A team of researchers from Oxford University is working on a new technique which would have the capability to synthesis human-like thoughts in AI machines. The new method which involves language guided imagination (LGI) network, could pave the way for AI machines to think like humans led by language.
The cognition process involves two things, firstly, the brain needs to grasp a specific language expression and secondly, it must utilize the information to develop ideas.
The team is currently attempting to create Natural Language Processing (NLP) tools that can answer to queries like the way humans do. But, these tools will not understand in a similar manner and with a similar depth as humans; because human brains are trained in a different way that changes as their brain develops.
Two researchers, Wenchuan Wu, and Feng Qi have developed an artificial neural based on the prefrontal cortex, a part of the brain in order to produce human-like thinking pattern in machines. The LGI system of the researchers comprises of three main components: a language system, a vision system, and a synthetic prefrontal cortex.
The vision system has an encoder inside it. The encoder is responsible for separating the information received by the network or imagined situations into abstract population representations.
Coming to the language system, it is also known as the sub-system. It is responsible for imitating a function of the human brain by taking quantity information and afterwards, converting the information into text symbols.
The synthetic prefrontal cortex mimics the cortex of the human brain.