Thank you to all the 2015 participants and speakers! Please see video footage here and photos from the event here.
The School of Information Theory will bring together over 100 graduate students, postdoctoral scholars, and leading researchers for four action-packed days of learning, stimulating discussions, professional networking and fun activities, all on the beautiful campus of the University of California, San Diego (UCSD) and in the nearby beach town of La Jolla.
- Tutorials by some of the best known researchers in information theory and related fields
- Poster presentations by student participants with feedback and discussion
- Panel discussion on “IT: Academia vs. Industry Perspectives”
- Social events and fun activities
9:00am-12:00pm: Syed Jafar, UC Irvine
Interference Management -- An Information Theoretic View
2:00pm-5:00pm: Stephen Boyd, Stanford
Convex Optimization and Applications
7:00pm: Diego Mesa, UCSD
Introduction to IPython Scientific Computing Environment
Todd Coleman, UCSD
Mentoring & Success Panel Discussion
9:00am-12:00pm: Venkatesan Guruswami, CMU
List and local error-correction
2:00pm-5:00pm: Urbashi Mitra, USC
Biological Communication Channels: Engineered and Natural
6:30pm: Banquet at 15
9:00am-12:00pm: Paul Siegel, UCSD (Padovani lecturer)
Constrained Codes for Multilevel Flash Memory
» Click here for presentation
On Thursday, leading researchers from major telecommunication companies will describe their vision of the next wireless communication systems. Please join us for this exciting preview of things to come.
9:00 - Farooq Khan, President, Samsung Research America
ARCHITECTING Tb/s WIRELESS
Wireless communications along with the Internet has been the most transformative technology in the past 50 years. We expect that wireless internet growth driven by the need to connect all humankind (not just 1/3) as well as Billions of things to the Internet will require Terabit/s shared links for ground based local area and wide area wireless access, for wireless backhaul as well as access via unmanned aerial vehicles (UAVs) and satellites. We present a new scalable radio architecture that we refer to multi-comm-core (MCC) to enable low-cost Terabit/s wireless using both traditional and millimeter wave spectrum.
9:50 - John Smee, Senior Director, Qualcomm
5G Design Across Services
Mobile communications is evolving quickly with significant increases in the number of connected devices and applications demanding higher levels of performance. To support these expanded connectivity needs, the industry is actively engaging in 5G research and development and upcoming standardization. This presentation will introduce a 5G Unified Air Interface (UAI) design approach for supporting diverse requirements across licensed, unlicensed and shared licensed spectrum in both sub-6GHz and above-6GHz bands including mmWave. We describe several key UAI PHY/MAC techniques, including Massive MIMO, self-contained subframes, device-centric MAC, and low latency and high-reliability design, as well as network architecture improvements. The combination of advanced techniques is focused on providing not only the flexibility to support growing use cases and spectrum bands but also the performance, capacity, and scalability required for cost effective 5G services.
10:40 - Bin Li, Senior Technical Expert, Huawei
POLAR CODES FOR 5G
We present an class of enhanced Polar code structures: the Polar code is concatenated with CRC and is decoded by the adaptive SC-list decoding. Such code structures have better performance, lower power consumption, lower latency than the 4G-LTE turbo code. Polar code is a potential candidate for channel coding in 5G system.
11:30 - Lunch
12:50 - Tom Marzetta, Alcatel Lucent
MASSIVE MIMO AND BEYOND
Massive MIMO, a Fifth Generation wireless technology, owes an enormous debt to Shannon Theory. But the foundations of Massive MIMO, in fact, comprise three engineering principles: 1) the employment of directly measured channel state information yields beam-forming gains that are proportional to the number of service antennas irrespective of the noisiness of the CSI, 2) the effective channel which is created for a user is substantially frequency independent, so simple frequency independent power control ensures uniformly great service everywhere in the cell, 3) pilot contamination imposes an ultimate limitation on the number of mobile users who can simultaneously be served. Future social media services, such as ubiquitous sharing of holographic video among augmented-reality equipped acquaintances, will motivate the commercial introduction of Massive MIMO. More serious, non-entertainment applications will emerge. Physical data governed by partial differential equations should be sampled at the Nyquist rate over both space and time, and Massive MIMO permits continuous collection of uncompressed data from unprecedented numbers of sensors. Massive MIMO as well as Point-to-Point MIMO can potentially work in non-electromagnetic media. The single most critical question in communication theory is the following: does Massive MIMO represent the end of the line for wireless communications, or is there something better?
1:40 - Ivana Maric, Senior Researcher, Ericsson
INFORMATION THEORETIC CONCEPTS OF 5G
I will review several 5G scenarios that lead to information theoretic problems, and will then focus on two of them.
In the first scenario, I will consider communications in wireless networks with multiple relays. These networks will find their applications in 5G systems to enable mesh networking, e.g. in wireless backhaul at high frequencies and massive machine-type-communications. For such networks, noisy network coding (NNC) (and quantize-map-and-forward) can achieve the cut-set upper bound within a constant gap. In 5G multihop applications, however, this gap may not be negligible. Furthermore, the implementation of NNC is challenging. I will present a low complexity NNC scheme that overcomes these problems and can lead to practical solutions for 5G mesh networking.
I will then discuss channel coding problems arising in 5G, and will present a design for rate-compatible polar codes.
Based on joint work with Song-Nam Hong, Dennis Hui and Giuseppe Caire.
2:30 - Closing
- Massimo Franceschetti, UCSD
- Sree Goparaju, UCSD
- Tara Javidi, UCSD
- Young-Han Kim, UCSD
- Victoria Kostina, Caltech
- Alon Orlitsky, UCSD
- Paul Siegel, UCSD
- Gerhard Kramer, TU Munich
- Ramesh Rao, UCSD
- Alexander Vardy, UCSD
- Aylin Yener, Penn State University