Fields of innovation
Self-aware and self-developing systems – novel machinery
Cooperative machines will serve us responsibly
The science fiction vision of machines building machines is approaching reality. First, machines have to exhibit intelligence, autonomy, evolution, collaboration, and responsibility. A common research agenda between academia and industry is necessary to develop these traits in machines and maintain a competitive lead in cognitive system development.
Machines are increasingly being endowed with more intelligent capabilities. Eventually, machines will even have self-awareness and the ability to develop themselves, otherwise known as self-development. The combination of self-awareness and self-development is referred to as self-x, and it is the trajectory for machines that will comprise systems relevant to our business.
will be self-aware and capable of developing themselves further
Self-awareness and self-development
The two pillars of self-x each include several traits that we expect machines to adopt in the future. Properties of self-awareness include self-monitoring, self-explaining, self-certification, self-pricing and self-understanding. These traits will enable machines to reflect on their own status, explain their decisions, validate their own integrity, negotiate values, and consider their experiences to draw conclusions.
The traits of self-development include self-limitation, self-healing, self-organizing, self-learning and self-extension. They will make it possible for machines to limit their own features in the interest of collaboration, recover from system failures, include and exclude systems in cooperative efforts, learn from their own experiences or that of other systems, and extend their own application with a context of purpose.
Life with advanced machines
The machines we interact with in the future will appear to us to be intelligent, cooperative and responsible. Their knowledge of users, from demands to the situational context, will allow them serve in a way that demonstrates purpose. At Bosch, we anticipate that by the year 2025, 50 percent of the systems that are responsible for generating our revenue will have self-x capabilities, underscoring the importance of these developments to our business.
Importantly, as machines begin to mimic abilities that are currently the domain of humans alone, interesting questions arise. For example, what are the ethical and social values of these systems? How can we create trust in autonomous systems in a human-machine collaboration environment? How can we guarantee systems’ integrity when they change structures, features, knowledge and behavior independently?
These questions comprise both technical and philosophical perspectives and their answers will influence the way we interact with self-x machines.
Technology to-do list
A number of technologies are needed to bring these capabilities to life. They include explainable artificial intelligence, unsupervised learning, digital clone technology, verification and validation for machine learning, knowledge representation and resilient and self-validating architectures. Very likely, we will need additional technologies that are not yet defined to fully realize self-x capabilities.
In order to develop these technologies, we need to adopt a common research agenda in our industry in cooperation with academic efforts. Doing so will help us maintain a competitive lead in cognitive system development and ensure the horizon of possibilities for Bosch remains bright.
Bosch envisions a day with self-aware and self-developing machines. We are focused on developing the technologies that will revolutionize machines as we know them.