 I'm here checking out the Lattice semiconductor FPGA right here, and hi, so who are you? Hi, I'm Yuzo Shimoyama from Lattice Semiconductor Japan, and this we are demonstrating our latest FPGA demonstration, which is an increment, a convolutional neural network into our smallest FPGA. So you do AI? Yes, AI inside of this chip, without using any cloud service. So how does it work? So we implement a neural network into this small FPGA. You can see it detecting up. And gesture, it will detect three signs of it. Log, scissors, and power, paper. So there's a little camera there? Yes, this IR sensor is located here. This IR sensor? Yes. And because of the neural network, it can recognize different hands, different shape. If the customer is training another network, they can detect other objects. Nice, is this already in the market or is it brand new? It's brand new. It's brand new demonstration. It's because FPGA, you can do everything you want. So you can put the end neural network or you can do other things with the same chip. Yes. So can you show some other demo with the chip? Sure. Why don't we go over there? This is more larger. So inside here, what is it doing? It's also another neural network AI is doing. So kind of we are detecting a human presence. There's a number of humans detecting. And one of them goes showing it up. Well, we detect a human being. So you can detect all the humans with a small FPGA? Yes. Very small one. It's still a little bit larger. This is our largest FPGA in our lineup. Yes. But it still has a 10 by 10 millimeter. It's second ball. Second ball. Do you have it here also? This one is completely another one. It's not a neural network AI. It's a security. Is this the big chip? Is that your chip? Yes. It's our test chip. This one is implemented. There's some security hardware engine inside the FPGA. So it will protect the entire system. The entire system will protect multiple farmers from any hardware attacks. Is this the same as the one before? It's also using different FPGA, but different implementations. It's using a 3D depth sensor. It's using still camera here to camera here. And detecting the human species here. Some graze game is showing us some distance difference. So many implementations. Here you're doing some IoT stuff. IoT stuff. Is this neural network called something? Neural, yes. IoT signal application. There's a multiple temperature sensor here. It goes to this FPGA. It aggregates similar chip. So, Lattice has been doing FPGA for a long time? Yes, we are around 35 years. 35? Yep. So Lattice invented FPGA? No. The last one is... Is this top 3 or something? FPGA for a few companies. This is the fourth position. The fourth position. And then here's just another Lattice chip right here. Yes. This is also part of this making some kind of an assistant tool to designing a neural network. Here's a camera. It's using FPGA here. And they are using FPGA here. And using it combined with PC, users can easily start designing your network. Two, three years ago, Lattice acquired a silicon image, right? Yes. So that means all this stuff is also part of... That's a different section. But it's also part of the company. Yes, part of the company. But we are really focusing on FPGA recently. In this booth, it's just about FPGA? Just FPGA. Here's some other stuff going on here with this kind of stuff. This is... This one is a human detection. Also the human detection. Another human detector. This is a face. So I can recognize my face. Recognize how many faces come through. And what is this? This is what you call Sensei. So the Sensei is part of your portfolio. It's under FPGA. Sensei is our neural network development stack. So including hardware, operation board, and the neural network IP core, and the compiler, and the different design, and the support and design service partners. Open source? Yes, open source. And also some parties, which are some IPs. But it's completely stuck to the AI. So people can customize it for exactly what they want? Yes. Accelerate even more. Some other faces. Face solution right here. This one... This one is also... This one is also... Video interface bridge. We have two hardware and MIPI interface. Transceivers. And another. Another direct interface we have on our... IOS. So you can bridge MIPI interface to another. Or another to MIPI interface. So you can bridge MIPI interface. And that's useful for... for VR or camera, or something on video interface. Which not directly connect to camera to system on chip. Translate. Put it over. Here. The design. They're all in production. And they are stitching four cameras. And stitching those images. So it's a real-time stitching. So real-time stitching. Camera going on here. So is Lattice Semiconductor is American company? Yes, American company. But what's the office in Japan? Yes, we have Japanese office, Japanese office here. What's happening in Japan? All this stuff? Yes. All the AI? No, no. Of course we are part of the company. So headquarters. And our Silicon Valley office. And the Shanghai office. And the Japan office. We work together to design those solutions. So very advanced. And millions of devices are using in the world, right? It's big. It's everywhere. Is it cars? Yes, some cars. Industry, consumers. Wide variety of customers we have.