How Google’s AI is Helping Neuroscience Evolve
How Google’s AI is Helping Neuroscience Evolve
Artificial intelligence is a concept that has been dreamt about for decades now. You’ve had the chance to see robots in books, movies, and other forms of science-fiction entertainment. Thanks to decades of research and the evolution of technology, AI is becoming more science than fiction.
We’re getting closer to developing machines that can have human characteristics. Even though we won’t witness this for quite some time, there’s no doubt that technology is moving in that direction.
The problem is, before we make robots with brains similar to those of humans, we need to understand how the human brain works. Even though we can see it, and understand some of it, there’s still a lot of things to be discovered about it.
As one of the world’s biggest tech companies, Google has always wanted to be the leader in AI, as the first company to bring this technology to fruition will likely dominate the tech sector, if not the world. Towards this end, Google’s research aims to understand more about the human brain so that the knowledge can be applied to AI. This is, of course, music to the ears of neuroscientists.
Visualizing Neural Pathways
To put it bluntly, neuroscientists have struck out trying to understand the mechanism of neuron interactions and neural activation patterns. Even though some progress has been made, there’s still a lot of work to do before they can explain exactly how neurons work.
On the other hand, Google’s technology is striving towards making this happen. Their artificial neural network should be able to create a comprehensive model of every synapse and neuron in the brain.
Teams at Google have been working on training the artificial neural network to recreate a 3D model of the brain’s neural pathways. The AI processed 663 gigabytes of images containing zebra finch’s brain, and successfully constructed a 3D model of every synapse and neuron.
How Does It Work?
Obviously, the complete mechanism is incredibly complex, so let’s break it down:
Electronic microscope images of the brain were sent to Google in 2012. The team used AI to compile a 3D image consisting of thousands of 2D ones. When all of those pictures stack onto one another, they form a 3D model.
The reason why this couldn’t be done effectively before is the amount of data. Just imagine what 663GB of data looks like. The neurons could not have been seen in detail because of other data that tangles it.
According to Google’s estimates, it would take neuroscientists around 100,000 hours to extract the images of neurons from a cube-shaped sample of 1mm in size. On the other hand, Google’s AI has been trained to do this in seven days.
The algorithm automated the process of slicing and extracting important data. The way it works is by looking for neurons in each image and tracing them through the sample.
This isn’t the first time that a machine has been able to do this. However, Google’s solution is far more capable and accurate than all previous approaches. Google’s researcher Viren Jain states that the breakthrough happened when they managed to teach the AI to look for one neuron structure at a time, instead of trying to make it look for every neuron at once.
How Does This Impact The Future Of Neuroscience?
It’s safe to say that this technology will prove very useful in the field of neuroscience. Further development of this technology will bring neuroscientists much closer to understanding how neural pathways actually work.
It took Viren Jain and his team 12 years to develop this AI solution. Now that they’ve finally managed to apply it to neuroscience, it’s only a matter of time before we’re able to get an in-depth look at brain functions that have thus far eluded us.
Of course, this also means that we’re getting closer to seeing AI develop to the point where tech companies will be able to create machines similar to humans.
Is Science Fiction Our Future Reality?
The answer to this is – very likely. Robots that have features of humans are already being made, and the technology continues to develop as you’re reading this. We’re witnessing a primitive version of humanized robots that seem to get smarter pretty quickly.
However, this doesn’t mean that this will happen in the near future. The human brain is incredibly complex, so there’s a chance that decades will pass before it is fully understood.
The main reason behind this is human psychology. No AI has been able to have most of the features that make us human. In order to understand how the brain works, we need to understand the mind, and that’s still out of science’s reach.
There are many predictions about when AI will evolve enough to be considered similar to humans. Some believe that this won’t happen before 2080’s, which could very well be a realistic prediction, considering the sizable gap in knowledge.
Most experts agree that robots are highly unlikely to express emotions the way humans do. This makes sense, since emotions are hard to understand and turn into something that you can teach to a machine. Nevertheless, there’s no doubt that AI will keep improving, and that we’ll be able to see some pretty impressive machines in the future.
The Final Word
As you can see, the way that technology is evolving can help us understand a lot about the human brain. What was once considered impossible to grasp, it is just a matter of research and time today.
Neuroscience will only continue to advance, inching closer and closer to answering all remaining questions about the human brain.
And no, there’s no reason to believe those science fiction writers who fantasize about machines taking over the world. The technology that is being developed serves the purpose of helping us understand the world that we live in better, and it can only be of use to us. The only thing left to do is to wait in excitement for the new breakthroughs.