There are many researches going on to make the machines as smart as humans. In some of the researches, it is proved that machines can perform the task better than human or in other words they can outperform humans in some of the predictive analysis.
Last year, an Artificial Intelligence Algorithm was accurate to identify breast cancer. It is also performing well in self-driving cars. Some of the companies are using Artificial Intelligence and Machine learning in various aspect.
Now, the University of Oxford published new research in which they tried to recreate the human-like thinking pattern in machines.
The purpose of the research is to build a human-like neural network by removing the neuroscience unsupported components from AI architecture while introducing the novel neural mechanism and algorithm into it.
Human-like thinking pattern is based on the mental ideas that are guided by the languages to achieve a goal. For example: if you are leaving a house and saw a heavy rain before, then you would think of having an umbrella before leaving your house to avoid getting wet if the rain comes.
Now, while thinking the above example, how a human mind think? It is like, drops of water in huge amount is rain while holding an umbrella is something that doesn’t let the water come to you.
This continual thinking capacity distinguishes us from the machine, even though the latter can also recognize images, process language, and sense rain-drops. Continual thinking requires the capacity to generate mental imagination guided by language, and extract language representations from a real or imagined scenario.
The researchers proposed a Language guided imagination (LGI) network to incrementally learn the meaning and usage of numerous words and syntaxes, aiming to form a human-like machine thinking process.
The LGI contains three subsystems
- A vision system that contains an encoder to disentangle the input or imagined scenarios into abstract population representations, and an imagination decoder to reconstruct imagined scenario from higher level representations.
- Language system, that contains a binarizer to transfer symbol texts into binary vectors, an IPS (mimicking the human IntraParietal Sulcus, implemented by an LSTM) to extract the quantity information from the input texts, and a textizer to convert binary vectors into text symbols.
- A PFC (mimicking the human PreFrontal Cortex, implemented by an LSTM) to combine inputs of both language and vision representations, and predict text symbols and manipulated images accordingly.
Feng Qi and Wenchuan Wu, the two researchers carried out the recent study.
Qi and Wu evaluated their LGI network in a series of experiments and found that it successfully acquired eight different syntaxes or tasks in a cumulative way. Their technique also formed the first ‘machine thinking loop,” showing an interaction between imagined pictures and language texts. In the future, the LGI network developed by the researchers could aid the development of more advanced AI, which is capable of human-like thinking strategies, such as visualization and imagination.
“LGI has incrementally learned eight different syntaxes (or tasks), with which a machine thinking loop has been formed and validated by the proper interaction between language and vision system,” the researchers wrote. “Our paper provides a new architecture to let the machine learn, understand and use language in a human-like way that could ultimately enable a machine to construct fictitious mental scenarios and possess intelligence.”
For this experiment, researchers gave some input text and input movie and after 5000 steps training, LGI could not only reconstruct the input image with high precision but could also predict the ’mentally’ moved object with specified morphology, correct manipulated direction and position just after the command sentence completed.
The research paper is published arXiv.org with the title “Human-like machine thinking: Language guided imagination”
Tags: machine learning,machine learning algorithms,machine learning applications,machine learning examples,human-like thinking,machine learning algorithms, SPOKEN by YOU