12.5 C
New York
Saturday, July 4, 2020
Home News Good at StarCraft? DARPA wants to train military robots with your brain...

Good at StarCraft? DARPA wants to train military robots with your brain waves

Douglas Levere, University at Buffalo

The 1984 movie The Last Starfighter tells the story of a teenager whose calling in life seems to be nothing more than to play arcade games. Fortunately, he’s spectacularly good at it. The game he’s best at is a video game called, as the movie’s title would have it, Starfighter. In it, the player must defend their homestead, The Frontier, from the perils of Xur and the Ko-Dan Armada, by way of a series of wireframe laser battles.

But there’s a twist. It turns out that Starfighter isn’t simply a game; it’s actually a kind of test. The war with Xur and the Ko-Dan Armada is real, and the arcade game — with its demands on rapid-fire reaction times on the part of players — is a stealth recruiting tool, intended to seek out the best of the best to become genuine starfighters.

More than 35 years after The Last Starfighter hit theaters, engineers from the University at Buffalo, New York, Artificial Intelligence Institute have received funding from DARPA, the U.S. Defense Advanced Research Projects Agency, to carry out research that’s… well, let’s just say that it’s extremely similar. They have built a real-time strategy game, currently unnamed, that’s reminiscent of existing games like StarCraft or Stellaris in style. In this game, players must use resources to build units and defeat enemies; manipulating large numbers of agents on-screen to complete their mission objectives.

But this isn’t any ordinary gaming experience. When people play the University at Buffalo’s new strategy game, they first have to agree to be hooked up to electroencephalogram (EEG) technology so that the game’s designers can record their brain activity. As they play, their eye movements are also tracked by way of special ultra high-speed cameras to see exactly how they respond to what they’re doing. This information, which can be teased out using machine learning algorithms, will then be used to develop new algorithms that can help train large numbers of future robots. In particular, the hope is that these insights into complex decision-making can improve coordination between large teams of autonomous air and ground robots. You know, should the game be brought to life.

Patrik Stollarz/Stringer/Getty Images

For anyone who grew up on movies like The Last Starfighter, this will seem strangely familiar. Although there’s a twist here, too. In The Last Starfighter (and other sci-fi stories which tread similar ground, such as Orson Scott Card’s Ender’s Game and Ernest Cline’s Armada), the goal is to train humans to have the kind of lightning fast reflexes that would normally be found in a machine. In this case, it’s different. The purpose of the University at Buffalo’s new gaming project isn’t to make players more machine-like.

Just the opposite, in fact. It’s all about trying to make machines that think more like humans.

Training tomorrow’s swarms today

“We’re trying to recruit [participants] who have strong gaming experience,” Souma Chowdhury, assistant professor of mechanical and aerospace engineering in the School of Engineering and Applied Sciences, told Digital Trends.

Chowdhury is one of the lead investigators on the project. He pauses and gives a nervous chuckle; the slightest hint of an apology creeping into his voice. “I myself do not have gaming experience,” he said. “I’m not a computer gamer at all. But many of our students are into games like crazy.”

“We’re trying to recruit [participants] who have strong gaming experience.”

Chowdhury’s own area of interest is swarm intelligence, a branch of computer science dating back to the late-1980s. Swarm intelligence is all about the collective behavior of decentralized, self-organized systems, both virtual and robotic. “It’s a real hot topic,” he said. “It’s becoming known that there are a lot of different applications which could be done by not using a single $1 million robot, but rather a large swarm of simpler, cheaper robots. These could be ground-based, air-based, or a combination of those two approaches.”

Some researchers in swarm robotics try and create swarms that can carry out complex procedures by hand-crafting the actions of every agent involved; the way you might coach each member of a dance troop so they can master a complex routine. Put them all together and you’ll get something that looks like emergent collaboration, although it’s actually a collection of individuals doing their own thing. The idea of using modern machine learning artificial intelligence is that it could give robot swarms the ability to more autonomous function as a meaningful collective.

Douglas Levere, University at Buffalo

But that’s easier said than done. Training one robot to do something requires a significant amount of training. Training a swarm, potentially with varying abilities, to complete tasks in complex, uncertain environments is a whole lot trickier. It means running tens of thousands of simulations, making the process extremely time-consuming and expensive. The idea driving this new project is that watching humans play the game will make it easier for machines to learn.

“Imagine walking into a classroom where there’s no teacher, and saying ‘let’s learn algebra,’” Chowdhury said. “You can learn just using exercises and textbooks. But it’s going to take a lot more time. If you have a teacher you can follow it’ll make it faster. In this case, we want to see how humans play this game and then use that to significantly speed up the A.I. in learning the behavior. Before it would be necessary to run 10,000 simulations to learn. Now we only need to run perhaps 1,000 simulations and augment this with data from humans.”

The researchers believe that, by observing the type of tactical or strategic decisions humans take when they play a strategy game, it will be possible to work out which features and events motivate these actions.

Teaching the machines

“The project is ongoing, at a pretty aggressive pace,” Chowdhury said. “We are around the halfway mark.”

At present, they’ve yet to start the start the data-gathering phase of the project, although Chowdhury has a good idea of the format that it will take. The plan is to carry out experiments with around 25 participants. Each participant will play between six and seven games with different randomized settings and levels of complexity. Unlike games such as StarCraft, which can last for hours, in this case each game will go on for only last only between five and ten minutes. That will be sufficient to measure decision-making strategies, and for these features of interest to be extracted using algorithms and scripts developed by the team.

“Humans can come up with very unique strategies that an A.I. might not ever learn.”

“At this point, it is difficult to comment on the amount or size of data that will be eventually collected,” Chowdhury said. However, the aim is reportedly to eventually scale up to 250 aerial and ground robots, working in highly complex situations. One example might be dealing with sudden loss of visibility due to smoke. The team plans to develop algorithms, modeled on human behavior, that will allow them to adapt to challenges such as this.

“Humans can come up with very unique strategies that an A.I. might not ever learn,” he continued. “A lot of the hype we see in A.I. are in applications that are relatively deterministic environments, like chess or Go. Those are very complicated games, but the number of possibilities are still countable. Maybe they’re in the trillions, but they’re countable. But in terms of contextual reasoning in a real environment to get stuff done? That’s still at a nascent stage.”

Humans make the strategies

In Daniel Kahneman’s 2011 book Thinking, Fast and Slow, the Nobel-winning economist and psychologist describes two different modes of thought. The first system is fast and instinctive, the kind of thing we might call intuition. That might be locating the source of a specific sound, completing the phrase “war and…” or, yes, blasting Ko-Dan ships out of the air (or lack thereof) in Starfighter. The second system is slower, more deliberate, more logical. It’s centered on conscious thinking — which in this case might very well refer to forming strategies.

Chowdhury doesn’t cite Kahneman’s work when he discusses the project. But it’s hard not to be reminded of it. As he points out, machines are already capable of an impressive number of autonomous features. A $10,000 drone possesses some impressive smarts when it comes to navigating between locations. The same is true with agents in a strategy game. Units are often governed by low level rules which allow them to react to their surroundings. That could mean attacking or defending if they are confronted by an enemy. It might also mean being able to maintain formations as they move around the map. But in both cases what’s missing is the overarching strategy needed to execute tasks.

“You don’t need a human to do low level control, controlling each agent,” Chowdhury said. “That’s not what we’re interested in. They’re not controlling every single robot and where they’re going. The human role is more that of a supervisor or a tactician. A good analogy would be that, in a disaster response environment, you have a supervisor. They might have a team of 100 rescuers working under them. There’s a hierarchy, but the supervisor does not tell each of those team members exactly what they should do. The rescuers make a lot of independent decisions, but the supervisor creates the overall tactics. That’s what we want to build.”

If Chowdhury and his team get their way, the robot of swarms of tomorrow will be a whole lot smarter. And they’ll have gamers to thank for it.

Latest

Sennheiser GSP 670 headset review: premium price, subpar performance

The search for a new headset can really get frustrating. Sure, there are a million options on Amazon for under $50, but when you want something premium, where do you start? If you’re looking for the best possible audio quality, you start with the Sennheiser GSP 670 and hope you can find it on sale because these things don’t come cheap.The GSP 670 is a premium headset with sound quality and a price tag to match. Launching at $350, you’re paying for the Sennheiser name and quality. We’ve tested multiple Sennheiser headsets throughout the years and have almost always come away impressed. That’s the same story here.The first thing you may notice about this headset is just how big it is. It looks big before you pick it up and it feels big once you put it on. Coming in at just shy of 400g, it has the weight to make those extremely long gaming sessions taxing, but luckily Sennheiser included one of the best headbands I’ve seen in a headset yet. It’s big and comfortable without looking too ridiculous.The earcups are equally nice with large plus fabric cups that will keep your ears away from the driver covers. If you prefer leatherette cups you’ll want to find another option, but I did find these to be one of the most comfortable headsets to just sit and listen to music on. The clamping force is just right (although uneven; more on that later) and the earcups provide a wonderful seal to keep the noise of the world away from your ears.One the outside of the headset, there’s a small tactile wheel to adjust chat volume if you’re using a gaming console, a large volume knob, and a multifunction button that will provide audio prompts for battery level and put you into pairing mode when you hold it down. The only thing we’re missing here is a physical switch to move between Bluetooth and 2.4ghz connection standards, and we’ll tell you why that matters in a bit.The microphone is on the left side of the headset and provides a nice tactile click when you flip it all the way up. This is how you mute your microphone and comes in handy when you need to have a quick conversation and get back to whatever you were doing before.I wish I could report that the microphone provided better audio quality but I was pretty disappointed. It’s been a struggle to find a wireless headset that really gives great performance in this area (I’m guessing there’s a bandwidth issue) and the Sennheisers fall disappointingly short. I think they sound much the same as every other headset released in the last decade, which isn’t saying a lot.Both Bluetooth and 2.4ghz connection standards are here. Plugging the USB dongle into my computer, the headset paired almost instantly and opened up a world of opportunity to tune through the Sennheiser app. There are options to tune your EQ, how the microphone sounds, and even provide a noise gate in case you have a noisy background. I didn’t find much difference in how the microphone sounded using these options so hopefully, they continue to be tuned in future updates.The sound that comes through these headphones is a completely different story. This has been one of the best audio experiences I’ve had in my time reviewing tech. I’d put it up there with the Sony WH-1000xm3 in terms of enjoyment. Where Sony offers amazing noise cancelation, the Sennheiser GSP 670 takes the crown in terms of audio quality.I found music pleasingly bass-y without feeling like I’m slogging through the mud just to listen. Mids are very clear while highs are crisp without being piercing.I just wish I enjoyed wearing these more. I can’t overstate how heavy these things are. At just under 400g, they’re one of the heavier headsets I’ve tested and it can be exhausting during long sessions. With 16 hours of battery life, those sessions can last all night, but you’ll need breaks.Additionally, I don’t like wearing these because of how the cups sit on my head. While the cups themselves are large enough that my ear doesn’t touch anything, the clamping is uneven and annoying. You can use the sliders in the headband to adjust your clamp, but I always end up with more pressure on the bottom of the cups than at the top.Frankly, these don’t look great and certainly don’t look like something I’d pay over $300 for. They’re big and bulky with muted colors and an … aggressive? design. I’m not entirely sure what to call this design language but there are definitely better-looking options on the market. This won’t matter to some, but for those who do care, it’s a bit of a killer and makes the cost harder to justify.ConclusionThere are always trade-offs when you’re using a wireless headset. Sennheiser smartly did not skimp on the audio quality and if you’re looking for a wireless headset that sounds great, this is definitely where you want to start. I put it at the top of the list in that respect.But, where it falls apart is pretty much everywhere else. Tradeoffs become pretty obvious when you use these for more than a few hours.Yep, they’re built solidly and the plastic design means they’ll hold up to some abuse. But, these look cheaper than competing options like the Astro A50s and Arctis Pro Wireless. Plus, as I’ve said a few times, they’re heavy.It’s awesome that they have both 2.4ghz and Bluetooth standards. But there’s no way to manually switch between them and the second that your computer plays audio via the USB dongle, the Bluetooth cuts out completely. If you’re using these to take a phone call or listen to music on your phone and you accidentally click on a YouTube link on your computer, say goodbye to your audio. This would be an easy fix with a manual switch and we hope to see that in a future revision.Best over-ear headphones (spring 2020)I can’t state enough how crappy the audio from the mic is. Maybe I’m spoiled by streamers who invest hundreds and hundreds of dollars into their audio equipment, but this sounds like every headset I’ve heard the last decade of gaming and that’s pretty disappointing.If your voice quality matters to you at all, I’d suggest getting a standalone mic. But you have to ask yourself if you’re grabbing something like a Blue Yeti, is there a justification for the GSP 670 when you can buy a wireless headset for far cheaper?I know it probably looks like I hate the Sennheiser GSP 670 but I don’t. In true dad fashion, I’m not mad, I’m just disappointed. While they’re best in class in terms of audio quality, the things they miss on are a killer and make them harder to recommend over other competitors.After a bit of searching, I’ve found the Sennheiser GSP 670 around $300 and sometimes cheaper on sale. I think if you can find these cheaper than that, go for it. Your ears will thank you. At full price, they’re a tough sell.Buy the Sennheiser GSP 670 at Amazon

Best bird identification apps for iOS and Android

The best bird guide may be on your smartphone. Here are some of our favorite birding apps for iOS and Android.

How long does your laptop last on a charge? Here’s how to find out

Finding exact values for laptop battery life is tricky. Here's what to know about precise battery measurement.

How to watch Nathan’s Famous Hot Dog-Eating Contest live from anywhere

The 2020 Fourth of July International Hot Dog-Eating Contest may not be in front of a live audience this year,