Dr. Ryan Camacho, an Associate Professor in the Electrical and Computer Engineering department at Brigham Young University, is at the forefront of new technology and innovation.
Dr. Ryan Camacho, an Associate Professor in the Electrical and Computer Engineering department at Brigham Young University, is at the forefront of new technology and innovation. In his latest work, Accelerating Silicon Photonic Parameter Extraction Using Artificial Neural Networks, which was published on June 12, 2019 in The Optical Society Continuum Journal, he and his co-authors explored novel silicon photonic parameter extraction tools that use artificial neural networks. “In this department we design a lot of circuits, digital logic, and computing things using electronics,” says Camacho, “What if we could do everything that electronics does but with photons instead of electrons?’ We’re in [the] early days of this.”
Dr. Camacho predicts that this exciting new technology will be especially useful in communications. He explained that similar to fiber-optic cables, which transmit data by turning a light on and off or shifting the phase of the light back and forth, silicon photonics utilizes the same basic principle by creating “little wave guides, almost like wires, that guide the light and create interference between other bits of light [which then transmit] data that we can then process..” Though this advancement has many promising applications, Dr. Camacho and his team have faced many challenges in the design and development processes. “In contrast electronic circuits, we don’t know exactly how to design complex photonic circuits yet, even at the component level” Camacho says. “A lot of the time in electronics we’ll say ‘oh we have a capacitor or an inductor or some transistors’ and people are working to design faster smaller versions of those. For the most part, however, the basic design was worked out50 years ago, and now we just work to make them smaller and faster, and there’s a lot of existing research. On the other hand, with light, we are in the early days and the best way to design components and circuits things is by no means obvious—how we’re going to do it, what’s going to work, what’s the equivalent of a capacitor for light, etc.” Dr. Camacho explained the method which he and his co-authors used to work through design issues by saying, “one thing you can do in electronics that you also do in photonics is you test your circuit, see how it’s working and then you’ll know—I tried to design one thing, but I actually got something different. This parameter extraction piece [helps us answer questions like], ‘what did I actually make? Did I make what I thought I designed? Is it working the way I thought?’”
Dr. Camacho credits one of his students, Alec Hammond, with proposing the initial idea on which the Article was based. “The whole machine learning approach is something Alec and I have learned together over the last year and a half. He came to me and said ‘I want to do something in machine learning.’ My specialty is the optics and designing photonics. . . [Alec] said we could do some machine learning here, so we just kind of learned it together and this is our first paper to come out as a result of that. It’s exciting for us to see our work come so far.” Hammond is one of two students who are credited as co-authors on the article. Easton Potokar, an undergraduate student in the math department at Brigham Young University, also contributed greatly by creating and running the neural networks for the project.
Dr Camacho expressed his excitement about his team’s discoveries by saying, “we’ve increased the ability to extract parameters tremendously. We did some benchmarking in this research and it’s almost ten thousand times faster than the next best approach. So if you think about how many things you can design and look at, it’s a real game changer. We’ve had some interest from companies and we’re cautiously optimistic that we’ll have a real impact.” Looking forward, Dr. Camacho hopes his innovations will be able to produce more efficient and more capable circuits that are better equipped to process information. He also hopes to be able to apply these new technologies to quantum computing and quantum information processing. Camacho and his team were thrilled to receive the Editor’s Pick Award from the OSA Continuum and are excited to see how their work continues to impact existing technology.