa City, Gunma Prefecture, Global Pig Farm is a company created with the aim of bringing together family-owned pig breeders to carry out the highest-level technological innovation in Japan. We have achieved sixth industrialization

to its own brand “Wabuta Mochibuta” by continuously undertaking breeding (breeding improvement), feed design, production management, transportation and sales.

* 6th industrialization means working across all major industries, agriculture, forestry and fisheries, secondary industry, manufacturing and tertiary industries, sales and service.

Currently, in the livestock industry, forecasts for livestock using AI (artificial intelligence) are being made. On the other hand, the pig industry has lower unit cost than cow raising, so it is difficult to build systems with high cost. Therefore, the breeders decided on the transport date based on past experience, arranging the transport and adjusting the number of animals. However, that will depend on a particular person for the experience and know-how, and there is a problem in expanding the business.

Use cloud-style AI services

Microsoft’s cloud-based AI service “Azure Machine Learning” was chosen by Global Pig Farm to address the cost problem of introducing AI systems and standardizing individual shipping forecasting operations.

By providing AI systems in the cloud, the service was introduced at a lower cost than when it was built on its own. In addition, various algorithms (computational and processing methods that lead to solutions to problems and problems) are prepared in advance, which can therefore be introduced in a short period of time. Furthermore, because this is a cloud service that uses a powerful Microsoft AI system through the Internet, you can use it as much as you want, when you need it and save money.

[Bài viết liên quan]Quick detection of signs of disease in cows Modern livestock management methods use artificial intelligence

Validated with 10 years of shipping data

When introducing the AI ​​system, Knowledge Communication Co. Ltd., which provides business support using cloud AI, conducts pre-rollout verification (PoC) for the first time using operational data moved in the past 10 years.

Knowledge Communications has modeled transport prediction using AI systems by combining actual shipping data with external data, such as weather data that could affect pig shipments. As a result, we were able to obtain the following results.

・ AI system allows for prediction with error of at least one week.

Experimental demonstration of the actual relevance of the data supposedly related to the transit time.

-Set up the method (data cleansing method) to build the correct model from variable data.

・ Building linkage with existing systems will be completed in about one and a half months.

・ Pre-introduction verification is completed in approximately 3 months.

In the future, we will further promote the introduction of IoT and AI for swine production in cooperation with shipping forecasting, such as real-time temperature data collection of pig farms and management air machine.

Using the latest AI technology, anyone can transport pigs more accurately, which often relies on the experience of the person in charge. The use of AI, IoT, etc. has the potential to significantly expand the capabilities of agriculture.