Tag Archives: level

#431920 If We Could Engineer Animals to Be as ...

Advances in neural implants and genetic engineering suggest that in the not–too–distant future we may be able to boost human intelligence. If that’s true, could we—and should we—bring our animal cousins along for the ride?
Human brain augmentation made headlines last year after several tech firms announced ambitious efforts to build neural implant technology. Duke University neuroscientist Mikhail Lebedev told me in July it could be decades before these devices have applications beyond the strictly medical.
But he said the technology, as well as other pharmacological and genetic engineering approaches, will almost certainly allow us to boost our mental capacities at some point in the next few decades.
Whether this kind of cognitive enhancement is a good idea or not, and how we should regulate it, are matters of heated debate among philosophers, futurists, and bioethicists, but for some it has raised the question of whether we could do the same for animals.
There’s already tantalizing evidence of the idea’s feasibility. As detailed in BBC Future, a group from MIT found that mice that were genetically engineered to express the human FOXP2 gene linked to learning and speech processing picked up maze routes faster. Another group at Wake Forest University studying Alzheimer’s found that neural implants could boost rhesus monkeys’ scores on intelligence tests.
The concept of “animal uplift” is most famously depicted in the Planet of the Apes movie series, whose planet–conquering protagonists are likely to put most people off the idea. But proponents are less pessimistic about the outcomes.
Science fiction author David Brin popularized the concept in his “Uplift” series of novels, in which humans share the world with various other intelligent animals that all bring their own unique skills, perspectives, and innovations to the table. “The benefits, after a few hundred years, could be amazing,” he told Scientific American.
Others, like George Dvorsky, the director of the Rights of Non-Human Persons program at the Institute for Ethics and Emerging Technologies, go further and claim there is a moral imperative. He told the Boston Globe that denying augmentation technology to animals would be just as unethical as excluding certain groups of humans.
Others are less convinced. Forbes’ Alex Knapp points out that developing the technology to uplift animals will likely require lots of very invasive animal research that will cause huge suffering to the animals it purports to help. This is problematic enough with normal animals, but could be even more morally dubious when applied to ones whose cognitive capacities have been enhanced.
The whole concept could also be based on a fundamental misunderstanding of the nature of intelligence. Humans are prone to seeing intelligence as a single, self-contained metric that progresses in a linear way with humans at the pinnacle.
In an opinion piece in Wired arguing against the likelihood of superhuman artificial intelligence, Kevin Kelly points out that science has no such single dimension with which to rank the intelligence of different species. Each one combines a bundle of cognitive capabilities, some of which are well below our own capabilities and others which are superhuman. He uses the example of the squirrel, which can remember the precise location of thousands of acorns for years.
Uplift efforts may end up being less about boosting intelligence and more about making animals more human-like. That represents “a kind of benevolent colonialism” that assumes being more human-like is a good thing, Paul Graham Raven, a futures researcher at the University of Sheffield in the United Kingdom, told the Boston Globe. There’s scant evidence that’s the case, and it’s easy to see how a chimpanzee with the mind of a human might struggle to adjust.
There are also fundamental barriers that may make it difficult to achieve human-level cognitive capabilities in animals, no matter how advanced brain augmentation technology gets. In 2013 Swedish researchers selectively bred small fish called guppies for bigger brains. This made them smarter, but growing the energy-intensive organ meant the guppies developed smaller guts and produced fewer offspring to compensate.
This highlights the fact that uplifting animals may require more than just changes to their brains, possibly a complete rewiring of their physiology that could prove far more technically challenging than human brain augmentation.
Our intelligence is intimately tied to our evolutionary history—our brains are bigger than other animals’; opposable thumbs allow us to use tools; our vocal chords make complex communication possible. No matter how much you augment a cow’s brain, it still couldn’t use a screwdriver or talk to you in English because it simply doesn’t have the machinery.
Finally, from a purely selfish point of view, even if it does become possible to create a level playing field between us and other animals, it may not be a smart move for humanity. There’s no reason to assume animals would be any more benevolent than we are, having evolved in the same ‘survival of the fittest’ crucible that we have. And given our already endless capacity to divide ourselves along national, religious, or ethnic lines, conflict between species seems inevitable.
We’re already likely to face considerable competition from smart machines in the coming decades if you believe the hype around AI. So maybe adding a few more intelligent species to the mix isn’t the best idea.
Image Credit: Ron Meijer / Shutterstock.com Continue reading

Posted in Human Robots | Tagged , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , | Leave a comment

#431873 Why the World Is Still Getting ...

If you read or watch the news, you’ll likely think the world is falling to pieces. Trends like terrorism, climate change, and a growing population straining the planet’s finite resources can easily lead you to think our world is in crisis.
But there’s another story, a story the news doesn’t often report. This story is backed by data, and it says we’re actually living in the most peaceful, abundant time in history, and things are likely to continue getting better.
The News vs. the Data
The reality that’s often clouded by a constant stream of bad news is we’re actually seeing a massive drop in poverty, fewer deaths from violent crime and preventable diseases. On top of that, we’re the most educated populace to ever walk the planet.
“Violence has been in decline for thousands of years, and today we may be living in the most peaceful era in the existence of our species.” –Steven Pinker
In the last hundred years, we’ve seen the average human life expectancy nearly double, the global GDP per capita rise exponentially, and childhood mortality drop 10-fold.

That’s pretty good progress! Maybe the world isn’t all gloom and doom.If you’re still not convinced the world is getting better, check out the charts in this article from Vox and on Peter Diamandis’ website for a lot more data.
Abundance for All Is Possible
So now that you know the world isn’t so bad after all, here’s another thing to think about: it can get much better, very soon.
In their book Abundance: The Future Is Better Than You Think, Steven Kotler and Peter Diamandis suggest it may be possible for us to meet and even exceed the basic needs of all the people living on the planet today.
“In the hands of smart and driven innovators, science and technology take things which were once scarce and make them abundant and accessible to all.”
This means making sure every single person in the world has adequate food, water and shelter, as well as a good education, access to healthcare, and personal freedom.
This might seem unimaginable, especially if you tend to think the world is only getting worse. But given how much progress we’ve already made in the last few hundred years, coupled with the recent explosion of information sharing and new, powerful technologies, abundance for all is not as out of reach as you might believe.
Throughout history, we’ve seen that in the hands of smart and driven innovators, science and technology take things which were once scarce and make them abundant and accessible to all.
Napoleon III
In Abundance, Diamandis and Kotler tell the story of how aluminum went from being one of the rarest metals on the planet to being one of the most abundant…
In the 1800s, aluminum was more valuable than silver and gold because it was rarer. So when Napoleon III entertained the King of Siam, the king and his guests were honored by being given aluminum utensils, while the rest of the dinner party ate with gold.
But aluminum is not really rare.
In fact, aluminum is the third most abundant element in the Earth’s crust, making up 8.3% of the weight of our planet. But it wasn’t until chemists Charles Martin Hall and Paul Héroult discovered how to use electrolysis to cheaply separate aluminum from surrounding materials that the element became suddenly abundant.
The problems keeping us from achieving a world where everyone’s basic needs are met may seem like resource problems — when in reality, many are accessibility problems.
The Engine Driving Us Toward Abundance: Exponential Technology
History is full of examples like the aluminum story. The most powerful one of the last few decades is information technology. Think about all the things that computers and the internet made abundant that were previously far less accessible because of cost or availability … Here are just a few examples:

Easy access to the world’s information
Ability to share information freely with anyone and everyone
Free/cheap long-distance communication
Buying and selling goods/services regardless of location

Less than two decades ago, when someone reached a certain level of economic stability, they could spend somewhere around $10K on stereos, cameras, entertainment systems, etc — today, we have all that equipment in the palm of our hand.
Now, there is a new generation of technologies heavily dependant on information technology and, therefore, similarly riding the wave of exponential growth. When put to the right use, emerging technologies like artificial intelligence, robotics, digital manufacturing, nano-materials and digital biology make it possible for us to drastically raise the standard of living for every person on the planet.

These are just some of the innovations which are unlocking currently scarce resources:

IBM’s Watson Health is being trained and used in medical facilities like the Cleveland Clinic to help doctors diagnose disease. In the future, it’s likely we’ll trust AI just as much, if not more than humans to diagnose disease, allowing people all over the world to have access to great diagnostic tools regardless of whether there is a well-trained doctor near them.

Solar power is now cheaper than fossil fuels in some parts of the world, and with advances in new materials and storage, the cost may decrease further. This could eventually lead to nearly-free, clean energy for people across the world.

Google’s GMNT network can now translate languages as well as a human, unlocking the ability for people to communicate globally as we never have before.

Self-driving cars are already on the roads of several American cities and will be coming to a road near you in the next couple years. Considering the average American spends nearly two hours driving every day, not having to drive would free up an increasingly scarce resource: time.

The Change-Makers
Today’s innovators can create enormous change because they have these incredible tools—which would have once been available only to big organizations—at their fingertips. And, as a result of our hyper-connected world, there is an unprecedented ability for people across the planet to work together to create solutions to some of our most pressing problems today.
“In today’s hyperlinked world, solving problems anywhere, solves problems everywhere.” –Peter Diamandis and Steven Kotler, Abundance
According to Diamandis and Kotler, there are three groups of people accelerating positive change.

DIY InnovatorsIn the 1970s and 1980s, the Homebrew Computer Club was a meeting place of “do-it-yourself” computer enthusiasts who shared ideas and spare parts. By the 1990s and 2000s, that little club became known as an inception point for the personal computer industry — dozens of companies, including Apple Computer, can directly trace their origins back to Homebrew. Since then, we’ve seen the rise of the social entrepreneur, the Maker Movement and the DIY Bio movement, which have similar ambitions to democratize social reform, manufacturing, and biology, the way Homebrew democratized computers. These are the people who look for new opportunities and aren’t afraid to take risks to create something new that will change the status-quo.
Techno-PhilanthropistsUnlike the robber barons of the 19th and early 20th centuries, today’s “techno-philanthropists” are not just giving away some of their wealth for a new museum, they are using their wealth to solve global problems and investing in social entrepreneurs aiming to do the same. The Bill and Melinda Gates Foundation has given away at least $28 billion, with a strong focus on ending diseases like polio, malaria, and measles for good. Jeff Skoll, after cashing out of eBay with $2 billion in 1998, went on to create the Skoll Foundation, which funds social entrepreneurs across the world. And last year, Mark Zuckerberg and Priscilla Chan pledged to give away 99% of their $46 billion in Facebook stock during their lifetimes.
The Rising BillionCisco estimates that by 2020, there will be 4.1 billion people connected to the internet, up from 3 billion in 2015. This number might even be higher, given the efforts of companies like Facebook, Google, Virgin Group, and SpaceX to bring internet access to the world. That’s a billion new people in the next several years who will be connected to the global conversation, looking to learn, create and better their own lives and communities.In his book, Fortune at the Bottom of the Pyramid, C.K. Pahalad writes that finding co-creative ways to serve this rising market can help lift people out of poverty while creating viable businesses for inventive companies.

The Path to Abundance
Eager to create change, innovators armed with powerful technologies can accomplish incredible feats. Kotler and Diamandis imagine that the path to abundance occurs in three tiers:

Basic Needs (food, water, shelter)
Tools of Growth (energy, education, access to information)
Ideal Health and Freedom

Of course, progress doesn’t always happen in a straight, logical way, but having a framework to visualize the needs is helpful.
Many people don’t believe it’s possible to end the persistent global problems we’re facing. However, looking at history, we can see many examples where technological tools have unlocked resources that previously seemed scarce.
Technological solutions are not always the answer, and we need social change and policy solutions as much as we need technology solutions. But we have seen time and time again, that powerful tools in the hands of innovative, driven change-makers can make the seemingly impossible happen.

You can download the full “Path to Abundance” infographic here. It was created under a CC BY-NC-ND license. If you share, please attribute to Singularity University.
Image Credit: janez volmajer / Shutterstock.com Continue reading

Posted in Human Robots | Tagged , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , | Leave a comment

#431869 When Will We Finally Achieve True ...

The field of artificial intelligence goes back a long way, but many consider it was officially born when a group of scientists at Dartmouth College got together for a summer, back in 1956. Computers had, over the last few decades, come on in incredible leaps and bounds; they could now perform calculations far faster than humans. Optimism, given the incredible progress that had been made, was rational. Genius computer scientist Alan Turing had already mooted the idea of thinking machines just a few years before. The scientists had a fairly simple idea: intelligence is, after all, just a mathematical process. The human brain was a type of machine. Pick apart that process, and you can make a machine simulate it.
The problem didn’t seem too hard: the Dartmouth scientists wrote, “We think that a significant advance can be made in one or more of these problems if a carefully selected group of scientists work on it together for a summer.” This research proposal, by the way, contains one of the earliest uses of the term artificial intelligence. They had a number of ideas—maybe simulating the human brain’s pattern of neurons could work and teaching machines the abstract rules of human language would be important.
The scientists were optimistic, and their efforts were rewarded. Before too long, they had computer programs that seemed to understand human language and could solve algebra problems. People were confidently predicting there would be a human-level intelligent machine built within, oh, let’s say, the next twenty years.
It’s fitting that the industry of predicting when we’d have human-level intelligent AI was born at around the same time as the AI industry itself. In fact, it goes all the way back to Turing’s first paper on “thinking machines,” where he predicted that the Turing Test—machines that could convince humans they were human—would be passed in 50 years, by 2000. Nowadays, of course, people are still predicting it will happen within the next 20 years, perhaps most famously Ray Kurzweil. There are so many different surveys of experts and analyses that you almost wonder if AI researchers aren’t tempted to come up with an auto reply: “I’ve already predicted what your question will be, and no, I can’t really predict that.”
The issue with trying to predict the exact date of human-level AI is that we don’t know how far is left to go. This is unlike Moore’s Law. Moore’s Law, the doubling of processing power roughly every couple of years, makes a very concrete prediction about a very specific phenomenon. We understand roughly how to get there—improved engineering of silicon wafers—and we know we’re not at the fundamental limits of our current approach (at least, not until you’re trying to work on chips at the atomic scale). You cannot say the same about artificial intelligence.
Common Mistakes
Stuart Armstrong’s survey looked for trends in these predictions. Specifically, there were two major cognitive biases he was looking for. The first was the idea that AI experts predict true AI will arrive (and make them immortal) conveniently just before they’d be due to die. This is the “Rapture of the Nerds” criticism people have leveled at Kurzweil—his predictions are motivated by fear of death, desire for immortality, and are fundamentally irrational. The ability to create a superintelligence is taken as an article of faith. There are also criticisms by people working in the AI field who know first-hand the frustrations and limitations of today’s AI.
The second was the idea that people always pick a time span of 15 to 20 years. That’s enough to convince people they’re working on something that could prove revolutionary very soon (people are less impressed by efforts that will lead to tangible results centuries down the line), but not enough for you to be embarrassingly proved wrong. Of the two, Armstrong found more evidence for the second one—people were perfectly happy to predict AI after they died, although most didn’t, but there was a clear bias towards “15–20 years from now” in predictions throughout history.
Measuring Progress
Armstrong points out that, if you want to assess the validity of a specific prediction, there are plenty of parameters you can look at. For example, the idea that human-level intelligence will be developed by simulating the human brain does at least give you a clear pathway that allows you to assess progress. Every time we get a more detailed map of the brain, or successfully simulate another part of it, we can tell that we are progressing towards this eventual goal, which will presumably end in human-level AI. We may not be 20 years away on that path, but at least you can scientifically evaluate the progress.
Compare this to those that say AI, or else consciousness, will “emerge” if a network is sufficiently complex, given enough processing power. This might be how we imagine human intelligence and consciousness emerged during evolution—although evolution had billions of years, not just decades. The issue with this is that we have no empirical evidence: we have never seen consciousness manifest itself out of a complex network. Not only do we not know if this is possible, we cannot know how far away we are from reaching this, as we can’t even measure progress along the way.
There is an immense difficulty in understanding which tasks are hard, which has continued from the birth of AI to the present day. Just look at that original research proposal, where understanding human language, randomness and creativity, and self-improvement are all mentioned in the same breath. We have great natural language processing, but do our computers understand what they’re processing? We have AI that can randomly vary to be “creative,” but is it creative? Exponential self-improvement of the kind the singularity often relies on seems far away.
We also struggle to understand what’s meant by intelligence. For example, AI experts consistently underestimated the ability of AI to play Go. Many thought, in 2015, it would take until 2027. In the end, it took two years, not twelve. But does that mean AI is any closer to being able to write the Great American Novel, say? Does it mean it’s any closer to conceptually understanding the world around it? Does it mean that it’s any closer to human-level intelligence? That’s not necessarily clear.
Not Human, But Smarter Than Humans
But perhaps we’ve been looking at the wrong problem. For example, the Turing test has not yet been passed in the sense that AI cannot convince people it’s human in conversation; but of course the calculating ability, and perhaps soon the ability to perform other tasks like pattern recognition and driving cars, far exceed human levels. As “weak” AI algorithms make more decisions, and Internet of Things evangelists and tech optimists seek to find more ways to feed more data into more algorithms, the impact on society from this “artificial intelligence” can only grow.
It may be that we don’t yet have the mechanism for human-level intelligence, but it’s also true that we don’t know how far we can go with the current generation of algorithms. Those scary surveys that state automation will disrupt society and change it in fundamental ways don’t rely on nearly as many assumptions about some nebulous superintelligence.
Then there are those that point out we should be worried about AI for other reasons. Just because we can’t say for sure if human-level AI will arrive this century, or never, it doesn’t mean we shouldn’t prepare for the possibility that the optimistic predictors could be correct. We need to ensure that human values are programmed into these algorithms, so that they understand the value of human life and can act in “moral, responsible” ways.
Phil Torres, at the Project for Future Human Flourishing, expressed it well in an interview with me. He points out that if we suddenly decided, as a society, that we had to solve the problem of morality—determine what was right and wrong and feed it into a machine—in the next twenty years…would we even be able to do it?
So, we should take predictions with a grain of salt. Remember, it turned out the problems the AI pioneers foresaw were far more complicated than they anticipated. The same could be true today. At the same time, we cannot be unprepared. We should understand the risks and take our precautions. When those scientists met in Dartmouth in 1956, they had no idea of the vast, foggy terrain before them. Sixty years later, we still don’t know how much further there is to go, or how far we can go. But we’re going somewhere.
Image Credit: Ico Maker / Shutterstock.com Continue reading

Posted in Human Robots | Tagged , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , | Leave a comment

#431866 The Technologies We’ll Have Our Eyes ...

It’s that time of year again when our team has a little fun and throws on our futurist glasses to look ahead at some of the technologies and trends we’re most anticipating next year.
Whether the implications of a technology are vast or it resonates with one of us personally, here’s the list from some of the Singularity Hub team of what we have our eyes on as we enter the new year.
For a little refresher, these were the technologies our team was fired up about at the start of 2017.
Tweet us the technology you’re excited to watch in 2018 at @SingularityHub.
Cryptocurrency and Blockchain
“Given all the noise Bitcoin is making globally in the media, it is driving droves of main street investors to dabble in and learn more about cryptocurrencies. This will continue to raise valuations and drive adoption of blockchain. From Bank of America recently getting a blockchain-based patent approved to the Australian Securities Exchange’s plan to use blockchain, next year is going to be chock-full of these stories. Coindesk even recently spotted a patent filing from Apple involving blockchain. From ‘China’s Ethereum’, NEO, to IOTA to Golem to Qtum, there are a lot of interesting cryptos to follow given the immense numbers of potential applications. Hang on, it’s going to be a bumpy ride in 2018!”
–Kirk Nankivell, Website Manager
There Is No One Technology to Watch
“Next year may be remembered for advances in gene editing, blockchain, AI—or most likely all these and more. There is no single technology to watch. A number of consequential trends are advancing and converging. This general pace of change is exciting, and it also contributes to spiking anxiety. Technology’s invisible lines of force are extending further and faster into our lives and subtly subverting how we view the world and each other in unanticipated ways. Still, all the near-term messiness and volatility, the little and not-so-little dramas, the hype and disillusion, the controversies and conflict, all that smooths out a bit when you take a deep breath and a step back, and it’s my sincere hope and belief the net result will be more beneficial than harmful.”
–Jason Dorrier, Managing Editor
‘Fake News’ Fighting Technology
“It’s been a wild ride for the media this year with the term ‘fake news’ moving from the public’s peripheral and into mainstream vocabulary. The spread of ‘fake news’ is often blamed on media outlets, but social media platforms and search engines are often responsible too. (Facebook still won’t identify as a media company—maybe next year?) Yes, technology can contribute to spreading false information, but it can also help stop it. From technologists who are building in-article ‘trust indicator’ features, to artificial intelligence systems that can both spot and shut down fake news early on, I’m hopeful we can create new solutions to this huge problem. One step further: if publishers step up to fix this we might see some faith restored in the media.”
–Alison E. Berman, Digital Producer
Pay-as-You-Go Home Solar Power
“People in rural African communities are increasingly bypassing electrical grids (which aren’t even an option in many cases) and installing pay-as-you-go solar panels on their homes. The companies offering these services are currently not subject to any regulations, though they’re essentially acting as a utility. As demand for power grows, they’ll have to come up with ways to efficiently scale, and to balance the humanitarian and capitalistic aspects of their work. It’s fascinating to think traditional grids may never be necessary in many areas of the continent thanks to this technology.”
–Vanessa Bates Ramirez, Associate Editor
Virtual Personal Assistants
“AI is clearly going to rule our lives, and in many ways it already makes us look like clumsy apes. Alexa, Siri, and Google Assistant are promising first steps toward a world of computers that understand us and relate to us on an emotional level. I crave the day when my Apple Watch coaches me into healthier habits, lets me know about new concerts nearby, speaks to my self-driving Lyft on my behalf, and can help me respond effectively to aggravating emails based on communication patterns. But let’s not brush aside privacy concerns and the implications of handing over our personal data to megacorporations. The scariest thing here is that privacy laws and advertising ethics do not accommodate this level of intrusive data hoarding.”
–Matthew Straub, Director of Digital Engagement (Hub social media)
Solve for Learning: Educational Apps for Children in Conflict Zones
“I am most excited by exponential technology when it is used to help solve a global grand challenge. Educational apps are currently being developed to help solve for learning by increasing accessibility to learning opportunities for children living in conflict zones. Many children in these areas are not receiving an education, with girls being 2.5 times more likely than boys to be out of school. The EduApp4Syria project is developing apps to help children in Syria and Kashmir learn in their native languages. Mobile phones are increasingly available in these areas, and the apps are available offline for children who do not have consistent access to mobile networks. The apps are low-cost, easily accessible, and scalable educational opportunities.
–Paige Wilcoxson, Director, Curriculum & Learning Design
Image Credit: Triff / Shutterstock.com Continue reading

Posted in Human Robots | Tagged , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , | Leave a comment

#431836 Do Our Brains Use Deep Learning to Make ...

The first time Dr. Blake Richards heard about deep learning, he was convinced that he wasn’t just looking at a technique that would revolutionize artificial intelligence. He also knew he was looking at something fundamental about the human brain.
That was the early 2000s, and Richards was taking a course with Dr. Geoff Hinton at the University of Toronto. Hinton, a pioneer architect of the algorithm that would later take the world by storm, was offering an introductory course on his learning method inspired by the human brain.
The key words here are “inspired by.” Despite Richards’ conviction, the odds were stacked against him. The human brain, as it happens, seems to lack a critical function that’s programmed into deep learning algorithms. On the surface, the algorithms were violating basic biological facts already proven by neuroscientists.
But what if, superficial differences aside, deep learning and the brain are actually compatible?
Now, in a new study published in eLife, Richards, working with DeepMind, proposed a new algorithm based on the biological structure of neurons in the neocortex. Also known as the cortex, this outermost region of the brain is home to higher cognitive functions such as reasoning, prediction, and flexible thought.
The team networked their artificial neurons together into a multi-layered network and challenged it with a classic computer vision task—identifying hand-written numbers.
The new algorithm performed well. But the kicker is that it analyzed the learning examples in a way that’s characteristic of deep learning algorithms, even though it was completely based on the brain’s fundamental biology.
“Deep learning is possible in a biological framework,” concludes the team.
Because the model is only a computer simulation at this point, Richards hopes to pass the baton to experimental neuroscientists, who could actively test whether the algorithm operates in an actual brain.
If so, the data could then be passed back to computer scientists to work out the next generation of massively parallel and low-energy algorithms to power our machines.
It’s a first step towards merging the two fields back into a “virtuous circle” of discovery and innovation.
The blame game
While you’ve probably heard of deep learning’s recent wins against humans in the game of Go, you might not know the nitty-gritty behind the algorithm’s operations.
In a nutshell, deep learning relies on an artificial neural network with virtual “neurons.” Like a towering skyscraper, the network is structured into hierarchies: lower-level neurons process aspects of an input—for example, a horizontal or vertical stroke that eventually forms the number four—whereas higher-level neurons extract more abstract aspects of the number four.
To teach the network, you give it examples of what you’re looking for. The signal propagates forward in the network (like climbing up a building), where each neuron works to fish out something fundamental about the number four.
Like children trying to learn a skill the first time, initially the network doesn’t do so well. It spits out what it thinks a universal number four should look like—think a Picasso-esque rendition.
But here’s where the learning occurs: the algorithm compares the output with the ideal output, and computes the difference between the two (dubbed “error”). This error is then “backpropagated” throughout the entire network, telling each neuron: hey, this is how far off you were, so try adjusting your computation closer to the ideal.
Millions of examples and tweakings later, the network inches closer to the desired output and becomes highly proficient at the trained task.
This error signal is crucial for learning. Without efficient “backprop,” the network doesn’t know which of its neurons are off kilter. By assigning blame, the AI can better itself.
The brain does this too. How? We have no clue.
Biological No-Go
What’s clear, though, is that the deep learning solution doesn’t work.
Backprop is a pretty needy function. It requires a very specific infrastructure for it to work as expected.
For one, each neuron in the network has to receive the error feedback. But in the brain, neurons are only connected to a few downstream partners (if that). For backprop to work in the brain, early-level neurons need to be able to receive information from billions of connections in their downstream circuits—a biological impossibility.
And while certain deep learning algorithms adapt a more local form of backprop— essentially between neurons—it requires their connection forwards and backwards to be symmetric. This hardly ever occurs in the brain’s synapses.
More recent algorithms adapt a slightly different strategy, in that they implement a separate feedback pathway that helps the neurons to figure out errors locally. While it’s more biologically plausible, the brain doesn’t have a separate computational network dedicated to the blame game.
What it does have are neurons with intricate structures, unlike the uniform “balls” that are currently applied in deep learning.
Branching Networks
The team took inspiration from pyramidal cells that populate the human cortex.
“Most of these neurons are shaped like trees, with ‘roots’ deep in the brain and ‘branches’ close to the surface,” says Richards. “What’s interesting is that these roots receive a different set of inputs than the branches that are way up at the top of the tree.”
This is an illustration of a multi-compartment neural network model for deep learning. Left: Reconstruction of pyramidal neurons from mouse primary visual cortex. Right: Illustration of simplified pyramidal neuron models. Image Credit: CIFAR
Curiously, the structure of neurons often turn out be “just right” for efficiently cracking a computational problem. Take the processing of sensations: the bottoms of pyramidal neurons are right smack where they need to be to receive sensory input, whereas the tops are conveniently placed to transmit feedback errors.
Could this intricate structure be evolution’s solution to channeling the error signal?
The team set up a multi-layered neural network based on previous algorithms. But rather than having uniform neurons, they gave those in middle layers—sandwiched between the input and output—compartments, just like real neurons.
When trained with hand-written digits, the algorithm performed much better than a single-layered network, despite lacking a way to perform classical backprop. The cell-like structure itself was sufficient to assign error: the error signals at one end of the neuron are naturally kept separate from input at the other end.
Then, at the right moment, the neuron brings both sources of information together to find the best solution.
There’s some biological evidence for this: neuroscientists have long known that the neuron’s input branches perform local computations, which can be integrated with signals that propagate backwards from the so-called output branch.
However, we don’t yet know if this is the brain’s way of dealing blame—a question that Richards urges neuroscientists to test out.
What’s more, the network parsed the problem in a way eerily similar to traditional deep learning algorithms: it took advantage of its multi-layered structure to extract progressively more abstract “ideas” about each number.
“[This is] the hallmark of deep learning,” the authors explain.
The Deep Learning Brain
Without doubt, there will be more twists and turns to the story as computer scientists incorporate more biological details into AI algorithms.
One aspect that Richards and team are already eyeing is a top-down predictive function, in which signals from higher levels directly influence how lower levels respond to input.
Feedback from upper levels doesn’t just provide error signals; it could also be nudging lower processing neurons towards a “better” activity pattern in real-time, says Richards.
The network doesn’t yet outperform other non-biologically derived (but “brain-inspired”) deep networks. But that’s not the point.
“Deep learning has had a huge impact on AI, but, to date, its impact on neuroscience has been limited,” the authors say.
Now neuroscientists have a lead they could experimentally test: that the structure of neurons underlie nature’s own deep learning algorithm.
“What we might see in the next decade or so is a real virtuous cycle of research between neuroscience and AI, where neuroscience discoveries help us to develop new AI and AI can help us interpret and understand our experimental data in neuroscience,” says Richards.
Image Credit: christitzeimaging.com / Shutterstock.com Continue reading

Posted in Human Robots | Tagged , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , | Leave a comment