AI is the new Grandmaster of StarCraft II

by PETER DOCKRILL

Video games were invented for humans, by humans. But that doesn’t necessarily mean we’re the best when it comes to playing them.

In a new achievement that signifies just how far artificial intelligence (AI) has progressed, scientists have developed a learning algorithm that rose to the very top echelon of the esports powerhouse StarCraft II, reaching Grandmaster level.

According to the researchers who created the AI – called AlphaStar – the accomplishment of reaching the Grandmaster League means you’re in the top 0.2 percent of StarCraft II players.

In other words, AlphaStar competes at a level in this multi-player real-time strategy game that could trounce millions of humans foolhardy enough to take it on.

In recent years, we’ve seen AI come to dominate games that represent more traditional tests of human skill, mastering the strategies of chess, poker, and Go.

For David Silver, principal research scientist at AI firm DeepMind in the UK, those kinds of milestones – many of which DeepMind pioneered – are what’s led us to this inevitable moment: a game representing even greater problems than the ancient games that have challenged human minds for centuries.

“Ever since computers cracked Go, chess, and poker, StarCraft has emerged by consensus as the next grand challenge,” Silver says.

“The game’s complexity is much greater than chess, because players control hundreds of units; more complex than Go, because there are 1,026 possible choices for every move; and players have less information about their opponents than in poker.”

Add it all together and mastering the complex real-time battles of StarCraft seems almost impossible for a machine, so how did they do it?

In a new paper published this week, the DeepMind team describes how they developed a multi-agent reinforcement learning algorithm, which trained itself up through self-play, including playing against itself, and playing humans, learning to mimic successful strategies, and also effective counter-strategies.

The research team has been working towards this goal for years. An earlier version of the system made headlines back in January when it started to beat human professionals.

“I will never forget the excitement and emotion we all felt when AlphaStar first started playing real competitive matches,” says Dario “TLO” Wünsch, one of the top human StarCraft II players beaten by the algorithm.

“The system is very skilled at assessing its strategic position, and knows exactly when to engage or disengage with its opponent.”

The latest algorithm takes things even further than that preliminary incarnation, and now effectively plays under artificial constraints designed to most realistically simulate gameplay as experienced by a human (such as observing the game at a distance, through a camera, and feeling the delay of network latency).

With all the imposed limitations of a human, AlphaStar still reached Grandmaster level in real, online competitive play, representing not just a world-first, but perhaps a sunset of these kinds of gaming challenges, given what the achievement now may make possible.

“Like StarCraft, real-world domains such as personal assistants, self-driving cars, or robotics require real-time decisions, over combinatorial or structured action spaces, given imperfectly observed information,” the authors write.

“The success of AlphaStar in StarCraft II suggests that general-purpose machine learning algorithms may have a substantial effect on complex real-world problems.”

The findings are reported in Nature.

https://www.sciencealert.com/starcraft-ii-has-a-new-grandmaster-and-it-s-not-human?perpetual=yes&limitstart=1

Neuroscientists Have Revealed What Your Brain Needs to Really Erase a Memory

by PETER DOCKRILL

When bad things happen, we don’t want to remember. We try to block, resist, ignore – but we should perhaps be doing the opposite, researchers say.

A new study led by scientists in Texas suggests the act of intentionally forgetting is linked to increased cerebral engagement with the unwanted information in question. In other words, to forget something, you actually need to focus on it.

“A moderate level of brain activity is critical to this forgetting mechanism,” explains psychologist Tracy Wang from the University of Texas at Austin.

“Too strong, and it will strengthen the memory; too weak, and you won’t modify it.”

Trying to actively forget unwanted memories doesn’t just help prevent your brain from getting overloaded.

It also lets people move on from painful experiences and emotions they’d rather not recall, which is part of the reason it’s an area of active interest to neuroscientists.

“We may want to discard memories that trigger maladaptive responses, such as traumatic memories, so that we can respond to new experiences in more adaptive ways,” says one of the researchers, Jarrod Lewis-Peacock.

“Decades of research has shown that we have the ability to voluntarily forget something, but how our brains do that is still being questioned.”

Much prior research on intentional forgetting has focussed on brain activity in the prefrontal cortex, and the brain’s memory centre, the hippocampus.

In the new study, the researchers monitored a different part of the brain called the ventral temporal cortex, which helps us process and categorise visual stimuli.

In an experiment with 24 healthy young adults, the participants were shown pictures of scenes and people’s faces, and were instructed to either remember or forget each image.

During the experiment, each of the participants had their brain activity monitored by functional magnetic resonance imaging (fMRI) machines.

When the researchers examined activity in the ventral temporal cortex, they found that the act of forgetting effectively uses more brain power than remembering.

“Pictures followed by a forget instruction elicited higher levels of processing in [the] ventral temporal cortex compared to those followed by a remember instruction,” the authors write in their paper.

“This boost in processing led to more forgetting, particularly for items that showed moderate (vs. weak or strong) activation.”

Of course, forgetting specific images on demand in a contrived laboratory experiment is very different to moving on from painful or traumatic memories of events experienced in the real world.

But the mechanisms at work could be the same, researchers say, and figuring out how to activate them could be a huge benefit to people around the world who need to forget things, but don’t know how.

Especially since this finding in particular challenges our natural intuition to suppress things; instead, we should involve more rather than less attention to unwanted information, in order to forget it.

“Importantly, it’s the intention to forget that increases the activation of the memory,” Wang says.

“When this activation hits the ‘moderate level’ sweet spot, that’s when it leads to later forgetting of that experience.”

The findings are reported in JNeurosci.

https://www.sciencealert.com/to-forget-something-you-need-to-think-about-it-neuroscientists-reveal

A new form of neural communication in the brain has been discovered.

by PETER DOCKRILL

Scientists think they’ve identified a previously unknown form of neural communication that self-propagates across brain tissue, and can leap wirelessly from neurons in one section of brain tissue to another – even if they’ve been surgically severed.

The discovery offers some radical new insights about the way neurons might be talking to one another, via a mysterious process unrelated to conventionally understood mechanisms, such as synaptic transmission, axonal transport, and gap junction connections.

“We don’t know yet the ‘So what?’ part of this discovery entirely,” says neural and biomedical engineer Dominique Durand from Case Western Reserve University.

“But we do know that this seems to be an entirely new form of communication in the brain, so we are very excited about this.”

Before this, scientists already knew there was more to neural communication than the above-mentioned connections that have been studied in detail, such as synaptic transmission.

For example, researchers have been aware for decades that the brain exhibits slow waves of neural oscillations whose purpose we don’t understand, but which appear in the cortex and hippocampus when we sleep, and so are hypothesised to play a part in memory consolidation.

“The functional relevance of this input‐ and output‐decoupled slow network rhythm remains a mystery,” explains neuroscientist Clayton Dickinson from the University of Alberta, who wasn’t involved in the new research but has discussed it in a perspective article.

“But [it’s] one that will probably be solved by an elucidation of both the cellular and the inter‐cellular mechanisms giving rise to it in the first place.”

To that end, Durand and his team investigated slow periodic activity in vitro, studying the brain waves in hippocampal slices extracted from decapitated mice.

What they found was that slow periodic activity can generate electric fields which in turn activate neighbouring cells, constituting a form of neural communication without chemical synaptic transmission or gap junctions.

“We’ve known about these waves for a long time, but no one knows their exact function and no one believed they could spontaneously propagate,” Durand says.

“I’ve been studying the hippocampus, itself just one small part of the brain, for 40 years and it keeps surprising me.”

This neural activity can actually be modulated – strengthened or blocked – by applying weak electrical fields and could be an analogue form of another cell communication method, called ephaptic coupling.

The team’s most radical finding was that these electrical fields can activate neurons through a complete gap in severed brain tissue, when the two pieces remain in close physical proximity.

“To ensure that the slice was completely cut, the two pieces of tissue were separated and then rejoined while a clear gap was observed under the surgical microscope,” the authors explain in their paper.

“The slow hippocampal periodic activity could indeed generate an event on the other side of a complete cut through the whole slice.”

If you think that sounds freaky, you’re not the only one. The review committee at The Journal of Physiology – in which the research has been published – insisted the experiments be completed again before agreeing to print the study.

Durand et al. dutifully complied, but sound pretty understanding of the cautiousness, all things considered, given the unprecedented weirdness of the observation they’re reporting.

“It was a jaw-dropping moment,” Durand says, “for us and for every scientist we told about this so far.”

“But every experiment we’ve done since to test it has confirmed it so far.”

It’ll take a lot more research to figure out if this bizarre form of neural communication is taking place in human brains – let alone decoding what exact function it performs – but for now, we’ve got new science that’s shocking in all kinds of ways, as Dickson adroitly observes.

“While it remains to be seen if the [findings] are relevant to spontaneous slow rhythms that occur in both cortical and hippocampal tissue in situ during sleep and sleep‐like states,” Dickson writes, “they should probably (and quite literally) electrify the field.”

The findings are reported in The Journal of Physiology.

https://www.sciencealert.com/neuroscientists-say-they-ve-found-an-entirely-new-form-of-neural-communication

Something Really Fascinating Happens When You Give Plants Anesthetic

261-plant-anaesthetics_1024

by PETER DOCKRILL

When somebody mentions anaesthetics, we probably think straight away of pain relief, but there’s a lot more going on in these complex chemical compounds than the simple negation of discomfort.

While there’s a range of chemicals that can induce anaesthesia in humans, just how these unrelated compounds trigger a lack of consciousness remains somewhat unclear.

And the mystery deepens when you consider it isn’t only animals that are affected by anaesthetics – plants are, too.

Humans in ancient societies were using things like herbs for various sedative purposes thousands of years ago, but the roots of modern anaesthesia began around the mid-19th century, when physicians began administering diethyl ether to patients during surgical procedures.

It was only a few decades later that scientists realised plants were similarly affected by ether, leading French physiologist Claude Bernard to conclude plants and animals shared a common biological essence that could be disrupted by anaesthetics.

261-plant-anaesthetics-2

A century and a half later, scientists are still investigating this strange commonality – basically by slipping plants the mickey and seeing what it does to them.

In a new study by Japanese and European researchers, the team filmed a number of plants that exhibit the phenomenon of rapid plant movement to see what kinds of anaesthetic chemicals affected them.

The sensitive plant (Mimosa pudica) usually closes its leaves in response to touch stimuli; but when exposed to diethyl ether, the dosed-up plants completely lost this response, becoming motionless, with the movement response only returning to normal after 7 hours.

In a separate experiment with the sensitive plants, a lidocaine solution also immobilised the leaves.

Similarly, the Venus flytrap (Dionaea muscipula) lost its ability to close its trap when exposed to diethyl ether – despite repeated prongings by the researchers – but the mechanism recovered in just 15 minutes.

Another carnivorous plant, Cape sundew (Drosera capensis), captures prey via sticky tentacles on its leaves, but experiments showed they lost the ability to bend their leaves and tentacles when exposed to the ether.

As for why plants are incapacitated by these chemicals, the researchers hypothesise it is to do with the inhibition of action potentials, preventing electrical impulses that help plants’ biological systems function.

“[B]ioelectricity and action potentials animate not only humans and animals but also plants,” the researchers explain.

“That animals/humans and also plants are animated via action potentials is of great importance for our ultimate understanding of the elusive nature of plant movements and plant-specific cognition/intelligence based plant behaviour.”

Ultimately, the team thinks these similarities between plant and animal reactions to anaesthetic compounds could lead to future research where plants might function as a substitute model or test system to explore human anaesthesia – something scientists are still pretty uncertain about.

It’s not easy being green, perhaps, but at least they shouldn’t feel any pain.

The findings are reported in Annals of Botany.

https://www.sciencealert.com/plants-respond-anaesthetics-weird-movement-action-brain

Discovery Reverses Wrinkles And Hair Loss in Mice

by PETER DOCKRILL

The appearance of wrinkled, weathered skin and the disappearance of hair are two of the regrettable hallmarks of getting older, but new research suggests these physical manifestations of ageing might not be permanent – and can potentially be reversed.

New experiments with mice show that by treating a mutation-based imbalance in mitochondrial function, animals that looked physically aged regrew hair and lost their wrinkles – restoring them to a healthy, youthful appearance in just weeks.

“To our knowledge, this observation is unprecedented,” says geneticist Keshav Singh from the University of Alabama at Birmingham.

One of the focal points of anti-ageing research is investigating the so-called mitochondrial theory of ageing, which posits that mutations in the DNA of our mitochondria – the ‘powerhouse of the cell’ – contribute over time to defects in these organelles, giving rise to ageing itself, associated chronic diseases, and other human pathologies.

To investigate these mechanisms, Singh and fellow researchers genetically modified mice to have depleted mitochondrial DNA (mtDNA).

They did this by adding the antibiotic doxycycline to the food and drinking water of transgenic mice. This turned on a mutation which causes mitochondrial dysfunction and depletes their healthy levels of mtDNA.

In the space of eight weeks, the previously healthy mice developed numerous physical changes reminiscent of natural ageing: greying and significantly thinning hair, wrinkled skin, along with slowed movements and lethargy.

The depleted mice also showed an increased numbers of skin cells, contributing to an abnormal thickening of the outer layer of their skin, in addition to dysfunctional hair follicles, and an imbalance between enzymes and inhibitors that usually prevents collagen fibres from wrinkling skin.

But once the doxycycline was no longer fed to the animals, and their mitochondria could get back to doing what they do best, the mice regained their healthy, youthful appearance within just four weeks.

Effectively, they reverted to the animals they were before their mitochondrial DNA content was tampered with – which could mean mitochondria are reversible regulators of skin ageing and hair loss.

“It suggests that epigenetic mechanisms underlying mitochondria-to-nucleus cross-talk must play an important role in the restoration of normal skin and hair phenotype,” says Singh.

“Further experiments are required to determine whether phenotypic changes in other organs can also be reversed to wildtype level by restoration of mitochondrial DNA.”

Even though the mitochondrial depletion affected the entire animal, for the most part the induced mutation did not seem to greatly affect other organs – suggesting hair and skin tissue are most susceptible to the depletion.

But it could also mean the discovery here isn’t the fountain of youth for slowing or reversing the wider physiological causes of ageing – only its more surface, cosmetic symptoms. Although, at least some in the scientific community aren’t persuaded yet.

“While this is a clever proof of principle, I am not convinced of the clinical relevance of this,” biologist Lindsay Wu, from the Laboratory for Ageing Research at the University of New South Wales, who was not involved in the study, told ScienceAlert.

“The rate of mitochondrial DNA mutations here is many orders of magnitude higher than the rate of mitochondrial DNA mutations observed during normal ageing.”

“I would be really keen to see what happens when they turn down the rate of mutations to a lower level more relevant to normal ageing,” Wu added.

In that vein – with further research, and assuming these effects can be replicated outside the bodies of mice, which isn’t yet known – it’s possible this could turn out to be a major discovery in the field.

For their part, at least, the researchers are convinced mtDNA mutations can teach us a lot more about how the clocks in our bodies might be stopped (or wound back to another time entirely).

“This mouse model should provide an unprecedented opportunity for the development of preventative and therapeutic drug development strategies to augment the mitochondrial functions for the treatment of ageing-associated skin and hair pathology,” the authors write in their paper, “and other human diseases in which mitochondrial dysfunction plays a significant role.”

The findings are reported in Cell Death and Disease.

https://www.sciencealert.com/unprecedented-dna-discovery-actually-reverses-wrinkles-and-hair-loss-mitochondria-mutation-mtdna