New neural

In contrast to other AI systems, the neural network integrates newly learned words into existing vocabulary. 

According to Microsoft, “the neural network can learn new words at a much quicker pace than ChatGPT. We’re not only talking about learning the new words but also adding them to its vocabulary for general use.”

Paul Smolensky, a cognitive scientist specializing in language at Johns Hopkins University in Baltimore, Maryland, told Naturethat the neural network’s human-like performance marked a “breakthrough in the ability to train networks to be systematic.”

The technology's capability was shown through people's natural skill to apply newly learned words in different contexts seamlessly, Naturereported. 

Initially, scientists conducted experiments on humans, testing their ability to apply new words in diverse situations. They utilized a learning method involving training an artificial intelligence system in real-time as it completes tasks.

According to Nature, by replicating the patterns of human errors in its training, the neural network became capable of using new words in ways that closely resembled human understanding, surpassing the performance of traditional AI models like ChatGPT. 

Fostering interaction between humans and AI

This advancement shows promise in fostering more organic interactions between humans and AI systems. For example, if an individual comprehends the term 'photobomb,' they can effortlessly apply it in diverse scenarios, such as 'photobomb twice' or 'photobomb during a Zoom call.'

Similarly, comprehending the sentence "the cat chases the dog" enables understanding similar phrases like "the dog chases the cat" without additional effort.

The research aimed to test the network's performance in comparison to human cognition, particularly in using newly acquired words in new settings.

However, some challenges with this interface have been noticed in its implementation, including scaling the technology upwards to handle larger datasets and generalizing its applications to different contexts, such as images. 

Researchers are currently studying human developmental processes to enhance the neural network's systematic generalization abilities.

Elia Bruni, natural language processing specialist at the University of Osnabrück in Germany, told Naturethat this research could make neural networks more efficient learners. 

Consequently, this would decrease the massive volume of data required to train systems like ChatGPT and decrease instances of 'hallucination,' where AI detects non-existent patterns, leading to inaccurate results.

“Infusing systematicity into neural networks is a big deal,” stated Bruni. “It could tackle both these issues at the same time.”