As of January 2019, we’re changing our hiring process. We’re ditching technical presentations and the usual whiteboard interview, and adopting a process that reflects what we care about and how we want to treat people as an employer. This post shares our philosophy to hiring, what we hire for, and what our new interview process looks like.
It is clear that artificial intelligence will have a significant role to play in both improving the performance of existing wireless systems and in the fundamental capability of future systems, including 5G consumer wireless systems. DeepSig’s CTO and pioneer in the field, Tim O’Shea, is pleased to participate on a panel discussion on AI and 5G at IEEE Globecom 2018, from Dec 9-13 in Dubai. Dr. O’Shea will also give a podium talk on DeepSig's applied research and commercialization efforts, and is serving as the TPC for WS-18: Machine Learning for Communications.
DeepSig will once again be speaking at NVIDIA’s flagship conference, GTC Silicon Valley, between March 17th and 21st of 2019. The DeepSig team was previously invited to speak at GTC Silicon Valley 2018 and then at GTC DC 2018, and we are excited to once again be contributing to the conference program.
Edge deployments and low-SWAP operating environments have been an important usage model for DeepSig’s commercial software products since the beginning, and we are excited to share some of our work and lessons learned in this area. Embedded applications have always been challenging for RF sensing due to the data and processing rates traditionally required, but DeepSig’s approach using deep learning is able to both improve performance and reduce power consumption.
Our presentation is titled, “Machine Learning for Wireless Communications using TensorRT and NVIDIA Xavier”, with talk ID S9693. Once the conference agenda is finalized, we will post more information here about the date and time of the talk!
The well known publication Military Embedded Systems has published an article about the application of artificial intelligence to wireless systems, with a specific focus on military communications and radio systems. The article was co-authored with Ettus Research.
The article can be found in the Nov/Dec 2018 print edition of Mil-Embedded, and on the publication website, here: http://mil-embedded.com/articles/ai-military-systems/
DeepSig gave the first public demonstration of a channel autoencoder running over-the-air at IEEE DySPAN 2018 in Seoul, South Korea. Co-founder & CTO Tim O’Shea presented DeepSig’s OmniPHY commercial software product for learned physical layers, showing it running live between two software-defined radios. This uses the same software product that’s being tested over NASA’s satellite system, TDRSS.
We will be presenting at NVIDIA's GPU Technology Conference (GTC) DC 2018, which is taking place October 22nd - 24th in Washington D.C. Building on the work presented in our GTC Silicon Valley 2018 Talk, and expanded upon in an article featured by NVIDIA, we will be presenting "Deep Learning for RF Sensing and Communications". Information about the talk can be found on the NVIDIA GTC DC website.
The abstract for the talk is below:
Machine learning is rapidly advancing the state-of-the-art in algorithm performance for wireless telecommunications systems. Building on our work presented at GTC Silicon Valley, recasting fundamental wireless signal processing problems as data-centric deep learning problems, we present further evidence that learned signal processing algorithms can empower the next generation of wireless systems with significant reductions in power consumption and improvements in density, throughput, and accuracy when compared to the brittle and manually designed systems of today. This talk will introduce the core enabling technologies and fundamental approaches, share our latest work and results in deep learning-based sensing and learned communications, and discuss applications such as 5G and IoT, commercial cyber-threat sensing, and defense RF sensing to illustrate the wide range of fields these technologies will impact over the next several years.
If you'll be at GTC DC and are interested in chatting, please send us a note at email@example.com! We would love to meet up at the conference or host you in our offices just outside of D.C. while you're in town.
Technology invented at Virginia Tech that harnesses a new area of artificial intelligence to improve wireless performance and defend wireless devices has taken a major step toward the public market.
The university and an Arlington-based start-up, DeepSig, recently executed a licensing agreement that allows the company to further develop for consumer use the innovative wireless communications and cybersecurity technology invented by researchers at Virginia Tech’s Hume Center for National Security and Technology.
“This technology leverages a field of artificial intelligence called machine learning in a new way in order to design a next generation of powerful wireless communications systems,” said Virginia Tech researcher and DeepSig founder Tim O’Shea. “It will be faster, more cost efficient, more secure, and easier to deploy than today’s wireless systems.”
The complexity of wireless system design is continually growing. Communications engineering strives to further improve metrics like throughput and interference robustness while simultaneously scaling to support the explosion of low-cost wireless devices. These often-competing needs makes system complexity intractable. Furthermore, algorithmic and hardware components are designed separately, then optimized, and integrated to form complete systems. This approach makes globally optimizing the end-to-end communications link extremely difficult, if not impossible.
DeepSig overcomes this complexity barrier by designing neural networks that learn how to effectively communicate, even under harsh impairments. To accomplish this, we leverage our background in radio and signal processing, recent developments in deep learning, and technology from NVIDIA such as GPU hardware and software libraries optimized for machine learning. Our work in learned communications demonstrates that machines easily match the performance of human-designed systems in simple scenarios, as shown in Figure 1. In more complex scenarios, a deep learning-based system can dramatically outperform existing approaches by learning a physical layer (PHY) inherently optimized for the radio hardware and channel.
The recording of our tech talk at NVIDIA's GTC Silicon Valley 2018 is live! You can view it here on NVIDIA's website: http://on-demand.gputechconf.com/gtc/2018/video/S8791/
One of our Principal Engineers, Nathan West, gave an invited talk at NEWSDR 2018 in March of this year. The recording of the talk just went live, and includes our first public disclosure of some of our real-time RF sensing work and learned physical layers. If you're new to the field of deep learning for signal processing and communications, this video is a great way to catch up on the latest scientific advances.