« Ada - der intelligente Raum » - Universität und ETH Zürich an der Expo.02


		Fundamental principles
		Sensory organs
		Processing data
			Neural networks
			Brain vs. computer
			Memory
			Learning
			Dreaming
		Interaction

Artificial neural networks

Technical devices emulating individual sensory organs, such as cameras or microphones, are nothing new and have benefited from continual enhancements in performance over the years. The achievement of the "Ada” project consists of linking the components together in an intelligent way. Ada is thus more than just the sum of her individual parts. The data that Ada collects through her artificial sensory organs is not processed by means of conventional software, but in a way similar to the human nervous system. While Ada is made up of normal computers with ordinary operating systems, their organisation and linking and the way the information is processed is new: Ada was programmed as a hybrid network modelled after the way biological nervous systems function, performing neural and digital computations in a computer system.



Sub-section of the neural network of Ada

Ada is the first large-scale neuromorphic system of its kind. Many of the technologies employed have never been integrated in a common system. The software on which Ada is based not only simplifies the developing of neural networks, but also the integration of external devices, like cameras, loudspeakers, microphones etc. In addition, it is possible to study the neural networks in real time, that is, being able to simultaneously record and evaluate data and display the results with no time lag.

Neural network

In the neural network, there is no paramount authority issuing simultaneous commands to all the different functional units. Rather, the functional units adjacent to one another communicate much more extensively with each another than with those further apart. Communication thus by and large occurs locally. By contrast, the conventional computer features hierarchical organisation: A superordinate authority can simultaneously contact all points.

In the neural network, communication operates from functional unit to functional unit, as from nerve cell to nerve cell in the nervous system. One can think of dominos, with the signal transmitted from one domino tile to the next as they successively topple over.





		-	Neural network
		-	hybrid
		-	neuromorphic



		-	mantik
		-	Neural networks I
		-	Neural networks II
		-	Genesis
		-	Tutorial for programming

Technology that can learn



Sub-section of the neural network of Ada

Neural networks are information processing systems that are able to learn. Based on preceding examples, they can learn to choose the best possible solution for a particular situation and continue to refine it accordingly. Like a brain, a technical neural network is a complex system of autonomous parts, so-called neurons or units, whose local interactions yield global patterns of activity («answers»). The dynamics of the brain is not centrally directed by programs as with computers, but is capable of organising itself. Like brains, technical neural networks are flexible and able to learn, tolerating deviations from the norm and making use of parallel signal processing.

The neural networks developed for Ada are based on models that reflect the way real brains function. There are countless models explaining how the nerve cells in specific regions of the brain communicate with one another. All such models make assumptions as to how a specific brain region exchanges information with other regions of the brain. Yet today we are still far from fully understanding how the various regions of the brain all work together. Even less is known about how to link together the various available models so the whole can function as a unit. Ada represents a first step in this direction.