In order to be able to use artificial intelligence, a company does not necessarily need a qualified specialist. A Fraunhofer study shows how small and medium-sized companies can proceed instead.
Artificial intelligence (AI) helps to optimize production processes and thus save money. However, small and medium-sized companies often lack the expertise to use this future technology. You can collect the data you need, but fail to analyze it. Large cloud providers can help here. They offer simple digital tools that process large data sets and provide AI solutions. Experts speak of “machine learning as a service platforms”. Every company without much experience can get started with artificial intelligence and have models developed that automatically detect faulty workpieces, for example.
Most common use cases on four platforms in comparison
But which platform is suitable for which task? The Stuttgart Fraunhofer Institutes for Manufacturing Engineering and Automation IPA and for Work Management and Organization IAO compared the approaches of the four largest providers – AWS, Google, IBM and Microsoft. They implemented solutions for four use cases that are common in practice and include four categories of data: tabular data, text, image and time series data:
- Customer churn: It is beneficial for hotels to know early on which guests are at risk of cancellation. There may already be a note in the tabular booking data. AI can track it down and develop a corresponding algorithm.
- Text categorization: Texts can be assigned to different categories, such as culture, sport and politics. For example, a press agency can automatically maintain an archive.
- Image recognition: Image analysis plays an important role in production. In this way, defects on the workpiece can be detected with camera systems. AI helps automate this control. The AI learns to recognize errors from a large number of so-called annotated images provided with metadata.
- Tool wear: Replacing a milling head at the right time saves money. If you intervene too early, you give away material, if you intervene too late, you risk a long production downtime. AI learns to interpret the time series data of vibrations and power consumption in order to correctly assess the condition of the milling head.
As a rule, the following applies to AI solutions: the more data are available and the better the quality of the data, the more reliably the model obtained works. When comparing the platforms, the Fraunhofer scientists always chose the most accessible solution. Often only the data records had to be uploaded and annotated: In image processing, this would mean adding the suffix correct or incorrect to each image. The platform then delivered the desired model including the forecast accuracy.
Result
The Fraunhofer study has shown that the solutions from all providers show strengths and do not require in-depth specialist knowledge. Of course there is one difference or another. Some platforms can be operated more intuitively than others. Some AI models only run on the provider’s cloud, others can also be exported and installed on the company’s own servers.
Which platform can be recommended for which application is shown in the study “Cloud-based AI platforms – opportunities and limits of services for machine learning as a service”. It is available for download under the following link: https://www.ki-fortstiegszentrum.de/de/studien/cloudbasierte-ki-plattformen.html
