In podcast episode 2, Madeleine Mickeleit talks to Kerim Galal, CEO of InnoSEP GmbH from Hannover. Kerim talks about the possibilities of low-code & artificial intelligence and presents some use cases from practice.
Podcast episode summary
The big advantage of using a low-code platform is easier access to artificial intelligence. A low-code development environment enables the development of applications without deep knowledge of the programming material. Low-Code is based on the principle of prebuilt code modules. From these, a software with the desired properties is compiled. The actual development takes place via a graphical interface and is intuitive. At the same time, this enables very fast and cost-effective application development.
The demarcation between artificial intelligence and machine learning takes place through the task domain that the finished application is intended to perform. Artificial Intelligence is the umbrella term and Machine Learning is a subset of it. Machine learning systems do not rely on a specific data set as input. Rather, these learn from training data and use algorithms to predict outcomes.
Thus, the combination of machine learning and low-code platform ensure dynamic and innovative developments in IIoT. This is where the experience of industry meets the expertise of IT developers. In collaboration, systems are then created that take on very specific tasks in Industry 4.0. The technical input, which requirements and tasks the system must perform, comes from the industry. The development is then handled by an IT service provider, such as InnoSEP GmbH in this example.
There are thus virtually no limits to the practical implementation of the combination of low-code and AI. There are a particularly large number of possible applications in the Internet of Things sector. The innovative impetus in each case comes from the industry itself. Machine learning can manage data from very different sources. In InnoSEP GmbH’s use cases, vibration sensors or camera images serve as input. In collaboration with IT experts, digitization then succeeds on a whole new level.
The decision as to whether machine learning can be usefully employed is analyzed individually on a case-by-case basis. It is important that there is physical detectability. Concrete measurement data that can be evaluated and analyzed by a computer system form the basis for an AI system. Another prerequisite is a broad data infrastructure. Artificial intelligence can then be implemented as a solution based on comprehensive measurement data.
Artificial intelligence is well suited for evaluating physical parameters that are common in the Internet of Things. In InnoSEP GmbH’s use cases, these are, for example, data from vibration sensors or images of components in incoming inspection. Based on the learned parameters, the AI system decides on the quality of the measurement data. The system makes a “good-or-bad” decision based on experience.
Cloud computing also plays an important role in this context. Kerim Galal of InnoSEP explains that connectors make it easy to establish a direct link between the industry and the IT service provider. The measurement data is transferred to the cloud via these interfaces and processed further on the InnoSEP platform. If desired, the infrastructure for the analyses and calculations can also be provided by InnoSEP. For example, infrastructure requirements in the industry have been lowered thanks to cloud computing, making it easier to get started with such a system.
Scalability and further development are two other issues that are important in this approach. Often the requirements change over time or the evaluation is to become more complex and extensive. Here again the advantages of a low-code platform come into play. Since low-code development is relatively simple, the industry customer can have a hand in the application itself. In this way, new features can be added or data modeling can be advanced. As a rule, these are ongoing processes, since hardly any system directly provides all the desired functions. Data validation, feedback and fine-tuning, and further development are iterative processes that can continue for an extended period after implementation.
The future development of AI and low-code will further drive digitization and automation in the Internet of Things. An edge device enables connections to areas that at first glance have very little to do with IT. InnoSEP’s CEO addresses a use case in which machine learning is used to assess and predict the health of farm animals. In this context, IT technologies serve as tools that are adapted to the respective tasks in cooperation with experts in the field from the industry. Thus, machine learning has the potential to become a key technology of the future.
More about this implementation partner
“Be your own Data Scientist” – With our Machine Learning Platform we combine data processing, expertise and Artificial Intelligence and enable “Code-Free” Machine Learning for industry and the scientific environment to accelerate your business processes. The application of Artificial Intelligence is currently reserved for Data Scientists or specialists with programming skills and expertise in Machine and Deep Learning. However, the need for specific and scalable data analytics solutions in technical and scientific departments is very high and the necessary expertise is available. In addition, it is not possible for every company to train their professionals as Data Scientists or to create their own Data Science team internally. Thus, our vision is to realize a bridge between expert knowledge and applied machine learning. The solutions from our machine learning platform can be integrated into your infrastructure and are applicable across industries and for a wide range of industrial applications. Since its foundation in 2016, the company has been implementing projects in the fields of predictive maintenance, anomaly detection and pattern recognition in quality assurance for premium customers from the automotive and automation industries, among others.