Inspired by techniques used to train deep neural networks, a neuroscience professor has advocated a new theory of dreams: the overfitted brain hypothesis.
The hypothesis, by Erik Hoel of Tufts University, suggests that the strangeness of our dreams serves to help our brains better generalize our everyday experiences.
“Obviously, there are an incredible amount of theories as to why we dream,” says Hoel. “But I wanted to draw attention to a theory of dreams that takes the dream itself very seriously, which says that the experience of dreams is the reason why you are dreaming.”
A common problem when it comes to training AI is that it becomes too familiar with the data it trains on; you start to assume that the training set is a perfect representation of anything you can find. Data scientists get around this by introducing some chaos to the data; In one of these regularization methods, called “abandonment,” some data is randomly ignored.
Imagine if black boxes suddenly appeared on the internal screen of an autonomous car – the car that sees the random black boxes on the screen and focuses on the general details of its surroundings, rather than the specific details of that driving experience in In particular, it will probably be better to understand the general driving experience.
“The original inspiration for deep neural networks was the brain,” says Hoel. And while comparing the brain to technology isn’t new, he explains that using deep neural networks to describe the overfitted brain hypothesis was a natural connection. “If you look at the techniques that people use in deep learning regularization, it is often the case that those techniques have some striking similarities to dreams,” he says.
With that in mind, his new theory suggests that dreams make our understanding of the world less simplistic and more complete, because our brains, like deep neural networks, also become too familiar with the “training set” of our everyday lives. . . Hoel’s theory is exposed in a review of the magazine. Patterns.
To counteract familiarity, he suggests, the brain creates a strange version of the world in dreams, the mental version of abandonment. “It is the very strangeness of dreams in their divergence from waking experience that gives them their biological function,” he writes.
Hoel says there is already evidence from neuroscience research to support the overfitted brain hypothesis. For example, it has been shown that the most reliable way to provoke dreams about something that happens in real life is to repeatedly perform a new task while you are awake. He argues that when you overtrain on a new task, the overfitting condition is triggered and your brain then tries to generalize to this task by creating dreams.
But he believes that research could also be done to determine if this is really why we dream. He says that well-designed behavioral tests could differentiate between generalization and memorization and the effect of lack of sleep on both.
Another area that you are interested in exploring is the idea of ”artificial dreams.” The overfit brain hypothesis occurred to him while thinking about the purpose of works of fiction such as movies or novels. Now, he hypothesizes that external stimuli, such as novels or television shows, could act as “surrogates” for dreams, and that perhaps they could even be designed to help delay the cognitive effects of sleep deprivation by emphasize your dream nature (for example, using reality technology).
While you can just turn off learning in artificial neural networks, Hoel says, you can’t do that with a brain. Brains are always learning new things, and that’s where the overfitted brain hypothesis comes into play. “Life is boring sometimes,” he says. “Dreams are there to prevent you from adapting too much to the model of the world.”
Fountain: Cell press
SHARE this dream story on social media …