 Hello, everyone. Can't hear you. OK, so I know it's lunchtime, and you guys are hungry, but it's last Tokyo. So once again, thank you. Hello. Great. So my name is Gajan, and I'm from Reputed Robotics. And today I want to share some of my thoughts with you on how to make robots more accessible. I know the title was Democratizing Robots, and a marketing guy said, like, yeah, we should tone down a little bit. So science fiction, they gave us the name robots, and they have never failed in entertaining us and making our imagination run wild. These are the cool robots we thought of and animated. But what's the reality? This is a car manufacturing line in Slovakia, owned by the Kivas Motors. And the movies really talk about them, but these are one of the most common robots in the world as of today. We have figured out how to make them move in a way very precise and a speedy manner, but they're really dangerous to be around. You can see those fences, and that is to keep the humans away. And if a human accidentally enters, the whole system is designed to shut down or freeze. These are some other robots that are not in factory, but we see or hear about. One on the left is a robotic vacuum cleaner. The marketing team of those companies try to say they are smart, but we all know they're not, especially, like, even on the second, third day, it gets caught on the same bunch of cables. On the first day, you are like, oh, you poor thing. You got caught. Let me help you. But if it repeats the same thing on the third day, you are like, are you kidding me? Right? And then one on the right is the predator drone. It's an unmanned vehicle. Some people mistake it for autonomous vehicle. It's not the case. This is the predator drone control center far away, but miles away, controlled by really humans. So the drone actually just follows the human orders. So what am I trying to say here? Well, one thing is, of course, using robots for warfare is not very cool. But what I'm trying to say here is that we are expecting more from the robots. I'm sure you do, too. And so does our prime minister. He wants to create a revolution to support the aging population of robots, of Japan, to work more efficiently and live a very comfortable life. This is a great vision for robots. So what's keeping them from coming out of factories? So one is the robots, as I said, they're big, bulky, and dangerous to be around. They aren't as mobile and agile as we would like to have. The second thing is the real world is very chaotic. Look at this stage. This is very, very different from a factory floor, especially like, think of your living room. And especially if you have kids like me, that's very chaotic. For a robot, it has to constantly look at the environment, especially if the environment is changing. It has to constantly plan on what to do next in such a way it's safe for itself and the humans around it. The other problem of chaotic worlds is that you cannot plan ahead. And you cannot think of all the things that's going to happen. So the robot has to, by itself, think and do something about it. So there are challenges. But this is the last three years of data on funding and acquisitions for robots. So it's exponentially growing. Although there are challenges, what are these guys, especially the governments, the private equity funds, what are they actually betting on? They are betting on something great. Because we, I think, are living in a great time. We are seeing a confluence of technologies coming together and building components that are faster, cheaper, smaller, and lighter. So let's take a look at some of the key enablers that's going to make robots more accessible. So the first one is motors. The motors convert different forms of energy to mechanical energy. One on the left is the James Watt steam engine, which was a core element in the industrial revolution, occupied probably the whole stage, produced about 10 horsepower. But thanks to electricity, the battery technology, and our ability to make very strong permanent magnets, we are on the right. So these motors, using this drone I have, about five of them combined can produce one horsepower, and each of them cost about $100, which is great. And when you compare the power to volume ratio, that is, divide the power it produces divided by the occupied volume, these motors are more than 1,000 times better than the steam engine. The second one is sensors. Sensors help robot understand the environment or their state. So one on the left is a sensor that's used to measure the acceleration of the missile in 1980s. That's a ballistic missile from France. It's really bulky, and I don't think I can carry it. I'm sure it costs you an arm and leg. One on the right is a sensor encircled by the red circle is that it's the same sensor. Actually, it's a little bit much better than that. That will cost you less than 500 yen. And this is thanks to the microelectromechanical systems technologies we have now. The third enabler is the computation, onboard, especially the onboard computation we can have. One on the left is the world's first commercial computer, current pricing, 11 million US dollars. And it can do 800 instructions per second. But thanks to the more slow, thanks to the mass adaptation of cell phones, we are on the right. $10, and this thing can do 200 million instructions per second. Compared to the 800, you had to pay 11 million. Now, this is really great. So let's look at what this machine I have, built by my colleagues, can demonstrate. And it's a combination of these three enablers. And let's see what these things can do. So you can just throw a start. And they fly stably. And they can do flips. And they can do double flips, too. And you can be the Harry Potter, magically controlling the drone. You can shoot the drone, and the drone will try to avoid it. And the last thing is the coolest part is that it can balance the pole. I think some of us will fail, including me. So this kind of demonstrates that we are at a point of time that we can build these agile machines for a very fraction of the cost compared to 10 years ago. So now let's look at some of the other key enablers. So this is a picture of a data center owned by Google. And a data center is a combination of tens of thousands of powerful machines tightly connected together. And the cloud is, of course, a technology concept, but it's also a disruptive business model. What they're trying to do is that you don't have to own or carry your computer anymore. You can rent by the hour, computation, access service. This is great, right? Now, what if robots also have access to these things? Actually, they do. So these are two very low-cost robots placed here in my old office in Zurich, Switzerland. And they're trying to do one of the most computationally expensive tasks for robotics, 3D mapping. So what are they doing is they're just sending the sensor data to the Amazon's data sensor in Ireland miles away. And they're building awesome maps. And since they're all connected, the map gets merged, and it becomes much better. The last key enabler I want to mention is the software algorithm. The science of software algorithm is getting great, thanks to two things. One is abundant computation from the data centers. And the other one is the abundance of data that is available to train these algorithms. So the one on the left is called the ImageNet, which is about 10 million images. And every year, researchers get together and challenge algorithms to detect scenes and objects. And you can see the error going down, the accuracy going up. But in 2015, the algorithm beat the human ability to detect objects and scenes. This is crazy. So there are great key enablers. So what's holding robots back? So we saw the two problems I mentioned are kind of solved. The last one is that putting these five enablers, or even more something I didn't mention, is itself a big challenge. So I'm going to share some of my experience here. So during my PhD, I worked on a project called RoboEarth. It was funded by the European Union, and we had several partners across Europe. So one of our milestones was to operate three different robots in a hospital room. It was a mock hospital room. And so each partner developed their own component. We met together in Netherlands to put stuff together. So it took about 15 PhDs, two sleepless weeks, to put together a one-hour demo for the European Commission. Now, this is all the skills that we had, a great team. But what was missing is the integration. So one of the last things what's holding us back is that if it takes like 15 PhDs, two sleepless weeks to put one-hour research demo, this is not suited for business. Businesses need robots operating 24-7 with minimal support and maintenance because it's cost-sensitive. So is there a solution? I think there is one. That is, you build a platform. A platform that takes away the complexity of building and maintaining your hardware. It can produce a lot of a comprehensive set of tools for you to maintain your hardware. But in addition to that, it does something cool. It abstracts the hardware. So since there is an abstraction, what can happen, a hardware manufacturer doesn't have to deal with all the software folks out there trying to run software on this hardware. He can just focus on what he does best, building great hardware. Another great thing about this abstraction is that it levels the playground for robotic software developers. So up to now, if you want to do software for robotics, you have to know about robotics. This is very limiting because there are about maybe 1,000, even 10,000. But imagine this whole field opening up for all the software developers, millions. So you can have a computer vision expert, machine learning expert, a logistics expert, a health care expert, all contributing this field of robotics. This is very exciting. And the other thing the platform can do is it can allow robots to seamlessly tap into the power of computers, data centers. So this gives a robot an extended brain and a connected brain. With an extended brain, what can happen is that the robot can run very powerful algorithms to understand the environment and then act in a very smart way. And with a connected brain, what they can do is that they can learn from each other. They can improve their skills without actually them experiencing. This is something that we humans cannot do. My son has to go to school and start from scratch. So the last thing, now you have software developers having building blocks, abstracted hardware. Now let's think about the system integrator or someone who wants to build a solution. So here, he just have to do some connections. So this allows him to actually focus on the real problem, which is beyond robotics and all the integrity details. And we are building such a platform. So we are a bunch of about 50 entrepreneurs who are trying to make a dent in the universe, as we told, as was covered in the last thing. And we are engineering such a platform. So what is the last missing part? The last missing part is a strong community around that platform, which is you all. So here is a shout out to all the software developers, hardware developers, and anyone who wants to build an application to empower lives. Let's join together and make this happen. So we are looking for partners, and we are hurrying. Thank you very much. So I finished three minutes early. You can grab those awesome hamburgers. They're pretty good. Thank you.