Why is Bayesian machine learning always generative

Self-learning systems: how does machine learning work?

Machine learning already has important functions for marketing. Currently, however, it is primarily large companies that use their technologies internally, above all Google. Self-learning systems are still so new that they cannot simply be bought as an out-of-the-box solution. Instead, the major Internet providers are developing their own systems and are thus initiators in this area. Since some of them have one, despite the commercial interest Open source approach and cooperate with independent science, developments in the field are picking up speed.

In addition to the creative side, marketing always has one analytical aspect: Statistics on customer behavior (buying behavior, number of visitors to websites, use of apps, etc.) play a major role in the decision for certain advertising measures. The larger the amount of data, the more information can usually be extracted from it. Intelligent programs are required to process such a mountain of features. This is where self-learning systems come in: The learned computer programs recognize patterns and can make well-founded predictions, which people who are fundamentally biased with data are only capable of to a limited extent.

As a rule, an analyst approaches measurement data with certain expectations. These prejudices are hardly avoidable for humans and often cause the results to be biased. The larger the amount of data the analysts process, the greater the deviations are likely to be. Intelligent machines can also have prejudices because people have trained them inadvertently, but they are more objective when it comes to hard facts. Machines therefore usually deliver more meaningful analyzes.

Self-learning systems also improve and facilitate the presentation of the analysis results: Automated data visualization is the name of a technology in which the computer independently selects the correct representation of data and information. This is particularly important so that people understand what the machine has discovered and predicted. In the great flood of data, it becomes difficult to present measurement results yourself. Therefore, the visualization must also run on the calculations of the computer.

But machine learning can also influence content creation - keyword: generative design. Instead of designing the same customer journey for all users - i.e. the steps that the customer goes through to purchase a product or service - dynamic systems can design individual experiences based on machine learning. The content that is displayed to the user on a website is still provided by copywriters and designers, but the system assembles the components specifically for the user. In the meantime, self-learning systems are also used to design oneself: With the Project Dreamcatcher, for example, B. possible to have components designed by a machine.

Machine learning can also be used to e.g. B. Chatbots to design better. Many companies are already using programs that handle part of their customer support via a chatbot. But in many cases users are quickly annoyed by the machine employees: The capabilities of the current chatbots are usually very limited and the answer options are based on manually maintained databases. A chatbot that is based on a self-learning system and has good speech recognition (NLP) can give customers the feeling of really communicating with a person - and thus pass the Turing test.

Amazon or Netflix demonstrate another important development in machine learning for marketers: recommendations. A big factor in the success of these providers is predicting what the user might want next. Depending on the data collected, the self-learning systems can recommend further products to the user. What was previously only possible on a large scale ("Our customers like product A, so most of them will also like product B."), Is now also possible on a small scale thanks to modern programs ("Customer X liked products A, B and C, which is why you will probably like product D too. ").

In summary, it can be stated that self-learning systems will influence online marketing in four important ways:

  1. quantity: Programs that work with machine learning and have been well trained can process huge amounts of data and thus make predictions for the future. Marketing experts can thus draw better conclusions about the success or failure of campaigns and measures.
  2. speed: Analyzes cost time - if you have to do them by hand. Self-learning systems increase the working speed and you can react more quickly to changes.
  3. automation: Machine learning makes it easier to automate processes. Since modern systems can independently adapt to new conditions with the help of machine learning, complex automation processes are also possible.
  4. individuality: Computer programs can serve countless customers. Since self-learning systems collect and process data from individual users, they can also provide comprehensive support to these customers. Individual recommendations and specially developed customer journeys help to use marketing more effectively.