Machine learning

Make big data manageable!

Machine learning (ML) is a branch of artificial intelligence. By recognizing patterns in existing databases, IT systems are able to find solutions to problems independently.

Prototypen und Demonstratoren vorhanden

Branchenübergreifender Einsatz

KMU geeignet?

The current status

A large number of machine learning implementation and development projects can currently be found. The deep learning sub-category in particular brings numerous new developments, e.g. in the areas of facial and speech recognition.

If you would like to start your own development project, we recommend using the Python programming language. The open source availability of numerous programming libraries such as TensorFlow from Google, scikit-learn or Theano enables the rapid integration and utilization of machine learning techniques.

Recognizing patterns in large amounts of data requires both high storage capacity and advanced knowledge in the field of machine learning. Cloud providers (Amazon, IBM, Google, Microsoft, etc.) now offer pattern recognition as “machine learning as a service” within their platforms.

Technology and use

Technology description

As the fourth industrial revolution progresses, more and more company and process-related data is being collected. The evaluation and assessment of these is therefore becoming increasingly difficult for people to handle. Based on existing databases, machine learning is used to identify regularities and patterns within the structures and thus enables the development of solutions and conclusions.

The subcategory of deep leaning also describes a method for generating artificial neural networks. These are modeled on a human brain in terms of their function and structure. By reading large data sets, important features can be automatically extracted and classified. Intelligent evaluation and predictions are therefore possible.

Image: Based on Towards Data Science

Possible application scenarios

There are currently numerous applications relating to machine learning. Machine learning can be used wherever large amounts of data need to be analyzed and searched for patterns.

In everyday life, the best-known applications are probably the recommendation services of Amazon and Netflix, the sorting of spam emails, voice recognition (Siri, Cortana) or facial recognition (Facebook).

In the production environment, machine learning can also be used in the areas of quality assurance, customer satisfaction, autonomous systems and predictive maintenance.

There are also initial areas of application in the office environment. With chatbots you can, for example conduct communication. You can make inquiries and receive corresponding information, recommendations for action or suggested solutions.

Step-by-step introduction

Step 1: Definition of the application scenario

Draw up a concrete target description. What do you hope to achieve by using machine learning? Define criteria that are decisive for solving the problem.

Step 2: ML mapping

Break down your problem in such a way that it can be processed by machine learning methods and algorithms. Reduce the overall complexity to individual, easily manageable problem cases.

Step 3: Process data

Data collection:

A crucial step is to ensure that the data structure is suitable. Are you currently recording the correct data/characteristics?

Data exploration:

Based on your experience, carry out an assessment of the data to determine whether the data has been processed in such a way that it can be analyzed. This step is particularly important in order to gain an understanding of which categorizations of the data you should make and which input variables are required for the algorithms. If necessary, you can carry out initial investigations to determine which changes result from the variation of individual data.

Step 4: Modeling

Based on the problem, different algorithms for data evaluation can be considered (decision trees, random forest, etc.). If you already have skills in your company, you can now implement machine learning tasks. If not, it is advisable to commission experts, but also to use machine learning-as-a-service functionalities in cloud platforms if necessary.

Step 5: Evaluation

The ML model must now be continuously monitored in its application. It must be checked whether the data records have been evaluated correctly with regard to the objective. Adjusting screws that lead to an improvement in results:

– Increase in data volume

– Clear categorization

– Reducing complexity

– Improving data quality

Opportunities for SMEs

Recognition of complex relationships

Greater efficiency

Identification of optimization potential

Making predictions

Contact us

Do you need support with the introduction in your company?

Get in touch with us!

Keep an eye on the most important SME-relevant technologies with our technology radar!

Back to the technology radar