State space (neural network)

The state space ( Engl. StateSpace ) referred to in the cognitive science the whole one all states, artificial intelligence or cognitive can take biological system. A distinction is made between search and problem-solving spaces.

From a mathematical point of view, the state space of an artificial neural network is a vector space that describes the set of possible states of the neural network. The individual activation states are represented here by vectors . Vectors also form the values ​​of internal status variables.

The sequence of the previous entries can serve as an example. Each input value is represented by a component of the input vector.

Visual state spaces have a special meaning in robotics in the development of navigation systems . The objects in the environment that are essential for navigation (e.g. doors, walls, etc.) are entered in a coordinate system and displayed on a status monitor during navigation. Both artificial and natural landmarks are used that can be recognized by the technical system via a video camera. This procedure ensures a robust detection of obstacles.