EEC 290 – Seminar Series
Location: Kleiber Hall 3
All are welcome.
Prof. Robert Cui, University of California, Davis
Learning over Limited Data Samples: Detection of Cooperative Interactions in Logistic Regression Models
Cui Seminar Slides
An important problem in the field of learning is to identify interactive effects among profiled variables for outcome prediction. This becomes more challenging if the available data samples (i.e., labeled data) are limited. In this talk, a logistic regression model with pairwise interactions among a set of binary covariates is considered. Modeling the structure of the interactions by a graph, our goal is to recover the interaction graph from independently identically distributed (i.i.d.) samples of the covariates and the outcome. When viewed as a feature selection problem, a simple quantity called influence is proposed as a measure of the marginal effects of the interaction terms on the outcome. For the case when the underlying interaction graph is known to be acyclic, it is shown that a simple algorithm that is based on a maximum-weight spanning tree with respect to the plug-in estimates of the influences not only has strong theoretical performance guarantees, but can also significantly outperform generic feature selection algorithms, for recovering the interaction graph from i.i.d. samples of the covariates and the outcome. Our results can be also extended to the model that includes both individual effects and pairwise interactions by constructing an auxiliary model.
Shuguang Cui received his Ph.D in Electrical Engineering from Stanford University, California, USA, in 2005. Afterwards, he has been working as assistant, associate, and full professor in Electrical and Computer Engineering at the Univ. of Arizona and Texas A&M University. He is currently a full professor in Electrical and Computer Engineering at the Univ. of California-Davis. His current research interests focus on data oriented large-scale information analysis and system design, including large-scale distributed estimation and detection, information theoretical approaches for large data set analysis, complex cyber-physical system design, and cognitive network optimization. He was selected as the Thomson Reuters Highly Cited Researcher and listed in the Worlds’ Most Influential Scientific Minds by ScienceWatch in 2014. He was the recipient of the IEEE Signal Processing Society 2012 Best Paper Award. He has served as the general co-chair and TPC co-chairs for many IEEE conferences. He has also been serving as the area editor for IEEE Signal Processing Magazine, and associate editors for IEEE Transactions on Big Data, IEEE Transactions on Signal Processing, IEEE JSAC Series on Green Communications and Networking, and IEEE Transactions on Wireless Communications. He was the elected member for IEEE Signal Processing Society SPCOM Technical Committee (2009~2014) and the elected Vice Chair for IEEE ComSoc Wireless Technical Committee (2015~2016). He is a member of the Steering Committee for both IEEE Transactions on Big Data and IEEE Transactions on Cognitive Communications and Networking. He is also a member of the IEEE ComSoc Emerging Technology Committee. He was elected as an IEEE Fellow in 2013 and an IEEE ComSoc Distinguished Lecturer in 2014.
Prof. Daniel Feezell, The University of New Mexico
Nonpolar and Semipolar III-Nitride Optoelectronic Materials and Devices
Feezell Seminar Slides
III-nitrides, including alloys of AlN, GaN, and InN, have revolutionized light-emitting diodes (LEDs) for solid-state lighting, diode lasers for projection and high-density optical data storage, and electronics for power switching. Despite these significant advances, III-nitride materials and devices remain relatively immature compared to conventional III-V semiconductors and numerous opportunities for research on complex materials challenges and novel device concepts still exist. Conventional III-nitride devices are grown on the polar c-plane of the wurtzite crystal and their performance is adversely affected by the presence of internal polarization-related electric fields. Alternatively, growth of III-nitride structures on nonpolar and semipolar orientations presents a viable approach to reducing or eliminating the issues associated with polarization-related electric fields. These orientations also offer increased design flexibility and provide a multitude of unique physical properties. In this talk, I will highlight the unique properties and advantages of nonpolar and semipolar III-nitrides, present two techniques for the epitaxial growth of these materials, and discuss the application of this platform to important optoelectronic devices, including LEDs and vertical-cavity surface-emitting lasers (VCSELs).
Daniel Feezell is an Assistant Professor in the Electrical and Computer Engineering Department and the Center for High Technology Materials (CHTM) at the University of New Mexico (UNM). Dr. Feezell received the Ph.D. degree in 2005 at the University of California Santa Barbara (UCSB) for work on long-wavelength InP-based vertical-cavity surface-emitting lasers (VCSELs). Prior to joining UNM, he was a Project Scientist in the Solid-State Lighting and Energy Center at UCSB and a Senior Device Scientist and the first employee at Soraa, Inc. where he worked on III-nitride light-emitting diodes (LEDs) and diode lasers. His current research interests include epitaxial growth, fabrication, and characterization of III-nitride materials and devices, including nonpolar and semipolar orientations; solid-state lighting and high-efficiency LEDs; nanoscale selective-area epitaxy; and edge-emitting and vertical-cavity surface-emitting lasers. In 2013, he received a Defense Advanced Research Projects Agency (DARPA) Young Faculty Award, with a Director’s Fellowship Extension in 2015. He also received a National Science Foundation Faculty Early Career Development (CAREER) Award in 2015. Dr. Feezell is a Senior Member of IEEE and the Sources Thrust Leader in the Smart Lighting Engineering Research Center (ERC). He has authored or co-authored over 80 journal and conference publications and holds several U.S. patents. Additional information can be found at http://www.feezellgroup.com
Prof. Umesh Mishra, University of California, Santa Barbara
Yes we GaN!
Thanks to the ingenuity and perseverance of engineers and scientists we are facing the prospect of living with the end of progress in the performance of Si-based devices. The next dominant semiconductor to serve the market alongside Si is emerging to be Gallium Nitride. We are all familiar with the photonic applications of GaN-based LEDs, such as traffic lights, back lights for flat panel displays, automotive headlamps and solid state home lighting. The market of these applications is well over $10B, an amazing growth since its invention in 1993. The next wave behind GaN-based lighting is the market of GaN-based electronics. The megaproblem that GaN-based electronics solves is advanced performance at higher efficiency. Tremendous progress has been made in this area with the market now embracing GaN-based RF solutions for both military and commercial applications as evidenced by offerings from Wolfspeed, SEDI, Qorvo and other companies and with 650V-rated GaN devices offered by Transphorm, GaN Systems, Panasonic and others. The lifetime for GaN RF device has been firmly established by the players mentioned and Transphorm has demonstrated both JEDEC qualification and exceptional lifetime with activation energies under both field and temperature accelerations for 650V-rated GaN power switching products. GaN-based electronic and photonic devices in the market today are based on a form of GaN called Ga-polar GaN. The future of GaN-based electronics using the next advance, N-polar GaN will be presented in this talk. Of course to paraphrase the only thing constant in life is change and GaN will also play out its lifetime of growth and new materials such as diamond and gallium Oxide will rise to keep innovation alive
Umesh K. Mishra received his B.Tech degree from IIT Kanpur, India in 1979, his M.S. from Lehigh University in 1981 and his Ph.D. from Cornell University in 1984. He has over 800 publications, 100 patents and is the co-founder of two companies, Nitres and Transphorm. He has had the privilege of advising over 50 Ph.D. students and is the recipient of the IEEE David Sarnoff Award and the Welker Medal from International Symposium of Compound Semiconductors (ISCS). He is a Fellow of the IEEE and the National Academy of Inventors and a Member of the National Academy of Engineering.
Dr. Ashwin Gopinath, Caltech
Bridging the gap at the bottom: Interfacing individual molecules and nanofabrication
Gopinath Seminar Slide
Conventional top-down nanofabrication, over the last six decades, has enabled almost all the complex electronic and optical devices around us. Parallely, inorganic/organic self-assembly has also been studied extensively since they enable structures with properties unattainable by any top-down method. While both these fields have independently matured, ongoing efforts to create ”hybrid nanostructures” by combining the two has been fraught with technical challenges. One the main roadblock is the absence of a scalable method to precisely organize components built bottom-up (like molecules, single ions, quantum dots or nanoparticles) within top-down nanofabricated devices.
In this talk, I will introduce a lithographically directed self-assembly technique as possible solution to the aforementioned roadblock . The method primarily involves utilizing DNA origami, a technique to create arbitrary 2/3D DNA nanoparticles, as a modular adaptor to bridge the atomic/molecular legnthscale and lithographic lengthscale. Additionally, as a proof of principle, the technique is used to organize discrete light emitters inside nanofabricated photonic crystal cavities (PCCs) for measuring as well as precisely controlling the optical properties of the PCCs. I will conclude by giving a few examples of functional optical devices and high throughput biophysics tools that can be realized using this fabrication technique.
Ashwin Gopinath was born and raised in India, he obtained a BS in Electrical engineering from Visvesvaraya Technological University, India. Subsequently, moved to boston for his graduate research, working with Prof. Luca Dal Negro at Boston Univeristy on light-matter interaction in disordered 2D structures for which he was awarded the Best dissertation award form college of engineering at Boston university. Since 2011, he has been working at Caltech as a postdocal fellow with Dr. Paul Rothemund. His research is at the intersection of nanofabrication, photonics and synthetic biology focused on engineering novel physical properties by controlling the spatial arrangements of molecules/emitters/atoms in 2D and 3D environments.
Dr. Michael Fazio, SLAC National Accelerator Laboratory, Stanford
The 21st Century Frontier of Particle Accelerator Technology
Discovery science in fields such as high energy and particle physics, ultrafast chemistry, and modern biology rely on probes consisting of energetic particle beams and intense photon beams. These “tools” are typically based on particle accelerators that use radio frequency (RF) energy in resonant structures to produce electromagnetic fields that guide and accelerate particles. Rolf Wideröe demonstrated the first linear accelerator using RF energy in 1928. However, the principle of RF particle acceleration based on resonant cavities was first demonstrated in the 1940s by William W. Hansen, who invented the electromagnetic resonant cavity. His accelerator architecture has become the basis of all modern linear accelerator technology; the first demonstration achieved 1.6 MeV of electron acceleration in a 1-m-long section. The evolution of accelerator design and engineering over the ensuing decades has achieved tremendous improvements: higher energies, higher currents, and multiple ion species, etc. that continue to produce great discoveries. Using these particle beams as drivers for free electron lasers we have also been able to create very intense, coherent photon beams from infrared through x-ray wavelengths that have opened up new areas of research.
However we now find ourselves at the point where the next generation in performance is virtually unaffordable. To build the next generation of accelerators we must create a significant shift in the capability-cost curve for these machines and the sources of RF energy that power them. This talk will briefly describe current and proposed discovery-class accelerator-based facilities for particle physics and photon science to establish the perspective for the rest of the talk: the research and development underway at SLAC that seeks to create the next generation of accelerator and RF source technologies needed to advance the capability of discovery science tools and to enable a range of emerging applications from medicine to national security.
Dr. Michael Fazio is the Associate Laboratory Director for the Technology Innovation Directorate at SLAC. The Directorate comprises the Laboratory’s core capabilities in high power RF accelerator research and engineering, pulsed power, advanced instrumentation, detection systems from microwave through gamma-ray wavelengths, and high speed data acquisition systems. The Directorate’s mission is to develop advanced technologies that enhance SLAC’s discovery science programs and to innovate and incubate new technologies that address national needs and emerging applications. He also serves as a member of the Los Alamos National Laboratory’s Matter-Radiation Interactions in Extremes (MaRIE) External Advisory Board.
He came to SLAC in 2010 serving as the first director of the RF Accelerator Research and Engineering Division responsible for SLAC’s integrated high power RF capability, where he guided the development of innovative RF sources, accelerators, pulsed power systems, and ultra-high-power components from concept through design, fabrication, testing, and operation.
For 34 years, 1978 to 2012, he worked at the Los Alamos National Laboratory in numerous research and leadership positions. In 2005, he assumed the leadership of the Intelligence, Space & Response Division, holding line and program responsibility for satellite-based nuclear explosion detection and treaty monitoring instrumentation, NASA-sponsored planetary exploration and space science, proliferation detection technology, astrophysics, remote sensing, information science, and directed energy. He was responsible for the Nuclear-Nonproliferation R&D Program and the Center for Space Science and Exploration.
Prior to 2005 he led the Los Alamos High Power Electrodynamics Group engaged in RF source, free-electron laser, advanced accelerator, and compact pulsed power R&D, and the application of these technologies. Under his leadership this group developed the Laboratory’s FEL and microwave directed energy programs.
Dr. Fazio received two “Awards for Excellence” for contributions to counter-proliferation work from US Government sponsors, and in 2013 a commendation from the NASA Administrator for his role as a member of the Mars Curiosity (Rover) ChemCam Instrument Development and Science Team “for exceptional achievement defining scientific goals and requirements, developing the instrument, and operating it successfully on Mars.” He has served on advisory committees for the DOE Office of Science, DOD, DARPA, and on the Air Force Scientific Advisory Board. Dr. Fazio received his Ph.D. in electrical engineering from Rice University in 1979.
Dr. Ranveer Chandra, Microsoft
FarmBeats: IoT for Agriculture
Food requirements are expected to double by 2050 to meet the demands of the world population, but the amount of land fit for agriculture is shrinking. Data-driven techniques, such as precision agriculture, could help meet the increased demand. In this demo, we will present FarmBeats, an agricultural IoT system that uses a combination of unmanned aerial vehicles (UAVs) and wireless sensors to enable data-driven agricultural techniques. In doing so, we develop novel algorithms to maximize the coverage of UAV flights given limited battery power, to combine information from a UAV’s video and ground sensor data, and, finally, to achieve cloud connectivity of the farm’s monitoring system while respecting the harsh bandwidth constraints imposed by the farm’s backhaul link to the Internet.
Ranveer Chandra is a Principal Researcher at Microsoft Research. He is leading an incubation on IoT Applications, with a focus in Agriculture. He is also leading research projects on white space networking, low-latency wireless, and improving battery life of mobile devices. Ranveer has published more than 60 research papers and filed over 100 patents, 65 of which have been granted. His technology has shipped as part of Windows 7, Windows 8, Windows 10, XBOX, Visual Studio, and Windows Phone. Ranveer has won several awards, including the MIT Technology Review Top Innovators under 35 (TR35 2010), best paper awards at ACM CoNext 2008, ACM SIGCOMM 2009, IEEE RTSS, and USENIX ATC, the Microsoft Graduate Research Fellowship, and Fellow in Communications of the World Technology Network. Ranveer has an undergraduate degree from IIT Kharagpur, India and a Ph.D. in Computer Science from Cornell University.
Prof. Peter Asbeck, University of California, San Diego
Microwave and Mm-Wave Power Amplifiers for Fun and Profit
Efficiency and linearity of power amplifiers are critical for wireless communication systems. As the world moves from 4G smart-phones to even more broadband 5G systems, there is a major question about whether the III-V based power amplifiers in your pocket today will be replaced by something else. This presentation reviews some of the system considerations for emerging 5G technology, and their implications for the RF front-end circuits; it presents advances made in power amplifiers at 2, 5, 15, 28, 45 and 73GHz, with a particular focus on application of scaled CMOS; and it discusses pathways for future development of 5G IC solutions.
Peter Asbeck is the Skyworks Chair Professor in the ECE Department at UCSD, La Jolla, CA, and a member of the UCSD Center for Wireless Communications. His research focuses on high frequency transistors and circuits, particularly power amplifiers for wireless communications.
Dr. Asbeck is a member of the NAE and received the IEEE Sarnoff award based on his work in power amplifier transistor technology.
Some of most compelling application domains of the IoT and Swarm concepts relate to how humans interact with the world around it and the cyberworld beyond. While the proliferation of communication and data processing devices has profoundly altered our interaction patterns, little has been changed in the way we process inputs (sensory) and outputs (actuation). The combination of IoT (Swarms) and wearable devices offers the potential for changing all of this.
The Human Intranet proposes an open scalable platform that seamlessly integrates an ever-increasing number of sensor, actuation, computation, storage, communication and energy nodes located on, in, or around the human body acting in symbiosis with the functions provided by the body itself. The traditional set of senses and interactions is to be augmented by a set of new capabilities, some of which might be hard to even imagine today.
Jan Rabaey received his Ph.D degree in applied sciences from the Katholieke Universiteit Leuven, Belgium. In 1987, he joined the faculty of the Electrical Engineering and Computer Science department of the University of California, Berkeley, where he now holds the Donald O. Pederson Distinguished Professorship. He has founded and directed a number of influential research labs, including the Berkeley Wireless Research Center (BWRC) and the Berkeley Ubiquitous SwarmLab.
Prof. Rabaey has made major contributions to a number of fields, including advanced wireless systems, ultra low-power design, sensor networks, and configurable ICs. His current interests include the conception and implementation of next-generation integrated wireless systems over a broad range of applications, as well as exploring the interaction between the cyber and the biological world. He has authored a number of very widely used textbooks.
He is the recipient of a wide range of major awards, amongst which the IEEE CAS Society Mac Van Valkenburg Award, the European Design Automation Association (EDAA) Lifetime Achievement award, and the Semiconductor Industry Association (SIA) University Researcher Award. He is an IEEE Fellow and a member of the Royal Flemish Academy of Sciences and Arts of Belgium, and has been involved in a broad variety of start-up ventures.
Prof. J. Sebastian Gomez-Diaz, University of California, Davis
Advanced Manipulation of Electromagnetic Waves with Nanostructured Surfaces
The exponential expansion of the information society is continuously imposing stringent – and sometimes even contradictory – technological requirements to modern communication systems, including high communication data rates, efficient use of the spectrum, ubiquitous wireless connectivity, and many functionalities integrated into reconfigurable, miniaturized, and wearable devices. To meet such requirements, it appears today clear that novel and advanced approaches –rather than the simple optimization of existing solutions– are needed. In this talk, I will discuss the unprecedented possibilities offered by ultrathin nanostructured surfaces to overcome such limitations and exhibit exciting functionalities thanks to the use of tunable, active, and nonlinear materials and enhanced wave-matter interactions. I will first focus on graphene and other 2D materials as a powerful reconfigurable platform for THz and infrared plasmonics, describing novel components such as modulators, waveguides, and antennas as well as unusual non-reciprocal responses based on the spatio-temporal modulation of graphene’s conductivity. Then, I will present ‘hyperbolic metasurfaces’ able to exploit strong anisotropic behavior over uniaxial structures, demonstrating extreme topological transitions and a dramatic enhancement of light-matter interactions with application in imaging, hyperlensing, and communications. Next, I will introduce a ‘flat non-linear paradigm’ able to simultaneously exhibit a record high second-order nonlinear response from plasmonic metasurfaces tied to multi-quantum wells and sub-diffractive phase control. Such combination paves the wave to the efficient generation of pencil-beams steered in arbitrary direction in space, vortex beams, and focusing. The last example will propose a new type of ultra-fast infrared sensor based on ultrathin nanomechanical resonators able to provide unprecedented electromechanical performance and thermal capabilities. I will finalize the talk by outlining my vision for the near and long-term future of ultrathin metasurfaces and their potential impact on society.
Juan Sebastian Gomez Diaz is an Assistant Professor in the Electrical and Computer Engineering Department of the University of California, Davis. He received his Ph.D. degree in electrical engineering (with honors) from the Technical University of Cartagena (UPCT, Spain) in 2011. From October 2011 until March 2014 he was a postdoctoral fellow at the École Polytechnque Fédéral de Lausanne (EPFL, Switzerland). Then, from May 2014 to August 2016, he continued his postdoctoral work in the Metamaterials and Plasmonic Research Laboratory of The University of Texas at Austin (US). He has co-authored 50 journal papers, some of them published in highly selective journals such as Nature Communications, Physical Review Letters, and IEEE Transactions/Letters, 60 conference papers and 1 book chapter. His main research interests include multidisciplinary areas of electromagnetic wave propagation and radiation, metamaterials and metasurfaces, plasmonics, novel 2D materials, antennas, non-linear phenomena, and other emerging topics on applied electromagnetics and nanotechnology.