Neuronal network capable of understanding spoken words better than humans

2000/01/01 Elhuyar Zientzia Iturria: Elhuyar aldizkaria

• Ingeniaritza
• Patenteak/Asmakizunak

The Department of Biomedicine of the University of Southern California has created a computer capable of understanding words, which offers great advantages with systems developed to date.

In fact, this computer runs through a neural network-based system and is able to recognize words from any voice. This system is the first independent speech recognition system of the speaker and can be of great use in different fields, especially in the field of information technologies: voice-controlled computers, media for deaf people, instant interview printing systems, etc.

Neural networks are computer systems based on the functioning of the human brain that repeat the brain system to process information. Like humans, they are able to classify similar information into groups (for example, a word given by different people and in different situations) and treat similar information from the same group. Conventional neural networks consist of information processors called “neurons”, connected to each other forming a network with thousands and thousands of interconnections.

New neural network T. Berger and J. Developed by Dr. Liaw, it has yielded excellent results in word recognition tests. The system is able to understand voices and words that are not separated by the human ear, even if there is a lot of noise or the signal is very weak. In addition, the neural network of Berger and Liawa has only 11 neurons and about 30 links.