Technical Tuesday – 28 February 2017 – Dataflow Machines and GPUs in Signal Processing for Wireless Communications by Dr. Joseph Mitola of Virginia Tech’s Hume Center

Posted by on Feb 28, 2017 in Events Archive | No Comments

Conventional computing and communications expose myriad attack surfaces because of the Turing-equivalence of the instruction set architectures and the mathematical impossibility of forming a complete set of monitor functions to protect the contents of the registers from insightfully designed malware such as what NIST terms Advanced Persistent Threats. Dr. Mitola directs the Information Systems Laboratory of Virginia Tech’s Hume Center for National Security and Technology that is developing dataflow computing on FPGAs and RF Deep Learning.  Data flow flexibility is established via embedded machine learning, while trustworthiness is established by machine DNA built into the hardware. FPGAs and GPUs use about 1/10th the power for networked computing and signal processing tasks so they are both green and inherently more secure than COTS blade servers. This presentation will present current results in RF deep learning and dataflow for radio signal processing. Dr. Mitola also will explain his role as Distinguished Professor at Nanjing University of Posts and Telecommunications (NUPT) in China with highlights of NUPT’s 5G wireless and cybersecurity research.

Presented by: Dr. Joseph Mitola III

Dr. Joseph Mitola III is recognized globally as the “godfather” of software defined and cognitive radio having coined both terms, written the most widely cited peer reviewed technical papers, and written the first two graduate text books on software defined and cognitive radio architectures.  His career has included positions of technical leadership with the US DoD, ITT Corporation, E-Systems, Harris Corporation, Advanced Decision Systems, The MITRE Corporation, and Stevens Institute of Technology. Career highlights include serving as Special Assistant for Trustworthy Cognitive Systems reporting jointly to the Director of DARPA and to the Deputy Director of NSA (2001-2005). His education includes BS in EE (Northeastern University), MS in stochastic optimal control (The Johns Hopkins University) and Doctorate in Teleinformatics (KTH, The Royal Institute of Technology, Stockholm, Sweden). Dr. Mitola provides software defined and cognitive systems subject matter expertise to law firms, commercial entities and to the US DoD, currently serving on the Defense Science Board Task Force on MILSATCOM and Tactical Networking. As Director of the Information Systems Laboratory of Virginia Tech’s Hume Center, he directs research in machine learning and cybersecurity for next-generation cognitive information systems.