How companies can successfully implement their AI data management projects
The field of artificial intelligence is still considered one of the future-oriented technologies that will help to fundamentally change industries such as medicine, trade, transport or public administration. Projects in which data is to be analyzed with the help of artificial intelligence are particularly popular. However, there are a few stumbling stones to be considered.
So first of all, depending on the problem, the right model has to be found. Another important factor is the data to be used to train an artificial intelligence. Here, for example, questions such as where to get the data from in the first place or how to ensure high-quality test and training data in the first place have to be answered. The so-called 'labeling' is of particular importance. Because data must be labelled correctly. It is also important to avoid overfitting. Here, the right amount and distribution of training data makes the proverbial decisive difference.
Dr. Simone Braun
Head of Business Development
Business Development Manager
Source: Informatik Aktuell, 18th August 2020, authors Dr. Simone Braun, Dan Follwarczny , © Informatik Aktuell / Alkmene Verlags- und Mediengesellschaft mbH, Frankfurt 2020, All rights reserved