Neuromorphic Computing heralds the next generation of artificial intelligence. While the first generation of thinking machines was extremely rule-based and used pure logic to draw conclusions, this new type of artificial intelligence goes a step further.

The pure conclusion based on certain rules is useful for monitoring processes and improving their efficiency. However, to bring real benefits to Artificial Intelligence, it must go beyond that. The next generation will therefore expand into areas that correspond to human cognition, such as interpretation and autonomous adaptation. This will enable artificial intelligence to abstract problems in the future and also to better handle new situations. Through this step, automation can be advanced and deep learning algorithms can work better.

 

Pioneer of Neuromorphic Computing

Pioneers of Neuromorphic Computing are the Intel Labs, which have placed a research focus on neuromorphic research. In doing so, they imitate the neuronal structures and functions of the human brain and transfer them into algorithmic approaches. Topics of particular interest are, for example, ambiguity and contradictions with which the machines could be confronted.

The Neuromorphic Computing Chip Loihi (Source: Intel Labs)

 

To better test these functions, Intel has developed the fifth generation self-learning neuromorphic research test chip, “Loihi”, for this purpose. The Loihi chip contains a total of approximately 130,000 neurons that can communicate and learn from each other.

This chip is to be made available to the entire research society for a wide range of tests in order to advance neuromorphic computing as an interdisciplinary challenge.

LEAVE A REPLY

Please enter your comment!
Please enter your name here

This site uses Akismet to reduce spam. Learn how your comment data is processed.