Connect with us

TOP SCEINCE

AI system self-organizes to develop features of brains of complex organisms

Published

on

AI system self-organizes to develop features of brains of complex organisms


Cambridge scientists have shown that placing physical constraints on an artificially-intelligent system — in much the same way that the human brain has to develop and operate within physical and biological constraints — allows it to develop features of the brains of complex organisms in order to solve tasks.

As neural systems such as the brain organise themselves and make connections, they have to balance competing demands. For example, energy and resources are needed to grow and sustain the network in physical space, while at the same time optimising the network for information processing. This trade-off shapes all brains within and across species, which may help explain why many brains converge on similar organisational solutions.

Jascha Achterberg, a Gates Scholar from the Medical Research Council Cognition and Brain Sciences Unit (MRC CBSU) at the University of Cambridge said: “Not only is the brain great at solving complex problems, it does so while using very little energy. In our new work we show that considering the brain’s problem solving abilities alongside its goal of spending as few resources as possible can help us understand why brains look like they do.”

Co-lead author Dr Danyal Akarca, also from the MRC CBSU, added: “This stems from a broad principle, which is that biological systems commonly evolve to make the most of what energetic resources they have available to them. The solutions they come to are often very elegant and reflect the trade-offs between various forces imposed on them.”

In a study published today in Nature Machine Intelligence, Achterberg, Akarca and colleagues created an artificial system intended to model a very simplified version of the brain and applied physical constraints. They found that their system went on to develop certain key characteristics and tactics similar to those found in human brains.

Instead of real neurons, the system used computational nodes. Neurons and nodes are similar in function, in that each takes an input, transforms it, and produces an output, and a single node or neuron might connect to multiple others, all inputting information to be computed.

In their system, however, the researchers applied a ‘physical’ constraint on the system. Each node was given a specific location in a virtual space, and the further away two nodes were, the more difficult it was for them to communicate. This is similar to how neurons in the human brain are organised.

The researchers gave the system a simple task to complete — in this case a simplified version of a maze navigation task typically given to animals such as rats and macaques when studying the brain, where it has to combine multiple pieces of information to decide on the shortest route to get to the end point.

One of the reasons the team chose this particular task is because to complete it, the system needs to maintain a number of elements — start location, end location and intermediate steps — and once it has learned to do the task reliably, it is possible to observe, at different moments in a trial, which nodes are important. For example, one particular cluster of nodes may encode the finish locations, while others encode the available routes, and it is possible to track which nodes are active at different stages of the task.

Initially, the system does not know how to complete the task and makes mistakes. But when it is given feedback it gradually learns to get better at the task. It learns by changing the strength of the connections between its nodes, similar to how the strength of connections between brain cells changes as we learn. The system then repeats the task over and over again, until eventually it learns to perform it correctly.

With their system, however, the physical constraint meant that the further away two nodes were, the more difficult it was to build a connection between the two nodes in response to the feedback. In the human brain, connections that span a large physical distance are expensive to form and maintain.

When the system was asked to perform the task under these constraints, it used some of the same tricks used by real human brains to solve the task. For example, to get around the constraints, the artificial systems started to develop hubs — highly connected nodes that act as conduits for passing information across the network.

More surprising, however, was that the response profiles of individual nodes themselves began to change: in other words, rather than having a system where each node codes for one particular property of the maze task, like the goal location or the next choice, nodes developed a flexible coding scheme. This means that at different moments in time nodes might be firing for a mix of the properties of the maze. For instance, the same node might be able to encode multiple locations of a maze, rather than needing specialised nodes for encoding specific locations. This is another feature seen in the brains of complex organisms.

Co-author Professor Duncan Astle, from Cambridge’s Department of Psychiatry, said: “This simple constraint — it’s harder to wire nodes that are far apart — forces artificial systems to produce some quite complicated characteristics. Interestingly, they are characteristics shared by biological systems like the human brain. I think that tells us something fundamental about why our brains are organised the way they are.”

Understanding the human brain

The team are hopeful that their AI system could begin to shed light on how these constraints, shape differences between people’s brains, and contribute to differences seen in those that experience cognitive or mental health difficulties.

Co-author Professor John Duncan from the MRC CBSU said: “These artificial brains give us a way to understand the rich and bewildering data we see when the activity of real neurons is recorded in real brains.”

Achterberg added: “Artificial ‘brains’ allow us to ask questions that it would be impossible to look at in an actual biological system. We can train the system to perform tasks and then play around experimentally with the constraints we impose, to see if it begins to look more like the brains of particular individuals.”

Implications for designing future AI systems

The findings are likely to be of interest to the AI community, too, where they could allow for the development of more efficient systems, particularly in situations where there are likely to be physical constraints.

Dr Akarca said: “AI researchers are constantly trying to work out how to make complex, neural systems that can encode and perform in a flexible way that is efficient. To achieve this, we think that neurobiology will give us a lot of inspiration. For example, the overall wiring cost of the system we’ve created is much lower than you would find in a typical AI system.”

Many modern AI solutions involve using architectures that only superficially resemble a brain. The researchers say their works shows that the type of problem the AI is solving will influence which architecture is the most powerful to use.

Achterberg said: “If you want to build an artificially-intelligent system that solves similar problems to humans, then ultimately the system will end up looking much closer to an actual brain than systems running on large compute cluster that specialise in very different tasks to those carried out by humans. The architecture and structure we see in our artificial ‘brain’ is there because it is beneficial for handling the specific brain-like challenges it faces.”

This means that robots that have to process a large amount of constantly changing information with finite energetic resources could benefit from having brain structures not dissimilar to ours.

Achterberg added: “Brains of robots that are deployed in the real physical world are probably going to look more like our brains because they might face the same challenges as us. They need to constantly process new information coming in through their sensors while controlling their bodies to move through space towards a goal. Many systems will need to run all their computations with a limited supply of electric energy and so, to balance these energetic constraints with the amount of information it needs to process, it will probably need a brain structure similar to ours.”

The research was funded by the Medical Research Council, Gates Cambridge, the James S McDonnell Foundation, Templeton World Charity Foundation and Google DeepMind.



Source link

Continue Reading
Click to comment

Leave a Reply

TOP SCEINCE

New snake discovery rewrites history, points to North America’s role in snake evolution

Published

on

By

AI system self-organizes to develop features of brains of complex organisms


A new species of fossil snake unearthed in Wyoming is rewriting our understanding of snake evolution. The discovery, based on four remarkably well-preserved specimens found curled together in a burrow, reveals a new species named Hibernophis breithaupti. This snake lived in North America 34 million years ago and sheds light on the origin and diversification of boas and pythons.

Hibernophis breithaupti has unique anatomical features, in part because the specimens are articulated — meaning they were found all in one piece with the bones still arranged in the proper order — which is unusual for fossil snakes. Researchers believe it may be an early member of Booidea, a group that includes modern boas and pythons. Modern boas are widespread in the Americas, but their early evolution is not well understood.These new and very complete fossils add important new information, in particular, on the evolution of small, burrowing boas known as rubber boas.

Traditionally, there has been much debate on the evolution of small burrowing boas. Hibernophis breithaupti shows that northern and more central parts of North America might have been a key hub for their development. The discovery of these snakes curled together also hints at the oldest potential evidence for a behavior familiar to us today — hibernation in groups.

“Modern garter snakes are famous for gathering by the thousands to hibernate together in dens and burrows,” says Michael Caldwell, a U of A paleontologist who co-led the research along with his former graduate student Jasmine Croghan, and collaborators from Australia and Brazil. “They do this to conserve heat through the effect created by the ball of hibernating animals. It’s fascinating to see possible evidence of such social behavior or hibernation dating back 34 million years.”



Source link

Continue Reading

TOP SCEINCE

Good timing: Study unravels how our brains track time

Published

on

By

AI system self-organizes to develop features of brains of complex organisms


Ever hear the old adage that time flies when you’re having fun? A new study by a team of UNLV researchers suggests that there’s a lot of truth to the trope.

Many people think of their brains as being intrinsically synced to the human-made clocks on their electronic devices, counting time in very specific, minute-by-minute increments. But the study, published this month in the latest issue of the peer-reviewed Cell Press journal Current Biology, showed that our brains don’t work that way.

By analyzing changes in brain activity patterns, the research team found that we perceive the passage of time based on the number of experiences we have — not some kind of internal clock. What’s more, increasing speed or output during an activity appears to affect how our brains perceive time.

“We tell time in our own experience by things we do, things that happen to us,” said James Hyman, a UNLV associate professor of psychology and the study’s senior author. “When we’re still and we’re bored, time goes very slowly because we’re not doing anything or nothing is happening. On the contrary, when a lot of events happen, each one of those activities is advancing our brains forward. And if this is how our brains objectively tell time, then the more that we do and the more that happens to us, the faster time goes.”

Methodology and Findings

The findings are based on analysis of activity in the anterior cingulate cortex (ACC), a portion of the brain important for monitoring activity and tracking experiences. To do this, rodents were tasked with using their noses to respond to a prompt 200 times.

Scientists already knew that brain patterns are similar, but slightly different, each time you do a repetitive motion, so they set out to answer: Is it possible to detect whether these slight differences in brain pattern changes correspond with doing the first versus 200th motion in series? And does the amount of time it takes to complete a series of motions impact brain wave activity?

By comparing pattern changes throughout the course of the task, researchers observed that there are indeed detectable changes in brain activity that occur as one moves from the beginning to middle to end of carrying out a task. And regardless of how slowly or quickly the animals moved, the brain patterns followed the same path. The patterns were consistent when researchers applied a machine learning-based mathematical model to predict the flow of brain activity, bolstering evidence that it’s experiences — not time, or a prescribed number of minutes, as you would measure it on a clock — that produce changes in our neurons’ activity patterns.

Hyman drove home the crux of the findings by sharing an anecdote of two factory workers tasked with making 100 widgets during their shift, with one worker completing the task in 30 minutes and the other in 90 minutes.

“The length of time it took to complete the task didn’t impact the brain patterns. The brain is not a clock; it acts like a counter,” Hyman explained. “Our brains register a vibe, a feeling about time. …And what that means for our workers making widgets is that you can tell the difference between making widget No. 85 and widget No. 60, but not necessarily between No. 85 and No. 88.”

But exactly “how” does the brain count? Researchers discovered that as the brain progresses through a task involving a series of motions, various small groups of firing cells begin to collaborate — essentially passing off the task to a different group of neurons every few repetitions, similar to runners passing the baton in a relay race.

“So, the cells are working together and over time randomly align to get the job done: one cell will take a few tasks and then another takes a few tasks,” Hyman said. “The cells are tracking motions and, thus, chunks of activities and time over the course of the task.”

And the study’s findings about our brains’ perception of time applies to activities-based actions other than physical motions too.

“This is the part of the brain we use for tracking something like a conversation through dinner,” Hyman said. “Think of the flow of conversation and you can recall things earlier and later in the dinner. But to pick apart one sentence from the next in your memory, it’s impossible. But you know you talked about one topic at the start, another topic during dessert, and another at the end.”

By observing the rodents who worked quickly, scientists also concluded that keeping up a good pace helps influence time perception: “The more we do, the faster time moves. They say that time flies when you’re having fun. As opposed to having fun, maybe it should be ‘time flies when you’re doing a lot’.”

Takeaways

While there’s already a wealth of information on brain processes over very short time scales of less than a second, Hyman said that the UNLV study is groundbreaking in its examination of brain patterns and perception of time over a span of just a few minutes to hours — “which is how we live much of our life: one hour at a time. ”

“This is among the first studies looking at behavioral time scales in this particular part of the brain called the ACC, which we know is so important for our behavior and our emotions,” Hyman said.

The ACC is implicated in most psychiatric and neurodegenerative disorders, and is a concentration area for mood disorders, PTSD, addiction, and anxiety. ACC function is also central to various dementias including Alzheimer’s disease, which is characterized by distortions in time. The ACC has long been linked to helping humans with sequencing events or tasks such as following recipes, and the research team speculates that their findings about time perception might fall within this realm.

While the findings are a breakthrough, more research is needed. Still, Hyman said, the preliminary findings posit some potentially helpful tidbits about time perception and its likely connection to memory processes for everyday citizens’ daily lives. For example, researchers speculate that it could lend insights for navigating things like school assignments or even breakups.

“If we want to remember something, we may want to slow down by studying in short bouts and take time before engaging in the next activity. Give yourself quiet times to not move,” Hyman said. “Conversely, if you want to move on from something quickly, get involved in an activity right away.”

Hyman said there’s also a huge relationship between the ACC, emotion, and cognition. Thinking of the brain as a physical entity that one can take ownership over might help us control our subjective experiences.

“When things move faster, we tend to think it’s more fun — or sometimes overwhelming. But we don’t need to think of it as being a purely psychological experience, as fun or overwhelming; rather, if you view it as a physical process, it can be helpful,” he said. “If it’s overwhelming, slow down or if you’re bored, add activities. People already do this, but it’s empowering to know it’s a way to work your own mental health, since our brains are working like this already.”



Source link

Continue Reading

TOP SCEINCE

Another intermediate-mass black hole discovered at the center of our galaxy

Published

on

By

AI system self-organizes to develop features of brains of complex organisms


While researching a cluster of stars in the immediate vicinity of the supermassive black hole SgrA* (Sagittarius A*) at the centre of our galaxy, an international team of researchers led by PD Dr Florian Peißker has found signs of another, intermediate-mass black hole. Despite enormous research efforts, only about ten of these intermediate-mass black holes have been found in our entire universe so far. Scientists believe that they formed shortly after the Big Bang. By merging, they act as ‘seeds’ for supermassive black holes. The study ‘The Evaporating Massive Embedded Stellar Cluster IRS 13 Close to Sgr A*. II. Kinematic structure’ was published in The Astrophysical Journal.

The analysed star cluster IRS 13 is located 0.1 light years from the centre of our galaxy. This is very close in astronomical terms, but would still require travelling from one end of our solar system to the other twenty times to cover the distance. The researchers noticed that the stars in IRS 13 move in an unexpectedly orderly pattern. They had actually expected the stars to be arranged randomly. Two conclusions can be drawn from this regular pattern: On the one hand, IRS 13 appears to interact with SgrA*, which leads to the orderly motion of the stars. On the other hand, there must be something inside the cluster for it to be able to maintain its observed compact shape.

Multi-wavelength observations with the Very Large Telescope as well as the ALMA and Chandra telescopes now suggest that the reason for the compact shape of IRS 13 could be an intermediate-mass black hole located at the centre of the star cluster. This would be supported by the fact that the researchers were able to observe characteristic X-rays and ionized gas rotating at a speed of several 100 km/s in a ring around the suspected location of the intermediate-mass black hole.

Another indication of the presence of an intermediate-mass black hole is the unusually high density of the star cluster, which is higher than that of any other known density of a star cluster in our Milky Way. “IRS 13 appears to be an essential building block for the growth of our central black hole SgrA*,” said Florian Peißker, first author of the study. “This fascinating star cluster has continued to surprise the scientific community ever since it was discovered around twenty years ago. At first it was thought to be an unusually heavy star. With the high-resolution data, however, we can now confirm the building-block composition with an intermediate-mass black hole at the centre.” Planned observations with the James Webb Space Telescope and the Extremely Large Telescope, which is currently under construction, will provide further insights into the processes within the star cluster.



Source link

Continue Reading

Trending